Related Links:
Camera and Lens Testing
Lens Faults Hierarchy
Lens Test Chart and Info Site (John Chap)
Lens Testers Anonymous (Stephen Gandy's classic site)
Wide Variations in Samples of Nikon's 35-105mm Lens

One of the most common questions on rec.photo newsgroups is How Good is My Lens?.

My answer has to be that the only way to tell is by testing the lens you have. The above chart shows how much several used lenses from the same batch can vary. If these lenses were really identical, then you should see only three similar curves for the center, middle, and edge resolution of the two lenses. But the lenses vary, by quite a lot. So we end up with some very different curves.

This observation is also why photography magazine test reviews are less useful than many people suggest. Lenses from the same batch can easily vary significantly, as shown above and as documented in other tests below (for Kowa lenses).

So how can you tell if the lens they tested was an average one for the batch, or at the low end, or a superior lens from the high end in performance terms? You can't, and with only one lens tested, neither can they! Even worse, you can't tell whether your own lens is similar to the one tested, or worse, or perhaps a lot better too. The only way you can tell is by testing each lens!

Another lesson from the above curves lies in the performance differences between these two lenses. The first lens has fabulously high center and middle performance, one of the two best medium format lenses tested in this study. But this same lens has the worst edge performance, down to the point of being unacceptable. By contrast, the second lens has very good center and middle resolution, and is much higher in edge performance.

Can you also tell that the best overall performance for the second lens is at f/16 from the chart? Can you see how diffraction effects limit the lenses at f/22 and f/32? Or that the worst performance for the first lens is wide open? Clearly, individual lens test data can help you see where your lenses perform best, and how well.

Thanks to Pentaxnut for data from Modern Photography tests
1973 lens s/n 565xx versus 1974 lens s/n 651xx

Lens Variations

One reason this is true is that even new lenses get abused. Some new lenses will get dropped in the store or by the gorillas in the mailroom.

A lens that has been dropped or abused could be out of alignment, which would seriously compromise its optical performance. Unless you test it, how will you know?

Believe it or not, but mass produced lenses are usually only batch tested for quality.

That means only a few lenses per hundred are actually tested for performance and quality. The other 98.5% or so of the lenses go into the box without testing. You may be the first one to try the lens out.

Welcome to the world of consumer quality control - where you, the consumer, perform the final quality control testing!

Despite the marketing hype, you will find that many OEM lenses will also have a startling range of problems. OEM problems are similar in magnitude and number to the top third party lens makers too. Check any OEM specific newsgroup (e.g., Nikon digest, Canon Digest..). You will see frequent postings complaining of problems with lenses right out of the box! These problems range from minor ones such as dust in the lens to loose or even shattered front elements.

My point here is that you could pay big bucks for an OEM lens and get a lemon. It happens every day. You could pay big bucks for a third party lens, and get a lemon too. Nobody's perfect!

Did you ever stop to think of what happens to all those lenses that get returned to the mail order and camera store dealers? Do you think they test the lenses to see if there really is a problem? No, huh? Do you think they re-box them and send them back out to the next guy? Could that be you? Did you get stuck with somebody else's lemon?

That puts the burden on you and me to test out our lenses individually to ensure that they are functioning properly and don't have any internal, hard to see problems.

To help you do that, I have collected a bunch of lens and camera testing tips from many articles and online sources at my Camera and Lens Testing Page.

When I saw this interesting series of comparisons, I thought they would prove my point on this page. Lenses do vary, so you have to test yours to be sure it measures up to your needs and expectations.

These four Kowa 50mm f/1.9 lenses were randomly chosen, the first one having been reviewed (Oct. 1964 Modern Photography) and the last three from the same batch as shown by their serial numbers.

A quick check shows these lenses really do vary considerably. Lens #834 is the best wide open and stopped down, and has the best edge sharpness. the lens originally reviewed in October 1964 did a bit better in terms of edge sharpness than #464. But both #761 and #834 provide more high marks.

For me, the most telling point is that you might have rejected this lens and camera, based on the initial October 1964 review that got published. On the other hand, you might have bought this lens for its great edge performance, based on #834 ratings. How can you or I tell if the lens tested and reported on in any given magazine review is representative of the lens we ultimately buy? (Hint: test it!)

If you saw such comparative ratings, you would probably be willing to pay quite a bit extra for #834 versus the October 1964 review lens or #464. You might even be willing to pay up for the differences between #761 and #834, the best of the series.

But obviously, the last three lenses are all from the same batch, by the same manufacturer (Kowa), and for sale at the same price!

This table puts many magazine lens-test users into a considerable bind. Kowa Optical Corp. was (and is) one of the best lens makers in Japan, and acted as an OEM (like Nikon, Canon) for its own 35mm (Kowa SETr) and 6x6cm (Kowa 6/Super 66) cameras. You should probably expect similar variations from the other lens makers who rely on batch testing (i.e., most of the Japanese and Asian OEM and third party manufacturers).

Granted, this study is from an older series of lenses. But the 50mm lens has fewer elements (and fewer tolerance variation sources) than a typical zoom lens.

Moreover, if you are not buying new lenses (such as those tested above) but used ones, your lens variations would be much greater. Optical mis-alignment can arise due to sharp impacts (possibly without showing damage, as on lens hoods). Thermal cycling may have caused elements to creep or shift in their mountings too. The lubricants and greases may have picked up dust and sand grit, which has worn their mechanical focusing systems. In short, lenses probably don't get better with age!

So I suggest that this table is proof that you probably need to consider testing your lenses. Moreover, you and I should probably take those lens test charts and reviews with a bit of caution. They may reflect the status at one new lens, but not accurately reflect the one we have in our hands or on our cameras!

Many photographers will admit that budget lenses probably vary, due to cost constraints and sloppy tolerances. But they will then insist that the better and more costly lens brands are very carefully optimized and made to ensure each lens performs optimally. So here we have two lens tests of brand new lenses, by the same testers and standards, for two sets of Hasselblad Zeiss lenses just three years apart in production.

The first set were part of a review of the 500c (August 1965 Modern Photography) for six lenses. The 80mm f/2.8 zeiss planar (#3510705), 120mm f/5.6 Zeiss S-planar (#3335955), and 150mm f/4 Zeiss Sonnar (#3459707) results are reported above. The second lens series (labeled #2 above) are for the same focal length lenses tested with the 500EL some three years later in August, 1968. These lenses are the 80mm#2 (#4688517), 120mm#2 (#4301626), and the 150mm#2 (#4221295) also listed above. Note that the earlier 120mm lens did not have a f/45 setting on the test chart.

Take a look at the chart, comparing the center and edge data for each of the two lenses sampled above. You can hardly argue that these lenses are "as alike as peas in a pod", as many advocates would argue, right?

Wouldn't you be willing to pay a good bit extra for the first 80mm f/2.8 lens, with 12 our of 14 ratings of "excellent" (and the other two "very good"), over the 80mm#2 lens with only half as many excellent scores, and 4 scores that were just "acceptable"? That is a pretty big difference in performance, yes? Similarly, the 150mm lens lens outperformed the second 150mm#2 lens in six categories by one full rating.

Critics might suggest that a lens that is marginal might be rated downwards or upwards depending on the tester. Tf so, it is hard to explain why all the downward ratings are in the second lens series rather than randomly distributed, as you would expect if the lenses were alike. But the 120mm lens tests make it clear that the differences between lens samples can be quite major. We see above examples of not just one category difference, but two categories and even three(!) categories worth of difference. The second 120mm#2 lens under-performed the first lens in over half the categories, and by not just one but two ratings twice (e.g., excellent vs. good) and even three ratings once (e.g., exc vs. acceptable). Yikes!

I argue elsewhere that much of the higher price of Zeiss lenses is a result of markups and marketing tiers, specifically the extra markups taken by Zeiss as the manufacturer to Rollei and Hasselblad, who then add on their own markups. So by comparison, lenses for Bronica SQ (6x6cm leaf shutter SLR) are about 40% less than Zeiss 6x6cm leaf shutter lenses for Hasselblad and Rollei. In the USA and other countries with an "official Hasselblad" importer, the importer can add another layer of markups. So these lenses end up costing a lot more not because a lot more was spent in making or testing them, compared to say Bronica lenses. Rather, they cost a lot more because several more layers of middlemen and markups have been added on to the price. On our grey market gear pages, we document how savings of 40% to 60% are possible on a number of lenses from overseas sources which bypass some of these markups and costs.

What these lens tests do tell us is that even the most costly lenses vary, by an often surprisingly large degree, between samples of lenses. Folks who suggest that the more pricey and big name lenses are made to some high and identical tolerances have to explain how such lenses can vary so radically when tested by the same testers and charts and standards.

Finally, these kinds of differences can be readily seen in comparing lenses. For example, I have three 75mm f/2.8 Bronica nikkor lenses, which I have used in a blind lens test. The results show that two of these lenses have been selected repeatedly as among the three lenses out of a sample of lenses including Hasselblad and Rolleiflex zeiss lenses among others. But the third lens is consistently downrated among the worst lenses in the test (and it is so on test charts as well). Here again, the lenses vary, and so do the slides....

We hereby publish two complete lens tests for samples of this lens. The first just seemed too good to be true, so we tested a second as well... Ibid., p. 90.

Ah, pity the poor photographer who doesn't bother to test his lenses, placing his or her faith in the manufacturers. Here we see two Spiratone lenses with sequential serial numbers (#099E and #100E). They should be virtually identical, right? Ooops! The lens test data clearly shows they differ - and by a lot.

Many of us would be willing to pay a good bit more for the first lens (#099E), which garnered all "excellents" in edge performance (versus only 2 out of 7 for the other lens). The same lens (#099E) had only "very good" or "excellent" ratings for its central resolution scores too. Again, the other lens had five ratings of only "good", and only one "very good" and "excellent" rating for central resolution. Not only do these lenses differ, but they differ by a large margin, yes?

Most current users would pass over an offering of such a Spiratone 135mm f/2.8 lens. But thanks to the preset lens design, you can get a lot of lens for your money (here, $29-37 with T-mount). If you test your lenses and explore such potential lenses, you might luck out and find one as good as #099E! Then again, if you don't test your lenses, you may end up with one like #100E.

The 135mm lens is relatively easy to design and make (versus say a fast lens or an ultrawide lens). So you have surprisingly good odds of finding a decent or better 135mm lens, with a bit of testing, especially now that such lenses are out of favor due to the prevalence of zooms. So good hunting!...

Here is yet more proof that lenses vary, even when picked up new from warehouse stock. The three Konica 45mm f/1.8 Hexanon lenses are serial numbers #1072583, #1072577, and #1070667. Note that the two closest s/n lenses #1072583 and #1072577 varied somewhat, while #77 and #67 provided very similar test ratings. The latest lens (#83) varied from both its mate from the same batch and from the earlier batch lens too. While the #83 lens did generally slightly poorer in the corners past f/5.6, it was a bit sharper at f/2.8.

Now what happens if we take the same three lenses, and send them off to the National Bureau of Standards to be tested? We get the results reported above (Source: Modern Photography, October 1965, p. 124). As before, #77 and #67 performed very similarly. Again, #83 was the odd lens out. But the patterns are quite different. With the first table, #83 differed by being lower resolution at some of the smaller f/stops (e.g., f/11). But with the second table, #83 was higher esp. in the edges at the wider f/stops (and even centrally at f/2.8). The overall pattern, esp. for the #77 and #67 lens tests, are very similar despite many differences in charts, films, and development between the two test series.

Modern Photography concluded " Surprise! The tests conducted by MODERN and the N.B.S. tests produced almost identical evaluations for all three Hexanon lenses....

You may feel like me, and decide that the patterns for the #83 lens are a bit less similar than we might like. But their underlying point is that the overall patterns should be similar, if the tests are accurate, despite differences in the actual values between any given lens. We can see that #83 is generally similar to the other two lenses, but not identical to them (hence, more proof lenses vary). Moreover, you can see that different lens tests may give very dissimilar lpmm values. But the patterns should be similar between the two tests on the same lens(es) if they are to be consistent tests.

We should not be surprised, on the basis of this test, that different test charts will yield different results. The Sleicher chart is different from the USAF test chart, and so yields different numeric results. Moreover, there is a subjective element in deciding if a block of lens resolution bars is too blurry to count, or just on the edge of being resolved and so it gets counted. So you can't take a set of resolution values from Chris Perez's tests, for example, and directly compare them against resolutions published in Modern Photography which involved a different chart, film, development cycle, standards, and observers.

Most of us use a simpler approach. We study and collect sources of lens test data and ratings, on and off-line, published and posted. When I check Chris Perez's medium format lens tests, I see that the two highest rated lenses were for the Mamiya 6 and 7 rangefinders. Now I happen to know that these cameras has very positive user reports on how sharp or high resolution the lenses are. So these results seem pretty consistent with both user reports and other published test values. Now I see that Chris also rates the Koni-Omega 6x7cm rangefinder very highly too. This observations suggests to me that I ought to check out the Koni-Omega, as it is a budget 6x7cm rangefinder camera ($200 up) with interchangeable lenses and backs.

More Proof Used Lenses Vary - A Lot! Three Koni Omega 6x7cm 58mm Lenses Compared (Data Source)

Source: Feb. 1986 and Jan. 1982 Modern Photography lens reviews

Here again, we have two 50mm f/1.8 Series E lenses by Nikon, tested in 1982 and 1986. The later lens scored rather higher in edge resolution, but slightly lower in center resolution. So the high and low curves above belong to the same lens (oops!). Note also how relatively consistent the second (middle curves) lens is in performance, with only a modest falloff in resolution in the edges.

My other point here is that these series E lenses are the economy (hence, "E" for economy) lenses in the nikon line. They are plasticky, single coated, lack aperture 'ears' for older camera bodies, and sell for $20-35 used. So how come our cheapy Series E lens equalled or beat the Leica 50mm f/2 summicron in 12 out of 14 comparisons (despite the nikkor being a faster lens to boot)? Oops?! ;-)

More Examples Showing Lenses Vary

Okay, by now even you should be convinced! Lenses do vary, and often by a surprising amount, even if made from the same basic formula or design. We have abstracted the above charts from the full lens tests by Modern Photography, mainly because these tests were done using the same charts and standards over a long period of years. As new camera body models came out (e.g., Fujica ST801 and ST901), they tested the new bodies and the latest production model of lens on each body.

I have also included some of the most popular and big-name brands above, such as Nikon, Canon, Pentax, Olympus, Contax/Zeiss, Minolta, and Leica. The Leica chart also lets you compare the rangefinder (M series) lens against the reflex SLR (Leicaflex SL and R series) lenses for fun too. Chinon, Konica, Fujinon, and Miranda lenses are also used as examples of lens variations for their normal lenses.

Used Lenses

Used lenses can have all of the problems of new lenses, plus add on some problems of their own.

The obvious difference between a new and used lens is that the used lens has an unknown history of use and possible abuse by any number of past users before it reaches your hands.

Sometimes, this lens abuse can be pretty obvious. When I see a bad ding on a filter ring, it tells me that this lens was probably either dropped some distance to a hard surface or did the Arc de Tripod.

What's the Arc de Tripod? That's what happens when your camera is mounted on a tripod, and somebody (usually you) hits it and knocks it over. The camera goes through an arc and then hits the ground, usually pretty hard if it starts from chest or eye level.

But not every dropped lens is badly dinged. You can find posts by users of lens hoods praising their lens hoods for taking the impact hit and being destroyed, rather than the lens it was mounted on. But that lens may still have gotten enough of a hit to knock it a bit out of alignment.

Bad Lenses Change Hands

Just as there can be some lenses slightly out of alignment direct from the manufacturers, some lenses can also randomly end up in near perfect alignment.

Naturally, this lens variation causes arguments between the owners of the lousy alignment lenses with the owners of the near perfect lenses of the same model.

What you can understand is that the good lenses are keepers, and rarely change hands unless removed from the dying grasp of their owner's hands.

The bad lenses are losers, and get passed around from one user to the next. Finally, they end up in the camera bags of vision impaired photographers or those who never do enlargements beyond 3x5 inches.

One consequence of this reality is that a few bad lenses due to quality control problems causes a ripple-down effect. The bad news ripples down from one user to the next, spreading the bad experience around.

Slip-streaming

Slip-streaming happens when a manufacturer changes a lens design, usually to correct a defect or problem, and doesn't tell anybody about it. Hey, would you want thousands of photographers sending back their used lenses and demanding the new fixed ones at no cost? (cf. Intel Pentium bug)

I have posted a table of lens flare values for various brands of OEM and third party lenses from Popular Photography. In performing this study, they were shocked to discover that some of the older Pentax Takumar lenses performed just as well as the latest Super-Multi-Coated versions. What was going on here?

The answer was that Pentax mounted some of the newer lens elements in the older lens mounts to use up their surplus of older lens mounts. Some lucky buyers got the newer SMC coated lens elements in the older and cheaper lens mounts.

When Popular Photography bought an old and a new one to compare, they actually got two of the same SMC lens optics lenses, just in two different mounts. What if you happened to get one of those slip-streamed improved lenses in the old mounts? You would also be surprised at how much better they were than your older lens too!

The good news is that lens manufacturers are constantly trying to make small incremental improvements in their lenses during their production runs.

The bad news is you can't tell if your lens is one of the earlier ones from before they learned of some problems and how to fix them.

The dark side of slip-streaming is that some of those improvements are made to reduce the cost of the lens, and raise the maker's profits, rather than make the lens better. Sometimes an internal metal part will get replaced by a cheaper molded plastic part. Plastic can reduce not only the cost and weight of the lens, but even improve the lens if used properly. But you can also use plastics and other tricks to reduce the costs of lenses, sometimes at the expense of long-term usability.

You can also find cases where a slight redesign enabled the manufacturer to remove one or more elements out of a lens. In my book, those lenses are different optical designs, and different lenses. But you couldn't tell without looking at the serial numbers and having internal company information.

My point here is that even two lenses that look identical externally can be different internally, due to changes in design over time.

You can't tell without testing the lens if it is better or worse than the magazine or other user lens tests suggest.

Who made it, and Where

Lenses don't care who made them. All that matters is the quality of the design and its implementation in optical glass and mechanical tolerances.

You might be surprised to discover that there are over twenty third party lens makers in Japan alone. These independent third party lens makers, along with other Asian lens manufacturers in Korea, China, Malaysia, Indonesia, etc. make most of the various third party "Japanese" lenses imported into the U.S.

Suppose Vivitar Corp. comes up with a new lens design. They may contract out production to a local Japanese third party lens maker from the above list. Sales take off. Vivitar needs more lenses to sell. Their behind-the-scenes third party lens maker can't keep up with the demand. So they may hire another production line from another third party lens maker to make the same lens.

Perhaps that factory is in Japan? China? Korea? You won't know who made your lens, since they all say Vivitar on them. So long as the lens is made of the exact same stuff to the exact same specifications, neither you nor the lens should care. But do you think two factory production lines have exactly the same quality ethic?

Suppose Vivitar makes the lens in one factory in Japan using one of these behind-the-scenes independent third party Japanese lens makers. Later, they move the production to a lower labor cost country such as Korea. Labor costs in Korea are heading up. So they move production to China or Malaysia. All those moves and changes in labor and factories can impact the quality variations in their lens lines.

But you or I have no way to tell, except by testing our individual lenses.

Cherry-Picking

The idea is simplicity itself. You bring in your camera, some film, and some paper to take notes (and perhaps your favorite lens test chart too). Now simply take each lens, record its serial number, mount it on your camera and tripod, and shoot some sample photos at a series of settings. Repeat this process for as many lenses as you like or as may be available.

You can also do this same test with different brands and models of lenses to really see for yourself what the differences are between each lens, each model, and each brand.

Get the film developed. Label the slides so you can see which lens made which image. Check the slides with a closeup loupe (I use 15x, some use 30x). The test conditions are as nearly identical as possible. The same film roll is used for side by side tests at each aperture, controlling for processing too.

Now see if you can identify which lens is the worst model or example of the bunch you tested. Is there one lens which has better contrast, color fidelity, or superior sharpness in part or all of its range? Which one would you want to buy?

This approach will enable you to cherry-pick the best lenses from those lenses available in stock. For expensive professional lenses, retailers are surprisingly cooperative in making these tests. They can be really, really sure that after all that trouble, you are likely to buy the best of the lenses you tested by serial number from them, and not another mail-order or used lens dealer.

Who do you suppose gets the lenses that were duds, or also-rans? Duh?!!

If you buy via discount or mail-order, you won't be able to use this technique. But shouldn't you at least be testing the lenses you do get to be sure this isn't some defect? Before the 30-day return guarantee expires?

Summary

I have tried to summarize some of the reasons that explain why you just have to test the individual lens to ensure that you get a good one.

This recommendation is just as important for new lenses out of the box as it is for used lenses from a dealer or individual seller.

For used lenses, the history of the lens becomes increasingly important over time. Individual lens testing is even more important with used lenses.

So you just have to read our Camera and Lens Testing Page and use the tips and tricks you learn there to find out how good your lens really is.

Good luck!

Camera Lens Registration Variation Surprises!

How about lens registration distance as another source of variation in lens performance? Most cameras with the same mount (e.g., Nikon F) should have nearly identical lens registration distances. A small 0.2mm or less error in this critical focusing distance could cut potential maximum lens resolution in half!

Unfortunately, a series of Leica screw mount and Leica M mount camera bodies turn out to have substantial differences - over a millimeter! - in lens registration distances. See the table above for details.

Our point here is that such differences in lens registration distances can cause lens performance to vary hugely when the same lens is used on different M mount bodies. The lens may actually be a top performer, when used at the correct lens registration distance. But using a lens adjusted for an M-mount Voigtlander T camera on an M-mount Konica RF camera would place the lens 1.7 mm too far forward. The lens performance would be very poor (e.g., dropping from a potential of 57 lpmm to circa 22 lpmm or a 60% loss in wide open resolution performance!

The simple test for this problem is to use fine grain film and careful focusing with the lens wide open. Compare to a known good lens on the same camera body (e.g., OEM lens). If the wide open performance is rather less than expected, you should consider having the lens and camera body checked by a camera repair technician. An adjustment or a few shims may be needed to bring the lens into proper registration on your camera. The improvement in wide open and general performance from this adjustment could be very significant, as the above table suggests.

Now you also know why some folks may have reported poor performance from some optics that others were raving about how great they were. The same lens, used on different M-mount bodies, in this case, could yield greatly differing resolution performance. In the above table, switching the Leica 50mm f/2 lens from the Konica RF body to the Leica or Voigtlander T body provided a 250% (!) improvement in lens resolution wide open.

From other postings, I know that a variety of older lenses (e.g., screw mount) and especially Russian or Ukrainian lenses have been found to have lens registration distance problems. Knowing this, testing for it, and having a "bad" lens adjusted to yield peak performance can provide you with bargain lenses for your camera(s). Enjoy!

• Vivitar 24mm f/2.8 TX lens (#37xxxxxx)
  • less than +1 % distortion (vs. 4% standard)
  • aperture f/2.81
  • focal length 24.1mm
  • weight = 10 ounces
• Vivitar 28mm f/2.5 TX lens (#37XXXXXX)
  • less than -1 % distortion (vs. 2.5% standard)
  • aperture f/2.6
  • focal length 28.3mm
  • weight = 10.5 ounces
• Vivitar 35mm f/2.5 TX lens (#37XXXXXX)
  • less than -1 % distortion (vs. 2.5% standard)
  • aperture f/2.62
  • focal length 36.2mm
  • weight = 11 ounces
• Vivitar 135mm f/2.5 TX lens (#37XXXXXX)
  • less than +1 % distortion (vs. 2.5% standard)
  • aperture f/2.58
  • focal length 134.3mm
  • weight = 19 ounces
• Vivitar 200mm f/3.5 TX lens (#37XXXXXX)
  • less than +1 % distortion (vs. 2.5% standard)
  • aperture f/3.68
  • focal length 198.0mm
  • weight = 21 ounces

The point of listing the above figures is to illustrate how close to marked focal length and aperture most prime lenses really are, even old ones like these Vivitar TX lenses. Distortion is also under 1%, well under the typical 2.5% acceptable range (and 4% for wide angle lenses like the 24mm lens here). Few consumer zooms could match this precision!

How did these lenses perform on sharpness? The 24mm lens turned in center sharpness that was very good to excellent (57-64 lpmm), while corner sharpness was good (1) to excellent (5 out of 6; 32 to 51 lpmm). Usually, you would expect the 24mm to be the hardest lens to design, and the worst performer. Actually, it is rated as the best of the litter here!

The 28mm lens was also very good (3) to excellent (3 out of 6) at the center (53-66 lpmm). Surprisingly, the corner performance was above average, with one very good mark and 5 excellent ratings (33 to 47 lpmm). Surprise again, as usually you might expect corner performance to be the worst performer for a low-cost 28mm wide angle prime lens.

The 35mm was excellent on corner sharpness (39-50 lpmm) at all stops, but only excellent at f/2.5 and f/11 and f/16. The mid-range settings dropped to very good at f/4, acceptable at f/5.6, and good at f/8 at the center.

The latter 35mm lens performance is a good example of why I admonish folks to actually test their lens rather than rely on assumptions about lens performance. Perhaps this lens was a bad example, a lemon? But if you expected the 35mm to perform better than the 24mm or 28mm, being a simpler design, you would be wrong here.

Moreover, you would normally expect the best performance in the f/5.6 to f/8 mid-range apertures, right? For this 35mm lens, you would get the worst performance.

You would expect wide open to be the worst, but it was excellent.

You might also expect the corner sharpness to be rated worse if the center was only good or acceptable, but here it was uniformly rated as excellent.

In short, this lens shows why it is critical to test your lens for variations in performance!

How did the Vivitar TX telephotos fare for sharpness? The 135mm f/2.5 fast telephoto was excellent in the corner, but ranged from excellent wide open to good in the mid ranges, then excellent at f/11 and f/16.

Again, these results are somewhat surprising. You would normally expect a low cost telephoto non-APO lens to do poorly wide open and in the corners. But the 135mm f/2.5 did excellently in the corners, and also wide open or stopped all the way down to f/16 and beyond. Where you might expect it to be sharpest, it did the worst!

Similarly, the 200mm f/3.5 Vivitar TX lens is very good from f/3.5 to f/11 in the corner, but only good stopped down to f/16 and f/22. And the center sharpness is very good wide open but falls to good at f/8 and f/11 and f/22.

Surprise again! This telephoto lens is relatively sharpest wide open and gets worse as you stop down. The optimal range is from wide open to about f/8.

In summary, these Vivitar TX lenses violate most of our assumptions about how lenses generally should perform:

• The 24mm was the best performer, not the worst.
• The wide angle lenses did relatively best wide open or at f/11 and f/16, not in their mid-ranges.
• Both telephotos were best wide open (or stopped down to f/11-16 for the 135mm lens).
• Corner sharpness was rated generally excellent or very good on all lenses (except 200mm at f/16-22).
• Center sharpness varied more, usually with the middle f/stop ranges having the lowest rating.

Unless you tested these lenses, how would you know how they really performed?

My final point is about the relative value of these prime lenses. Note how many excellent and very good ratings they garnered. Check out the low distortion values. Consider sharpness wide open and in the corners. That's where you need it most. That's where these lenses perform best. So how come you can often buy these lenses for $20 to $35 US from dealers (without TX adapter)?

Related Postings

...

>Is there a way to determine the aperture value for a given lens
>that will provide the sharpest photograph?

Yeah, testing.....;-)

But, be aware that meaningful testing is not so easy as it would appear to be...

And, many lenses, especially very fast ones, and zoom lenses, vary considerably in their performance with both focus distance and with focal length with zooms, in addition to aperture...

Also, if reasonable sharpness in the frame corners is your standard for good sharpness (it is mine...), you will often come up with a different answer than center-only performance will give.

And, puh-leeze ignore that useless chestnut, "best at two stops down from wide-open" - it usually doesn't work except for f4 lenses! Without testing, with high-quality lenses, a reasonable assumption is that performance peaks in the center around f5.6-f8 for non-extreme lenses for the 35mm format (BR's mysterious findings notwithstanding... [I know he is a careful tester, which leaves me completely mystified by his findings, which are so different from mine, and from those of virtually all other testers...]).

If you do want to test casually, try running a detailed horizon line diagonally across the frame from corner to corner. Carefully (manually) focus. Without changing focus, shoot the same thing at all the available stops (compensating with the shutter speed to keep the exposures constant). Try the widest three stops again, with refocusing done for each set (do NOT assume that marked infinity on the lens is truly infinity-focus!) to see if you are checking focus accuracy or the lens... Use good tripod technique, or hand-holding with fast enough film and bright enough light to give you sufficiently fast speeds at all stops. Inspect the film directly, using a good 10X magnifier.

DO NOT get caught up in "tester's disease"! ;-)

David Ruether
ruether@fcinet.com
rpn1@cornell.edu
http://www.fcinet.com/ruether

From: Chip Louie chipl@thelouie.org
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How to determine sharpest aperture value?
Date: Thu, 29 Oct 1998

Hi Ed,

Testing on slow, fine grain, high resolution slide film and looking at the slides with a 20x loupe can easiliy show you which f/stop for a particular lens gives best sharpness. I've found that in GENERAL for primes up to about 135mm two, maybe three stops down from wide open will yield the sharpest images. Anymore than this and the image sharpness starts to fall off again. Beyond 200mm I have only tested fast telephoto lenses and they tend to be sharpest wide open or maybe stopped down one stop. I'm an EOS shooter and this is what I've found after testing a lot of different primes and zooms and combinations of extenders and telephoto and super telephotos. An interesting thing I found was that the Canon EF 70-200 2.8L is very good wide open and excellent at f/4 just like my EF 300 2.8L and a rented EF 400 2.8L. I don't know if this is true with Nikon glass or not.

HTH,

Chip Louie

From: "Bob Flood" jrflood@ricochet.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How to determine sharpest aperture value?
Date: Thu, 29 Oct 1998

>Testing on slow, fine grain, high resolution slide film and looking at
>the slides with a 20x loupe can easiliy show you which f/stop for a
>particular lens gives best sharpness.

I didn't quote the whole post in this case, but it's a good one. I'd just like to add one idea.

To test a lens or lenses, find an office building with its outside covered with windows that are all the same size - something like a 20-story (or more) building. Fill the frame with windows (lots of them) and shoot. You may need to move closer or farther to get a lot of windows in the frame when changing lenses.

When you evaluate the results, the window edges will give you a way to judge the sharpnes at various f-stops, and, because the frame is full of windows, you can judge sharpness all over the frame, including edges and corners. It even lets you compare one lens to another.

[Ed. note: changes in the middle of production run aren't new...]
Date: Fri, 13 Nov 1998
From: Marc James Small msmall@roanoke.infi.net
To: rollei@mejac.palo-alto.ca.us
Subject: Re: [Rollei] Zeiss coatings.

John Kufrovich wrote:

>    On the Zeiss rolleis, what was the reason  for multicoating the viewing
>lens and not the taking lens.  Why just the hard coat on the taking lens.

This was not the universal practice. By the books, the only Rollei TLR's to be multicoated are the 2.8GX's, but we have uncovered several 2.8F's with coated lenses dating from the early 1970's, which has led us to believe that Zeiss converted from a single-coating production line to a multi-coated line and, thus, the customer got multi-coated lenses whether they asked for them or not. (I would suspect they had to pay extra, though, for the "T*" marking.)

So, I suspect a camera with a multicoated viewing lens and single-coated taking lens would date from this transition period.

Marc

rec.photo.equipment.35mm
From: tom@graywolf.com (Tom Rittenhouse)
[1] Re: Highest Rated Prime Lens
Date: Mon Dec 28 1998

Grcolts grcolts@aol.com wrote:

: Regardless of brand name, which AF prime lens in the 28 to 50 mm range are
: considered the overall best??

Well, do to manufacturing tolarance, this changes minute by minute. Maybe Galen Grindes will do an article about this. A is better than B unless you get a particular good example of B. You guys remind me of those test reports. They test a lens with mediocre QC and think the will be all the same. Your 35-80 Nikon my be better than m 35-80 canon but my cousin's is not.

Pentax 50, 1.4.

Damn, I need to learn to sleep at night....

From: "Bob Salomon" bobsalomon@mindspring.com
Newsgroups: rec.photo.equipment.large-format
Subject: Re: lens testing
Date: Sun, 13 Dec 1998

Actually there isn't.

And even if there was you could not consistently and repeatably do 2 consecutive tests and probably could not maintain the standards accurately for a single session. That is why lens manufacturers and camera manufacturers stopped using this type of test and all went to MTF tests which are both consistent and repeatable. Even some magazines have invested in the equipment necessary for MTF testing.

The variables in chart testing are:

1: The lens may not have been designed and optimized to reproduce a flat target

2: The lens may not be optimized for the reproduction ratio required to reproduce the chart

3: The lighting on the chart can vary

4: The atmospheric conditions in the room varies - more or less dust for example

5: The exposure may vary from test to test

6: The film differs by emulsion number

7: The development varies due to the strength of the chemistry and its degree of oxidation

8: Developing times will vary.

9: Paper sensitivities will vary by emulsion number

10: Paper development will vary as will the exposure time

11: Your eyes will vary in how small a line pair they can resolve depending on how tired your eyes are

12: The loupe or magnifying system you use.

Lastly how many people buy lenses just for duplicating a flat chart?

No one if the use intended is for scenics, interiors, 3 dimensional subjects close up, etc.

The best way is to learn how to read an MTF chart, we can send you this or you can propably find it in Popular Photography, buy a lens based on these tests and try it. If you are happy then you mage the right choice. If you are not return it as unsatisfactory.

HP Marketing Corp. U.S. distributor for Amazon, Braun, Gepe, Giottos, GO Light, Heliopan, HP Combi Plan T, Kaiser fototechnik, KoPho cases, Linhof, Pro Release, Rimowa, Rodenstock,Sirostar 2000

From: kirkfry@msn.com
Newsgroups: rec.photo.equipment.large-format
Subject: Re: lens testing
Date: Mon, 14 Dec 1998

Sandy,

Bob might have also mentioned that almost every lens manufactured for large format cameras in the last 50 years or so, if it is not damaged, will resolve at the theoretical diffraction limit (about 50 lppm)at f22, a very frequently used f-stop for adequate depth of field. I did a bunch of lens tests and concluded after reading the negatives with a microscope that I didn't really learn much. I got as much variation with the same lens as I got between lenses, all pretty close the theoretical maximum. What amazed me was that some pretty old lenses are capable of turning in impressive numbers. The problem with Bob's MTF approach is that it is pretty hard to the get numbers on a 300mm f6.8 Dagor. There may be a reason for this since the goal after all is to sell you one of the super duper wonder lenses. Just remember that Ansel Adams and Edward Weston used lenses that today would be characterized as junk compared to modern lenses...... It's not the hardware.....

rec.photo.equipment.35mm
From: ttsalo@iki.fi (Tomi T. Salo)
[1] Re: tokina 17mm lens
Date: Fri Jan 15 1999

rpn1@cornell.edu (Neuman - Ruether) writes:

> It is not a fisheye, and for my purposes, it
> was also not very sharp... (ATX-AF...).

You mean your specimen was not very sharp ;-) (for your purposes, whatever these might be...) My specimen might also not be VERY SHARP in the extreme corners, but it certainly is sharp (considerably better than the accepted 0.03 (or 0.025) mm circle of confusion standard). Saturation, distortion and contrast are very good. Build is excellent. AF is pretty lame though.

--
ttsalo@iki.fi
Tomi T. Salo

From: "Donald D. Forsling" forsling@netins.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How is the Quality of Quantaray Lense?
Date: Thu, 17 Dec 1998

Jim Bisnett wrote

>I was stupid enough to buy two of the Quantaray lenses when I first   bought a
>camera 10 years ago. They suck. They suck. I can't believe that sigma made
>them. If they did, they still suck.
>
>If someone can tell me if sigma made them.. I would like to know. Since I have
>considered buying sigma before, and if they do make quantaray I won't even
>consider it.  

The fact that Sigma makes Quantaray lenses (generally conceded to be a pretty poor line of lenses) says nothing about the quality of the lenses Sigma sells under its own name. Quantaray lenses are made to a certain set of specs, obviously. Sigma lenses are made to another and somewhat higher and tighter set of specs. There's no reason to believe that if it wanted to sell in the high end, Sigma could manufacture lenses as good as Nikkors. Obviously they don't. Quantaray lenses aren't very good. Sigmas are better but are still not as good a Tamrons or Tokinas--and nowhere near as good as the average Nikkor. High quality does cost relatively big money. Quantarays are cheap, bad lenses. If you can afford something else, buy it.

Cheers,
--
Don Forsling forsling@netins.net

rec.photo.equipment.35mm
From: dave@ced.utah.edu (dave)
[1] Nikon QC problems?
Date: Mon Jan 18 1999

I recently purchased two new Nikon lenses both of which had problems which resulted in their return. The first was an 80-200 f2.8 zoom. This lens showed excessive pincushion distortion at all aperatures regardless of focusing distance. I tried a couple of friends lenses which showed a lesser amount when close focused and none at infinity. The second lens was a 20mm f2.8 which had a large number of particles on the inner lens elements. Light tapping cause the particles to move about. Am I just being too picky or has anyone else had any similar problems?

d-

Date: Wed, 20 Jan 1999
From: mbergma2@ix12.ix.netcom.com
Reply to: mbergma2@ix.netcom.com
To: Robert Monaghan rmonagha@hermes.seas.smu.edu
Newsgroups: rec.photo.equipment.35mm
Subject: Re: lens variations within batches Re: tokina 17mm lens

Robert,

Given the subjective nature of reading resolution charts it would of interest to see the results of multiple tests on the same lens.

I also noticed the same types of results as the Kowa lenses in old copies of Modern Photography Photo Buying Guides.

Marc

rec.photo.equipment.medium-format
From: lemon@lime.org (lemonade)
[1] Re: Evolution of MF to MF
Date: Thu Feb 25 1999

see.signature@bottom.com wrote:

> images (Provia 100 and Velvia) were soft.  The Leica and Contax/Zeiss
> were literally razor sharp.  Furthermore, the feel of the Nikon images
> was soft.  Color was good but resolution and definition were lacking.

Sounds like there might some have been maladjustment or miscalibration or sampling issue. There surely are differences in the optical signatures of the various lenses, but not to give this conclusion.

In terms of resolution and contrast alone, Pop Phot recently tested all the various 50mm f1.4s. Top (just barely) was the Contax; then Canon, Leica, Minolta, Nikon, Pentax, Schneider, in alphabetical order.

They noted there could easily be more within-brand variation than what they found between brands.

And this was resolution and contrast testing only.

--

rec.photo.equipment.medium-format
From: daverk@msn.com
[1] Re: Rollei TLR
Date: Mon Mar 22 1999

a.calciu@anent.com wrote:

> the xenotar is a lens superior in performance to the planar. as a  matter of
> fact, ALL zeiss glass has been outclassed by the schneider products in the
> past 40-50 years. if you look at current prices, the 80mm planar for rollei
> 6008i costs 1400 bucks and the 80mm xenotar is 4000. much better lens,  and it
> is the same formula as the one in the TLR rolleis.
>
> for some hard to explain reason, people seek the planar. i have both lenses
> in TLR and SLR and i can tell you that the xenotar gives more pleasing
> pictures to my eyes.

I've seen a fair amount of variation in Rolleiflex TLRs between various lenses of the same type, not to mention between Planars & Xenotars and Tessars & Xenars, so I don't think it's possible to accurately make a statement like the one above. I will say the best-performing 2.8 Rolleiflex I've ever used had a Xenotar lens, but the Planars have been no slouches either. My favorite 3.5 Rollei (the only one I currently own) has a Tessar but I've had or used some fine cameras equipped with Xenars. Overall I don't think there's much difference between the Zeiss and Schneider brands. IMO the lens-to-lens variation is more significant.

-Dave-

Date: Sat, 27 Mar 1999
From: Peters torx@nwrain.com
Subject: [Rollei] Soft coatings, Lens tests

I had a beater Steinheil lens for an Exakta and learned about those soft coatings: I was wiping down the lensbarrel with WD-40 on a cloth just to clean it, when I accidentally got some of it on the glass...POOF, no lens coating! Wow, just gone...

I think it's important to remember when testing lenses--as Marc said--that when you test only one lens against another, all you are really "proving" is how one particular specimen stacks up against another particular specimen. It's easy to forget that. It may be an indication, or may lead you to SUSPECT that the results may be the rule, but you really can't draw that conclusion. If you are considering selling one or the other, a "one against one" test can be a factor in deciding which to keep, or show any serious problems you weren't aware of. But also remember, for a good test, the playing field has to be level: I use the same roll of fine grain B&W; film (if possible) or same lot number, same lighting conditions, same subject, same development, etc. Otherwise, you introduce variables that may flaw your result.

I don't have the standard lens test resolution charts, but tape a page of the classifieds to a brick wall and set up on a tripod 10 feet or more away. I can't determine lines per millemeter, but I CAN see which is the smallest readable type, and I can see differences in contrast. It is better to do your own printing or use a microscope to look at the negatives, because you can't assume that a commercial printer will critically focus the enlarger for each print--or maybe for any of them!

And, I too was wondering where Bob Shell is...He's not usually this quiet! :-)

bob

Date: Fri, 2 Apr 1999
From: "Kotsinadelis, Peter (Peter)" peterk@lucent.com
Subject: RE: [Rollei] Coatings for TLR

Many lenses were hand assembled which may account for the variances if the sharpness in one to another. I have a Tessar from 1939 that is absolutely spectacular in sharpness and surprised the heck out of me. I have seen newer ones that were good to excellent.

Peter K

-----Original Message-----
From: D.O'Keeffe [mailto:lt01@dial.pipex.com]
Sent: Friday, April 02, 1999 12:54 PM
To: rollei@mejac.palo-alto.ca.us
Subject: [Rollei] Coatings for TLR

Dear all,

When were coated lenses first introduced for Rollei T TLR's? One of my 2 T's (which I have good reason to believe dates from c.1964) has single coated lenses but the other (c.1960, I think) is definitely uncoated. Both have Tessar 3.5's.

And curiously, the older uncoated camera certainly produces sharper pictures at apertures between f.3.5 and f.8. Why??

TIA,

Darren.

From: Anders Svensson Anders.-.Eivor.Svensson@swipnet.se
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Nikon Vs. Tamron
Date: Thu, 20 May 1999

Chuck Ross skrev:

> According to Photodo's tests, the Nikon 28-200 comes in at a 2.6 rating,
> and the measured focal length is 28-190mm.
>
> The Tamron 28-200 comes in at 2.7 rating, the measured focal length is
> 29-192mm.

I wonder what the standard deviation between samples are for each of these lenses?

As long as not a few samples are tested, and there is a statistically significant difference, such small differences may mean "about the same quality".

> --
> Chuck Ross
> http://www.enteract.com/~ckross
> Digital Photo Gallery

--
Anders Svensson
Anders.-.Eivor.Svensson@swipnet.se

From: Anders Svensson Anders.-.Eivor.Svensson@swipnet.se
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Nikon Vs. Tamron
Date: Fri, 21 May 1999

W Scott Elliot skrev:

> Anders Svensson wrote 
>
> >I wonder what the standard deviation between samples are for each of these
> lenses?
> >As long as not a few samples are tested, and there is a statistically
> significant
> >difference, such small differences may mean "about the same quality".
>
> Calculating standard deviation from a sample size of one would not be a  very
> useful exercise.

Sample (singularis), samples (pluralis)? Did I make a grammar mistake? - sorry for that, English isn't my native language.

Of course. Standard deviation from one sample is meaningless. I agree.

> These ratings can only be used as a rough indication of quality, not an
> absolute ranking.  It would be more useful to use a large random sample and
> show standard deviation, but that still wouldn't help.  How would you know
> where the individual lens you were thinking of buying fit on the bell  curve.
> It could be near the middle, but it might also be in the top or bottom
> decile.

OTOH, they could be helpful if it was explained what they could mean. A large variation could indicate several things like bad quality control, a lens construction that is hard to make consistently well, a zoom lens that is mechanically weak - things like that.

It could also be a memento to the buyer to check his individual sample and exchanging it until he/she got a decent one. Think about what the manufacturer would do if such findings were publicised - he might even spruce up the QC himself...

> My Canon 28-35 f3.4-4.5 ***(???, my remark)*** receives a higher 
> Photodo ranking than > the Canon 17-35 f/2.8 L that costs more 
> than twice as much.  Do I think it is a better lens?  Not likely.

No, but it does have a better ranking. :-)

I don't think you are wrong in your conclusion, but the tests are what us poor consumers has to rely on until we actually get the lens in our hands.

You are absolutely right that 3.5 might be a good grade on a hard to make lens (like a WA zoom), as well as a (extremely) low grade n a lens that is easy to make (like a 50 mm prime)

Anders

> Scott

-- Anders Svensson
Anders.-.Eivor.Svensson@swipnet.se

From Nikon MF List:
Date: Mon, 5 Apr 1999
From: Larry Kopitnik kopitnil@marketingcomm.com
Subject: Re: Nikon Noct Coma problem

I've heard of the coma problem you described with the Noct-Nikkor. Though not having used the lens, I'd not be the best one to offer comment on that portion of your post. But in regards to:

>Do Nikon lenses vary from unit to unit by much?

I'd answer, yes, they can. Some zoom lenses in particular can suffer from sample variation. For example, the first 24-120 I owned showed particuarly bad distortion. I sold it but later regretted that decision and bought another. On my current one, distortion is not nearly as bad. I now assume the first had a problem (a misaligned lens element, perhaps?).

With the Noct-Nikkor, the aspheric element is ground by hand rather than machine. So that lens may well be susceptible to sample variation.

Larry

[Ed. note: this highlights constant changes and shifts in lens models...]
From: Pete Stone ptstone@concentric.net
Newsgroups: rec.photo.equipment.35mm
Subject: Nikon AFS Tests
Date: 24 Apr 1999

I recently had the chance to test 3 samples of the Nikon 80-200 AFS and 1 sample of the 28-70 AFS and thought others might like to know what I found, along with some opinions about other lenses from competing manufacturers.

I've been shooting professionally for the last 20 or so years, and currently use 3 Nikon F5s, Canon EOS-1n & EOS 3, and 2 Leica R8s, along with a wide variety of optics for these systems, so I'm not romantically enamored with any particular brand. I just use what I feel is the best tool for a particular photographic situation.

My tests were done locked down on a Gitzo Pro-Studex Tripod on Provia ( RDP-II ) film looking across open water ( to minimize any thermal effects ) at some fine detail at infinity ( a grain shipping port actually ) . Two exposures were made at each aperture tested ( to minimize second frame effect unsharpness ), and the same fine field detail was placed in various areas of the frame ( center, edges, corners ).

The results with the Nikon 80-200 AFS ( all 3 samples tested similarly ) was that while the central region proved sharper at 180-200mm than the previous 2 ring model, the corner sharpness was visibly poorer, even stopped down to f 5.6 or so. The newer lens also had a higher level of light falloff into the corners at 2.8 & 4.0 than the older version, which could be a problem for anyone trying to shoot a tight vertical portrait on chrome film. Stopping down from 2.8 to 4.0 actually produced no difference in corner exposure, it just brought the central area exposure down approx. 0.5 stop ( Most 80-200 zooms actually t-stop out to around 3.2 or so effective aperture ). The results in the 80-180 range showed improved corner sharpness somewhat, but the level of light falloff into the corners was still high.

I would be interested to hear if any others have run tests comparing the old and new versions of these lenses. By way of comparison, the best performing medium zoom tele I've tested is the Leica APO Elmarit 70-180 2.8 ( performs like a prime at just about every focal length! ), and the Canon 70-200 EF 2.8 ( I use this lens frequently...excellent performance at 180-200mm, with just a slight drop of resolution into the corners, along with a lower level of light falloff than the Nikon AFS: in fact the level of falloff the Nikon has at f 4.0 just about equals the level of falloff the Canon has at f 2.8).

Sharpness wise, the 2 ring older Nikon 80-200 AF-D performance at 200mm at 2.8 is not as high as the Canon 70-200 2.8, or the 180 2.8 by my tests, but stopped down to 4.0 or so it gets better.

My test of the Nikon 28-70 AFS ( at 2.8, 5.6, 11 ) was done against my Canon 28-70 EF, and the results were also quite interesting. At 70mm and 50mm, both lenses produced very similar results, that is to say excellent overall sharpness with some slight softening at the edges at 2.8, and medium softening in the corners. At 35mm, the Nikons performance level was visibly worse than the Canon ( which looked very good into the corners esp. stopped down ) at just about all apertures tested ( it looked like curvature of field to me...since the corners could be refocused at a different point than the center ), which was a real surprise ( sample variation?...Ill try to borrow another sample to retest this focal length ). At 28mm, the Nikon performed better than the Canon due to the fact that some color fringing was visible at the edges of the image with the EF lens that was completely absent with the Nikon ( due possibly to the 2 ED elements incorporated ), although Id have to rate both lenses very good at this setting.

If you're thinking of buying the Nikon 28-70 AFS, definitely test it at 35mm to see if it meets your optical needs before committing youre cash.

A few quick comments about some other optics I own:

Best 20mm: Leica ( actually 19mm! ) pretty much sharp across the whole field, but with no provision for addl filters ( I had to make my own adapter )

Second best: Nikon 20 2.8 AF-D ( AIS=same optics ) sharp into the corners stopped down...nice and small, but needs a format cut hood!...and 77 or 72mm filter size. Worst: Canon 20mm 2.8 EF. Possibly Canons worst wide...edge performance poorer than the Nikon, and the corners arent sharp even down to f 11! Canon users: get the Nikon 20mm and use it on your EOS with an adapter ( stopped down only )

Best 24mm: Almost a tie between the Canon and Nikon 24mm 2.8s...both very good across the field. Both need a different filter size ( 72mm? ) and format cut hoods

Best 28mm: Close tie between Leica ( 28 2.8R ) and Nikons ( 28 2.8 AIS...NOT AF! ), both excellent right across the field.

Second best 28mm: Nikon 28mm AF-D ( the D version is a new design from the AF non D version, although it does not have CRC, as some report, only the AIS 2.8 does! ) The main optical difference between the AF-D and AIS versions, is the better performance of the latter in the corner areas ( dont know about close performance! ). Third best 28mm: Canon EF 28mm 2.8...EVERY sample ( 5 to date ) Ive tried on 3 different bodies has a soft area at the left side of the frame just slightly away from the extreme edge...go figure! Otherwise its a pretty good general performer...but not excellent. 35,50,85,100, etc...All 3 ( Nikon, Canon, Leica ) lines optics are very good to excellent, with differences in performance ( bokeh generally is better with Canon & Leica optics ) being less significant than at the other ranges mentioned. A couple of real standouts, however, are the Leica 100mm 2.8 APO macro ( superb quality even wide open ) and the Canon 135 2.0 EF ( excellent by 2.8).

Nikon/Canon wide zooms: ( they both have visible color fringing at the wider settings, and neither one is better than a comparable prime at the same aperture, despite what others say ...except maybe the Canon at 20mm ( the primes that poor! )...I can just about always tell on film if Im using a wide zoom or a good prime....but they ARE great for grab shots, travel work, and anytime its just not convienient to keep switching lenses. I keep one on a body, but use a prime if I have the time! )Canons 17-35: best range is 20-28mm...17mm has poorer edge/corner performance, 35mm has lower sharpness overall. Nikons 20-35 has good central sharpness across the range, but at 20mm it has poor corner sharpness ( the Canon zooms better here, due to its secondary aspherical element ), and color fringing quite visible at the edges...this improves as you zoom tighter...its performance at 35mm is better overall than the Canon at 35mm .

That's it for now.

Date: Sat, 31 Jul 1999
To: rmonagha@post.smu.edu
Newsgroups: rec.photo.technique.nature,rec.photo.misc
Subject: Lens Variations - or why magazine tests vary so much etc.

Do lenses vary, or are they all identical as the industry seems to want us to believe?

Obviously, lenses do vary, the only question is by how much and in which directions (standard deviation and skewness) from the desired central point.

Manufacturers also vary, depending on their cost and quality concerns, from Zeiss/Leitz at one end to perhaps the much maligned Kiev soviet era lenses (which have high levels of performance variability reported for them).

An elementary observation is that objective lens tests/measurements by various magazines and presumably experienced observers reveals a surprising amount of variation in ratings for the same lens types (but different individual lenses were tested) when tested by different observers. While items like weight may vary not at all between reviews, items like focal length and maximum aperture and other optical parameters often vary significantly between testing sites.

A more sophisticated analysis suggests that there is no clear pattern in these variations, so magazine XYZ isn't always more conservative than magazine ABC.

In other words, if you assume the testing procedures are performed reasonably consistently, you either have to accept my premise that lenses vary considerably or explain the large and rather random variability observed in the reported test results between observers and scientific tests.

The only scientific study which I have seen published of consumer 35mm lenses shows they vary by about a grade either way (e.g. +/- 1 on a 5 scale). Granted, this study is of older 1970s lenses from the same batch. But lots of us still have and use such lenses. Current technology of making lenses may or may not be better in any given case or model, but the optical results suggest that newer isn't always better. (see http://www.smu.edu/~rmonagha/third/variations.html for details).

There is a lot of anecdotal evidence that lenses vary significantly. For example, Roger Hicks (The Lens Book) reports on finding some lemons when trying to replace some lost Vivitar zoom lenses - some worked remarkably, while others were bad performers. Again, this observation of lens variability may help explain the many arguments from folks who used this or that lens, some finding theirs worked great while others were upset at such junk glass. Based on such professionals' observations, lens variability is clearly an issue, and lenses of the same model are not all alike or interchangeable, as the photo-industry might like us to believe.

Thanks to Heinz Richter, I can document some significant differences in tolerances (a source of variability in lenses, obviously), from my page at http://www.smu.edu/~rmonagha/mf/glassmfg.html on optical glass mfgering...

quote: Here is what I have been able to gather: The industry standard for camera equipment is tolerances to 1/1000". Canon, with the introduction of the F1n proudly claimed that they had increased tolerances to 1/1250". Leitz and Zeiss (when they still made cameras themselves) apply tolerances of 1/2500". The international standard for glass requires tolerances of +/- 0.001% for the refractive index and +/- 0.8% for the Abbe number (dispersion). The Leitz data are +/- 0.0002% for refractive index and +/- 0.2% for the Abbe number. Allowable tolerances during lens production for Leitz are 1/4 Lambda or 1/4 of the average wavelength of light, which amounts to 0.0002 mm. Minoltas tolerances in this regard are 0.0003 mm. Needless to say, the Leitz testing equipment, much of which is manufactured in house, is made to specifications exceeding the figures above. end-quote:

Despite some posters belief that today's lenses are somehow better built and tested than in the past, there is lots of anecdotal evidence otherwise.

Just today, I've gathered postings on a defective (pinhole) coating, a bubble in a modern glass lens element (not older or high Z glass), and a scratch on the internal surface of another (nikkor) lens. I have a page which has just some of the many posts in the rec.photo groups on such lens defects in brand new lenses as these at my lens faults page - see http://www.smu.edu/~rmonagha/bronfaults.html

Personally, I suspect that lens variability is probably greater now than ever, partly due to the lower price points of modern lenses (in constant dollar terms) esp. in the popular consumer models, where enlargements are rarely made, so tradeoffs in quality are less obvious against cost factors.

I also think that new technologies such as autofocus may result in other design constraints (speed of focusing, low mass of components, low friction) which make it harder to design high performance lenses. Given that AF lenses are now cheaper than MF lenses (which lack the motors etc.), I have to wonder if the tolerances are always as good or better?

I can also argue that the tendency towards using zoom lenses (let alone AF zooms) would seem to make lens variability more likely, given the larger number of elements and potential for variability that implies. So I argue that the average lens of today is more complex (# elements, zooms, wide angle zooms more prevalent etc.) and harder and more costly to produce to a given tolerance or level of variability, versus older simpler prime and more limited range zooms of the past. So I expect higher variability today, rather than less, as the industry would like us to believe.

If you get what you pay for, then just how much precision do you get in a 14 element 28-210mm AF zoom that sells for $169 on the street? Hmmm?

The highest quality lenses are not made by robots or automated lines, but by human interaction, as Zeiss and Leitz are quick to remind us. As the lens defects suggest, most current lenses are only batch tested (1-2% of lenses typically) and you are probably the first human to examine your lens or test it after you purchase it. Better go and shine a light thru those new lenses, you may discover some defects you just assumed couldn't be there - others have, as their unhappy posts have indicated ;-)

More anecdotal evidence from camera store owners/staff suggest that obvious lens faults and problems result in return rates from 1-2% (prime OEM lenses) to over 40% (low end imports, possibly a bad batch?). Clearly, many unsophisticated photo newbies and users who don't carefully test their lenses may be experiencing less than great lens performance without knowing it in their 4x6 prints. Are you one of them? For tips on testing, see my page at http://www.smu.edu/~rmonagha/broncameratest.html tips...

However, I agree that we don't have good evidence of what is really going on with variability among consumer photographic lenses. I think the reason is probably obvious, i.e., the photoindustry and the magazines have an interest in the myth that lens are essentially the same. If lens really do vary, then the magazine reviews aren't very useful, since they don't apply to the lens you have in your hand, which could be significantly worse (or perhaps better) than the one and only one lens tested by the magazines. The photoindustry doesn't want us insisting on "cherry-picking" or testing all the lenses in our local camera store to get the best one, leaving them with the returns and lemons right?

Naturally, the manufacturers could simply provide more info on the range of variability of their lenses, tolerances, and so on. I suggest that they don't do so because they would only lose by doing so and alerting us as consumers to the considerable variability and lack of quality control and testing of individual lenses in their current production processes (i.e., that's where the lenses with defects cited above got thru despite obvious visible defects because only a few lenses in each batch were actually tested). If bubbles, internal lens element scratches, coating faults, and so on are getting out of the factories, what makes you think their processes are so precisely controlled otherwise as to turn out otherwise identical and precisely aligned and figured lenses with minimal variability? Personally, as a systems engineer, I take these coating defects, bubbles in glass, and scratched elements as pretty convincing proof of processes that are "sub-optimal" or even out of control? ;-)

It is fairly easy for previous posters to dismiss my claims that lenses vary, and vary significantly enough that you really have to test each lens (new as well as used) and that the published magazine guides are not as useful as they seem precisely because the lens specimens being tested may vary significantly from the one you are buying.

What would it take for a magazine to test a stratified random sample of a dozen or so typical consumer zoom lens and see how much variability there is in a particular sample lens? Why has only one magazine (Modern Photography) done so in the last 3 decades? On a non-major advertiser, I might add? I think the answer to that question is pretty clear too ;-)

Perhaps Chris Perez or others can arrange to perform this simple experiment. If lenses really are as variable as the available scientific and anecdotal lens testing and defect rate data suggest, then don't we really need to change our approach to testing lenses - and buying them?

From Leica LUG Digest:
Date: Mon, 6 Sep 1999
From: "Frank Filippone" red735i@earthlink.net
Subject: [Leica] Leica, Process control, and Minolta bodies

..... No they do not cost the same, but
>not because of MORE  rejects, but rather because there are stricter limits

It is NOT an erroneous statement, because this is a reality. Leica rejected so many CL cameras, that Minolta gave up doing them. Zeiss rejected so many of their own Contax 80-200 f/4 lenses (which was a wonderful lens that could focus right down to the front element) that it cost 8-10 times the competition at the time. High rejection rates are well-known.

I think you missed the point....it is really a volume thing..... the rejects are caused by a too wide variation in individual piece part manufacturing tolerances to specification. These tolerances build up and eventually get out the finished goods out of spec. Manufacturing process control for the past 15-20 years states that you never allow the final product to get out of spec, you catch the errors or tolerances at the piece part point. These tolerances at the piece part point must be tightened ( which costs more on a per unit basis, and really costs more if you are doing very small runs plus the R+D costs to do this when amortized over a small volume) ) such that the final assembly is within spec. When this is done, you tighten the specs again. The idea is to never throw out a single part along this learning curve, you just make them more perfectly each time you make a run of parts. Continuous process improvement. The final result is that the final assembly has reached extremely low tolerances, therefore low reject rates.... but it takes volume and focus. I do not think that Zeiss or Yashica/Zeiss rejected the completed lens. They are familiar with manufacturing techniques and especially process control. They never made enough lenses (enough iterations) to figure out the problems and resolve the solutions.

A rule of thumb is that your manufacturing costs should go down at the rate of 25% per doubling of volume, as expressed in units per constant time period.

The example of Leica/Minolta and the CL is real good.. the big house could never spend the effort ( resources ) to locate and get in control the manufacturing tolerances... too small a production in an otherwise enormous factory..... it was doomed to fail. The same may be true of Zeiss and Contax. An interesting idea would be to take a MInolta body and change ONLY the lensmount and necessary prongs to accomodate the Leica R lenses. Given the S series of lenses this may not be as difficult as it first looks......Or, (Oh, the flames I may receive for this!) change the mechanical properties of the Leica Lenses to fit on a Minolta (Canon, Nikon, or other ) body. Keep the optics just like they are today. Now there is an idea! Leica optics on Nikon ( or other) body reliability!

Final note.... While most of the process control improvement theorems were based from the USA, the experts in the implementation are Japanese. They do understand the issues, but can not break the inevitible low volume issue with Leica et al.

Frank FIlippone

[Ed. note: sample variations between ratings services etc...]
rec.photo.equipment.35mm
From: Jakob Bisgaard jfb@dhi.dk
[1] Re: How to weight photodo in decision making
Date: Wed Sep 29 1999

Mark Stringer wrote:

> A lot of reference is made to photodo ratings.  Some seem to weight these
> ratings high and others pooh-pooh their use in their decision to buy a  lens.
>
> What weight do you assign to these ratings relative to construction and
> feel?
>
> Has anyone purchased a lense with a high photodo rating and then been
> disappointed in its actual performance?
>
> Does anyone ever disect the ratings on a zoom lense to get a modified
> rating? (i.e., take a Pentax 80-320 with a 2.5 rating and rate its 80-200
> performance as 3.1)

Weight it as ColorFoto, Practical Photography or Chasseur d'Image. They are using MTF-based test as well. According to Photodo site the Nikon AF 180 f/2.8N IF-ED recieves a 3.6, which is nothing exceptional. In fact it's a bit low considering it is a prime with ED glass. Before Photodo initiated their site I came across some MTF data on the Canon mailing list for the Nikon 180 mm. You can find a reference on them on Olle Bjernulf's site

The explanation for the test is given on http://www.geocities.com/Yosemite/8917/lenstest.html

and to me that sounds pretty identical to the explanation you can find on Photodo's site under "Understanding the MTF graphs, numbers and grades"

Nikon AF 180 f/2.8N MTF 20 lines/mm - Canon mailing list

        Center    9mm      18mm
f/2.8    75/77    71/77    54/70
f/8      83/82    86/79    80/70

Nikon AF 180 f/2.8N MTF 20 lines/mm - Photodo (read from their graph)

        Center    9mm      18mm
f/2.8    70/69    67/66    62/62
f/8      81/80    84/73    82/58

If we should take these numbers for granted, then we could expect 10% deviations, maybe due to sample variation, maybe due instrument variation the same rating as the 180, namely 4.4 out of 5. On Photodo's site the AF-I 300 f/2.8D is given a 4.2 compared to the 3.6 for the 180. Draw your own conclusion.

By the way the data for the AF-I 300 f/2.8D on Photodo's site is wrong:

The AF-I 300 f/2.8 has 11 elements in 9 groups (4 ED elements) weighs 2950 g has 2.5 m as minimum focus distance is 241 mm long and 124 mm in diameter and the image of the lens is wrong. The data and image fits with the AF-N 300 f/2.8 with the screwdriver focus mechanism. I wonder which lens they actually tested!

And by the way my Nikkor AF 180 f/2.8N performs excellent. No regrets from me there. Wouldn't trade it for anything. :-)

Cheers and good shooting
Jakob

--
Jakob Bisgaard

Danish Hydraulic Institute (DHI)

[Ed. note: Mr. Carroll makes a good point that ratings by users often vary, including info about changes in lenses within models (slipstreaming) referenced above in the main article]

rec.photo.equipment.35mm
From: myname ccarroll05@sprynet.com
[1] Re: How to weight photodo in decision making
Date: Thu Oct 07 1999

Artur Swietanowski swietanowski@bigfoot.com wrote:

> Only if you have nothing but this sample. The poster assumed
> (quite reasonably, too) small variance.

Actually, I'm not so sure how reasonable this assumption is. It is somewhat at "variance" with my own experience. You can see some info regarding sample variability of various Nikon lenses at, for example, David Reuther's site. Some models seem to be better than others. But even with thousand dollar items there is evidence from his testing that you could get mis-alignment right out of the box on two of three samples (I'm thinking here of the 70-180 Micro). I myself have an excellent sample of the aforementioned zoom Micro, but, unfortunately, have gotten a doggy example of an otherwise widely praised 35-70/2.8, Go figure.

From Nikon Mailing List:
From: "Fernando Martins" fer@caleida.pt
Date: Fri, 31 Dec 1999

this may be due to the lens too. when I bought a voigtlaender 19-35, I brought my F100 to the store and tried the lens. the fit was too tight, the 2nd lens was the same. the 3rd lens' fit was perfect. it is not my camera, I've mounted many many lenses in it, from 5 different brands, aging from the '70s until now, and all fit perfectly.

best regards,

Fernando Martins
NT Systems Administrator
fer@novabase.pt
http://fernando.somewhere.net/

Date: Mon, 27 Mar 2000
From: "David Glos" david.glos@uc.edu
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Mamiya 645 AF vs Contax 645 Lens Tests

T Thurston wrote

>Well, I've just read the review in Pop Photo (Apr 2000) of the Mamiya 645 AF
>camera and lenses.  I was particularly interested in what they reported
>regarding the lenses, since they just ran (Nov 1999) a similar review of the
>Contax 645 and lenses.
>
>They way I interpret their results, it looks like the Mamiya lenses might be
>somewhat sharper overall than the Contax lenses.  For example, for the 80mm
>lenses they report:

ACTUAL TEST RESULTS DELETED

>For the 210mm lens, Mamiya looks like a clear winner until you stop all the way
>down to f/22 (which I don't do too often with a long lens)
>
>Of course, many people are suspicious of tests by Pop Photo and such, but it is
>interesting to me that the Pop Photo results for Contax lenses match quite
>closely the results posted on PhotoDo, and the MTF figures reported for Contax
>645 lenses on the CARL ZEISS web site.  Thus, I wonder if perhaps these Pop
>Photo tests might be accurate, at least as regards sharpness.
>
>Any thoughts?

Sure. Having some experience in the world of peer reviewed scientific publishing, I can safely state that the results, as shown, have very little meaning. Give me a dozen examples, of each lens, randomly chosen from the production line, on a given day, and then lets perform the tests again. I would not be surprised that the standard deviations among the examples, of a single lens design, from a single manufacturer, exceed the noted differences between the single examples from Zeiss and Mamiya, as tested by Pop Photo. I'm not saying their info is bogus, just that there really isn't enough info to choose one or another camera based on the published test reports. I would be much more alarmed if they reported huge differences, or discovered a particular optical flaw in one of the designs.

Considering the level of refinement currently practiced in modern lens design and manufacuring, the price of these optics, and their chosen audience, I would be extremely surprised if Zeiss or Mamiya dropped the ball. As alluded to in another post, the real test would be in shooting several test rolls under the conditions under which you intend to use the camera. Particular handling quirks, as opposed to optical qualities, might just play a larger role than you might imagine in your final decision.

Respectfully,
David Glos

[Ed. note: slipstreaming...]
From Rollei Mailing List:
Date: Fri, 17 Sep 1999
From: Bob Shell bob@bobshell.com
Subject: Re: [Rollei] OT - Cannon EOS 3?

> The reason I was asking about using lenses other than Canon is that I would
> probably have used it
> with my Tamron 28-200, which I love, but I don't know how the '3' would
> react with the lense.  I
> always feel that my Tamron is more sluggish than the Canons, and some of
> the 'hiccups' I have had
> have been to do with my Tamron rather than the camera itself.

Tamron does not have access to ultrasonic motors and has to use traditional small electric motors as a result. This makes their lenses slower and noisier than Canon and Sigma. I'm using the Tamron 28-300 on the EOS-3 and am quite happy with the combination for many occasions. I just sold one of the shots taken with this combo to a publisher who is using it as a full page bleed in an 8 1/2 X 11 book, so you know it is plenty sharp.

> Bob - you said you were now using a mid-production one, I am assuming that
> the ones in the store are
> further along, or are they still developing the '3'?

Canon, as with most major manufacturers tends to refine products as production progresses. There were some reports of metering problems with early production ones, and I know Canon jumped on that right away and straightened out the problem, which was apparently a software problem and not a hardware problem. It should be far enough along in production now that any major problems would have been cured.

Bob

[Ed. note: good example of why you can't rely on single point lens tests]
From: tut@ishi (Bill Tuthill)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Value of photodo ratings?
Date: 7 Jun 1999

It's always interesting to watch people criticize free information. The Easy Lens Guide and Photodo are two of the finest network resources for photographers, and I'm happy that they are available.

Billy R (willreed@worldnet.att.net) wrote:

> Can you elaborate on these multiple conflicting test results and provide
> evidence that Photodo's ratings are incorrect?

Here are some instances where the Easy Lens Guide and Photo scores diverge by more than half a point. In some cases, one wonders whether they even tested the same lens!

focal length & speed wgt len cost close filtr EZavg Photodo

Sigma   macro 50mm f2.8 EX 1:1  358g   63mm   $236  19cm   �52  4.66  3.9
Minolta 300mm f4.0 APO G        1350g  220mm  $900  250cm  intg 4.09  3.4
Minolta 50mm f1.4 fixed         235g   40mm   $200  45cm   �49  3.73  4.4
Tokina  28-70mm f2.6-2.8 ATX    760g   110mm  $500  70cm   �77  3.63  3.1
Pentax  macro 50mm f2.8 FA 1:1  385g   70mm   $350  19�cm  �52  3.62  4.6
Tamron  28-105mm f2.8 LD        869g   112mm  $799  49cm   �82  3.32  2.4
Canon   100-300mm f5.6 L USM    695g   167mm  $600  140cm  �58  3.05  3.6
Pentax  80-320mm f4.5-5.6 FA    550g   129mm  $250  150cm  �58  3.01  2.5
Canon   75-300mm f4-5.6 USM     495g   122mm  $215  150cm  �58  2.58  3.1
Tamron  28-80mm f3.5-5.6        237g   71mm   $119  70cm   �58  2.38  3.0
Pentax  28-80mm f3.5-5.6        275g   78mm   $140  50cm   �58  2.33  3.3
Tamron  28-300mm f3.5-6.3 LD    570g   94mm   $499  82cm   �72  2.00  2.6

From Contax Mailing List:
Date: Thu, 20 Apr 2000
From: "Bob Shell" bob@bobshell.com
Subject: Re: [CONTAX] important info

With final QC in Wetzlar and later in Solms. Too many lenses would not pass QC, according to Leica, so the cooperation was dissolved.

I'd guess the QC on the 80-200 from Kyocera is also done in Solms, but my bet is that not many lenses are rejected.

I may have told this story before, but don't recall. I was in Solms leaning over the shoulder of a Leica technician who was unpacking zoom lenses from a shipment from Minolta and checking centration on an automated machine. Each lens was checked and then put in either of two bins, one bin to be sold and one to be sent back to Minolta. I had heard that they rejected 75% of the lenses Minolta shipped them, so I asked him, via a translator, if this was true. He thought a quick moment and then shot back, "not from every batch!"

I also watched shutter modules for the R series being unpacked from a Seiko shipment and tested, and similarly divided into two batches.

Bob

- ----------

>From: "wei zhang" milklover2@hotmail.com
>To: contax@photo.cis.to
>Subject: Re: [CONTAX] important info
>Date: Thu, Apr 20, 2000, 5:04 PM
>

> Some of Leica zoom lenses were built by minolta as well.

[Ed. note: more slipstreaming and versions variations...]
From Contax Mailing List:
Date: Sun, 07 May 2000
From: "Bob Shell" bob@bobshell.com
Subject: Re: [CONTAX] Decisions decisions....

- ----------

>From: "Mark F Dalal" mfdalal@inreach.com
>Subject: Re: [CONTAX] Decisions decisions....
>Date: Sun, May 7, 2000, 5:51 PM
>
>
> One of the fellows on the Pentax List had that Tamron lens. He recently
> switched to the the SMC-A* 300/2.8. When comparing the two, he stated that
> the Tamron was a good and sharp lens but it lacked the snap/contrast of the
> Pentax lens. I would expect that you would experience the same thing
> comparing it to the Zeiss equivalent.

Depends on which version of the Tamron. It has been revised several times and the current version has a LOT more contrast than the first version. We just re-tested the latest version, as well as the latest 2X to go with it, and found both greatly improved over their predecessors.

> I think I would second Bob Walkden's recommendation. Perhaps a Pentax
> A*300/2.8 (~$3000 from KEH) plus a Super Program (~$200) would give you some
> really excellent results. The Pentax A*300/4 is also an excellent lens and
> goes for ~$600.  I don't much about Nikon or Canon offerings because I have
> little interest in either system but I imagine that their offerings are more
> economical also.

If you are going to recommend something other than Zeiss, the only reasonable suggestion if the Canon 300mm F/4 IS. Not only is it optically superior to the other offerings in this focal length and aperture but it offers IS technology which makes a real practical difference.

Bob

[Ed. note: RE: slip-streamed changes/improvements...]
From Rollei Mailing List:
Date: Thu, 11 May 2000
From: Bob Shell bob@bobshell.com
Subject: Re: [Rollei] 2000F Vs 3003 again

Rollei always had a practice of revising internal parts and subassemblies as problems were found. Kept repairmen busy keeping track of all the upgrades and installing them. I've been told that late 2000F cameras were substantially improved internally over early ones, and if early ones had come in for overhaul they would have been automatically upgraded at no additional charge. How can you tell? You can't.

Bob

....

From Contax Mailing List:
Date: Sun, 28 May 2000
From: "Bob Shell" bob@bobshell.com
Subject: Re: [CONTAX] Long lenses in MM/AE mount?

Also, don't forget sample variation, which is generally greater with lower priced lenses. After all one way to keep prices down is to set broader tolerances.

A good example is in 300mm f/2.8 lenses. I've used Tamron, Sigma and Tokina. The Tamron I tested was super, the Tokina not so hot and the Sigma really awful. George Lepp tested the same trio, but different samples, and his sample Tokina beat the socks off his sample Tamron. His Sigma sample was a close third, much better than mine.

The only way to buy one of these lenses is to work with a dealer who will let you shoot some film and return the lens if you are not happy.

BTW, we recently tested the latest version of the Sigma and it is far better than the old one.

Bob
- ----------

>From: Shel Belinkoff belinkoff@earthlink.net
>To: contax@photo.cis.to
>Subject: Re: [CONTAX] Long lenses in MM/AE mount?
>Date: Sun, May 28, 2000, 10:40 AM
>
>I heard some very good things about certain Tokina lenses, and a couple
>of months ago I had a chance to use one.  IMO, it was a POS - light fall
>of wide open and up to two stops down was horrible.   Unacceptable for
>my needs for that particular  lens.  Yet the lens owner was very
>pleased with it.

From Leica (topica) list:
Date: Mon, 29 May 2000
From: "David W. Almy" dalmy@mindspring.com
Subject: Sample variations

All,

There have been several mentions over the years of high-performing and marginal-performing lens types. By that I mean variations in optical performance by the same lenses, lens-to-lens -- the "all 35/1.4 ASPHs are NOT created equal" argument.

If you ask Leica NJ, they will tell you that all lenses that leave the factory "meet minimum performance standards." Joy. Others, users particularly, will spend years on the hunt for lenses that exceed these "minimum standards," and then jealously guard their precious, high-performing samples. I am the fortunate owner of an Apo-Televid spotting scope purchased from just such a (laudable) fanatic. He tested six before buying the one I now own, rigorously testing the optical performance of each before deciding. I, alas, do not have the time or temperament for such homework.

Is Leica optical performance, lens-to-lens within type, noticeably variable? If so, is there a scientific source anyone can identify where my lenses can be individually checked for performance? My 35/1.4 ASPH, when sent to Leica NJ for just such a check, came back "meeting minimum standards" and with glass covered with fingerprints (true story), so my suspicions remain.

David W. Almy
Annapolis

From Leica (topica) list:
Date: Mon, 29 May 2000
From: Godfrey DiGiorgi ramarren@bayarea.net
Subject: Re: Sample variations

David Almy wrote:

>Is Leica optical performance, lens-to-lens within type, noticeably
>variable?

Certainly. But then so is everyone else's. When I was at JPL, we'd get 10 or 15 Nikkor lenses and put them each on our optical bench for testing, pick the two out of the batch that were the best.

Such testing is mostly ephemeral for lenses used for pictorial photography.

Godfrey

From Leica (topica) list:
Date: Mon, 29 May 2000
From: Eric Welch ewelch@flashcom.net
Subject: Re: Sample variations

David Almy at dalmy@mindspring.com wrote:

> Is Leica optical performance, lens-to-lens within type, noticeably
> variable?

Only by using them for most of us so it seems. I remember listening to National Geographic photographer Nathan Benn talk about finding just the right 50 Summicron M and said he went through several (all used) looking for one that was what he was looking for. At the time I doubted him, but back in the 60s, the vintage he was interested in, the variability of the lenses is no doubt more than of today - or so one would think.

--

Eric Welch
Carlsbad, CA

..one sees the glass half full, another, the glass half empty. The engineer sees the glass twice as big as it has to be.

From Leica (topica) list:
Date: Mon, 29 May 2000
From: Godfrey DiGiorgi ramarren@bayarea.net
Subject: Re: Sample variations

Joe Stephenson wrote:

>the optical bench settles the problem at one level, but leaves it
>wide open at another.

Absolutely. That's why I said, "Such testing is mostly ephemeral for lenses used for pictorial photography."

>I'm curious to know the magnitude of the differences between lenses of the
>same type. Has anyone got a quantitive measure to offer?

At the Jet Propulsion Lab, we were grading lenses based upon resolution, contrast and color fidelity. Our interest was in data collection: we needed lenses of high resolution and accuracy to supply research data needs. Bokeh and pictorial qualities were not considered.

(Please remember that I'm citing this from memory of having done these tests in 1985-1986; I don't vouch for 100% accuracy in my recollection.)

Given a typical sample of 15 lenses, the results of testing demonstrated variation between samples of around 8% at 3 sigma, 6% at 2 sigma and 3% at 1 sigma. 1 sigma means that 85% of the lenses in the sample differed by up to 3% on the combined parameters of resolution, contrast and color fidelity. We cranked up the criteria sufficiently to select two or three lenses out of that sized batch, which meant that they performed the same to about a 1% variation level.

To give a simple example measure: if you considered 100 lpmm to be the peak, 3% variance gave us as low as 97 lpmm, 8% variance gave us 92 lpmm.

In practical terms, that means that of the 15 lenses in the sample, it was probable that 12-13 of them differed by up to 3% from the peak performance, with the remaining 2-3 of them differing by up to 8%. 3% difference was not noticeable to the naked eye, 8% difference became apparent when we viewed photographs taken with them back to back under fairly rigorous circumstances.

All of the lenses met very high quality standards for pictorial photography and produced more than acceptable photographs. We were looking for the peak performers to achieve accuracy in data gathering.

Godfrey

From Pentax Mailing List:
From: "Roy L. Jacobs" rljacobs@pipeline.com
Date: Fri, 2 Jun 2000
Subject: 67 Lens Quality

This is a more complicated area than I first believed, since based on my tests and feedback from others, it appears that 67 lenses vary a lot from sample to sample. For instance I once had the 200mm (new), which I purchased new. It did not have good image quality, which I did not realize until much later. (I t was sharp wide open and got worse as you stopped it down.) I got rid of it. Recently I purchased a used 200mm (new) and it has very good image quality. (It is fairly sharp wide open and gets better as you stop down, with fine detail very well rendered. I have the 75 mm and, while I do not use it alot it seems ok. The 55mm (new) and the 105mm are great lenses, although I have read criticisms of the 105. The 165mm f. 2.8 is o.k. and the 400EDIF is a great lens. I recently tested the 45mm and was not impressed, but the test was run on chrome and I am not convinced the film development was acceptable.

Bottom line -- you must test each lens you purchase as I am convinced there are large sample variations. (On ebay I obtain in advance a 14 day return privilege or I will not bid.) If anyone is interested in my testing procedure, email me privately.

[Ed.note: Mr. Puts is a noted Leica lens and optical testing expert...]
From Leica Topica Mailing List:
Date: Tue, 30 May 2000
From: "Erwin Puts" imxputs@knoware.nl
Subject: Sample variations of lenses

Any lens, however rigorously manufactured, tested and checked, will show some variation in quality. Today ,and in the past also, you compute a lens with the exact specifications that you want. If your design program or manual computation tells you that the thickness of a lens should be 4.87367 mm (yes five digits) and the radius of curvature 179,27691mm, that is what you use and the subsequent analysis of image quality (spot diagrams, point-spread functions and what have you) do calculate with these figures. In the next stage you do a tolerance analysis to state real world manufacturing and product variations and feed these quantities into the calculation to note the maximum variation that can be expected, including the amount of image degradation to be expected. A comparison of these values with the 'ideal' values will give you an indication. It is now common practice to use the MTF values as the test figures. Leica uses the contrast value, measured at 8 locations (45 degrees steps) around the lens at a location 15mm from the center. In this way they can test for absolute performance and check decentring and other tolerances. They then specify a percentage that the readings may vary around the 'ideal'value.

The tolerance analysis has specified the maximum allowable deviation to ensure that the specified image quality is within limits. To give an example. You can design a lens and specify that it will deliver (ideally) a contrast of 60%. Tolerance analysis does show that the variation is from 50% to 65%. Now a contrast difference of 10% is hardly visible at the spatial frequency used for the test. Leica might specify that a lens should have a minimum value of 55% to pass the test and accept that a larger part of the production lenses will have to be rejected, adding to the cost of the lens, which will be reflected in the end-user price as we all know.

The testing procedure will be outlined in more detail in my book. We may conclude that every lens, even from Leica has a certain spread around the ideal value and of course only a few will be spot on at all 8 measurements.

So Leica lenses do show variations in measurements for optical quality. Do we see it in practise? Yes, we can. But only under very exacting circumstances. Again a true life example: the lens performance is of couse calculated for the ideal image plane of maximum sharpness. If you defocus by 0.2 mm the contrast will drop already visibly. So the performance chain is: tolerance of the lens, tolerance of the focusing mechanism, tolerance of the distance from film plane to lens flange, tolerance of register. The focusing tolerance is the bad guy here and add to this our physiological limits to focus accurately and we have to accept that the body and rangefinder tolerances have a larger impact on image quality than the lens variations in itself. Still if I have a lens at the minimum level and a lens that happens to be perfect, I will notice a slight contrast drop in the field or on axis, IF all other variables are under control and IF I can compare side by side. Every Leica lens is individually checked and will perform to the design specificaions, within a small margin. Humans are humans and even machines have quirks, so an occasional out of tolerance deviation will occur, even at Leica.

Leica USA will tell you that the lens is within limits, if the specimen is above the minimum level.You might have the nagging doubt that the minimum is too low below the norm and that there must be a 'better'lens somewhere around there in the outside world. And that this better lens gives visually better results. It will on the bench and it will be marginally visible in practise. But an overexposure of 1/3 stop and a focussing error of 5 cm at a 1 meter distance will be more devastating to image quality. I do not believe the stories that someone will test six or more lenses (or binoculars) and handpick the best. This story assumes two premisses: that it is possible to find a lens where all parameters are in the plus direction and that any sample of six lenses will contain one of these 'best' lenses. It is true that if you test six randomly picked lenses, you will find a few that have higher values on the bench, if you are able to use MTF equipment.

If you do resolution comparison you need to add at first a 10% variablity because of the psychology involved. A resolution measurement of 100lp/mm can in fact be anything from 90 to 110, without the observer knowing it or seeing it. So you select a lens of allegedly 110lp and assume that is is a better one than the others with 100 or 95lp. In reality all may be the same or even the 95 may be better than the 110.

I have checked by MTF and other testing equipment literally hundreds of Leica lenses and I am very reluctant to make strong statements here. In real life you will find lenses that will score differently on different parameters. If I find a lens that has on axis contrast of 65 and in the field 56, it is within tolerances. But is it better than a lens which has 57% all over the picture area and is also within tolerances but on the low side. I would prefer the last sample. Of course we all hope to have the lens that has 65% all over, but if this lens has a focus shift of 0.09mm, where as the other one has 0.05mm (again both within tolerances) which one to choose.

I would prefer to hone my technique, use the optimum combination of material, and try to find my personal optimum with the equipment I have. If I consistently can get this optimum very time I take pictures, then the study of the optical properties of the lens are in order. And forget about doing your own benchmarking equipment. It takes a study of weeks and an experience of years to make reliable observations and draw relevent conclusions. There is much more here than meets the eye. If you happen to have a lemon lens, you will surely note it.

Erwin

From Leica Topica Mailing List:
Date: Tue, 30 May 2000
From: Eric Welch ewelch@flashcom.net
Subject: Re: Sample variations of lenses

Erwin Puts at imxputs@knoware.nl wrote:

> I do not believe the stories that someone will test six or more lenses (or
> binoculars) and handpick the best.

That isn't what I said. What Nathan Benn said was that he tested five or six lenses before he found one that met his standards, though I'm sure he could have tested 10 or 20 if that had been the case. He was just lucky to stop at six. All I am reporting is what he said. It's obviously not proof. His complaint was that Leica manufactured a lot of sub-standard 50 Summicrons according to what he looked for in performance. Maybe he was unlucky to hit a bad sample of lenses, and in other cases he would have found it in the first or second sample. The law of averages is no guarantee that every sample of two or six or 20 will fit the bell curve.

And the proof is in the pudding. His photographs reflect the quality of his standards. He also shoots with Hasselbads, and that shows in his photos as well. The larger negative is a factor, lens performance is another. In the end, it's the photographs that count, and image quality is one factor.

Bench tests are fine and good, but really only instructive if a statistically significant sample of lenses is large enough of say 60s vintage 50 Summicrons vs Nikkors or Canon or Contax competitors. Nobody has done that, except maybe Leica and you. A significant sample would be I'm guessing 400. That's about the minimum for any population to get reliable numbers. Anybody you know of actually do that besides Leica?

Even then, only a generalization can be made about the average performance of Leica lenses vs. Nikons, Canons and Contax. No single lens will be guaranteed to be good or a dog.

So in the end, a person has to test a lens and see if it fits their needs and standards. If not, send it back. But I suspect most people, if they stopped worrying about the niggling little differences that can only be seen in a bench test, they'd be out shooting more and making great photos (practice makes perfect as they saty) rather than shots of targets and brick walls, newspapers or whatever else they might use for testing.

And they'd end up being a whole lot happier. For someone like you, Erwin, who finds joy in your approach, then this is a significant area of study worth of the best efforts.

The two will never meet in the middle. Either one cares about it, or not. But if efforts are put into one area to the maximum, then the other is not likely to get adequate attention. I think that's just fine and good. But some people can't seem to pick their side. I picked mine long ago. I'm intersted in photograhs, and sometimes I sell my Leicas just to remind myself of that. That's as foolish as obsessing over whether one 35 Summilux ASPH outperforms another - only to find out one can't see the difference in practice, and has nothing to do with making great photographs at the end of the day.

--
Eric Welch
Carlsbad, CA

From Leica Topica Mailing List:
Date: Tue, 30 May 2000
From: Godfrey DiGiorgi ramarren@bayarea.net
Subject: Re: Sample variations of lenses

Yes, Erwin. You are absolutely right: there is much more here than meets the eye.

However, I'll assure you that the testing procedure we used at JPL was able to grade batches of 15 lenses quite effectively, and consistently, for the purposes we had in mind. We could always narrow the performance spectrum adequately to find the two best performers out of 15 that met *our* criteria for quality, perhaps not everyone else's.

I never try to do formal lens testing on my own. It's a waste of time. But that was not the situation at the Jet Propulsion Laboratory's facilities where we had every kind of optical bench in the world to work with as well as a few dozen optical engineers.

Me? I'm must a simple mathematician turned software geek, I did the data acquisition and statistics for the tests, collated the information that the setup produced. Each lens took about two hours time to do the data acquisition, plus setup and calibration time.

Godfrey

From Leica Mailing List;
Date: Sat, 1 Jul 2000
From: "Giorgio Ferrari" fergio@galactica.it
Subject: Leica vs. Nikon and Leica glass

Gentlemen,

I've read your messages about the Leica-Nikon comparison and about the Leica glasses. All right: Leica lenses are on average better than Nikon lenses (except for 35/1.4, 50/1,4, 105/2.5 and 180/2.8) but the quality of Leica lenses is not so constant inside of the producton.

One of the best italian Leica expert, Mr. Gino Ferzetti from Pescara - 83 years this year -, has demonstrated that the global quality of Leica lenses is variable, with high differences from the best to the worse. During the years, for each lens Mr. Ferzetti has tested three or plus models, verifying, substantially, three levels of quality.

I'm not sure about why: I think that the mechanical quality standard and the optical quality standard can be different during the years and for different markets. For example you can think about a 50 mm with an optical lenght of 50.3 mm and a mechanical lenght of 49.8 mm, and so on.

So, I'm sure that Leica lenses are better than Nikon lenses, but a worse Leica lens is not so diffrent from a best Nikon lens.

Excuse me for the confusion and for my bad language.

Giorgio Ferrari
"fergio@galactica.it"

From Nikon MF Mailing List:
Date: Tue, 11 Jul 2000
From: "MMA" aasem@ni.net
Subject: Lens review questions

Bjorn Rorslett commented on the Nikkor AF-S 17-35mm f/2.8D at

http://www.foto.no/nikon/lens_zoom.html

There, he noted sample variations that suffered from "optical decentering" and "focus shift." I have written to him for more information but have not recieved a response. Can anyone on the list explain what these terms mean?

How can I test my lens to see if it suffers from these problems?

Thanks,

-mark

[Ed. note: special thanks to Jerry for sharing these observations on "cherry picking" an ideal lens, and on lens variability...]
Date: Sun, 12 Nov 2000
From: Gerald Crum gwcrum@apk.net
To: rmonagha@post.cis.smu.edu
Subject: Lens Quality

I currently use an inexpensive, Tamron 70-210 f/3.8-4 zoom. An excellent lens and the third and best of the 3.8-4 series that Tamron made a few years back. The bottom line here is that it was "cherry-picked" from about 6 of these.

The design is excellent, but the lens is a cheap one and as a result suffers from A LOT of variation. How good is it? How about macro results comparable to a Tamron 90 f/2.5 macro lens? Believe it! How good at normal distances? Comparable to a Nikkor 200 f/4 AIS.

How much did it cost? Allowing for several trade ups from previous models to this one, about $160. The first one cost me $100 and it was selected from 2 the dealer had. Then I found a place that had a couple of used ones and traded up to the best of the the three for $20. We went through this process twice more at $20 per trade. After a while it became a sort of game. How far can I take this process?

I haven't found a better one now in over 4 years, so I probably have maxed the process. But it very nicely illustrates the issue of variability from unit to unit. It probably also illustrates the issue that the designers do a pretty good job of getting the design center values right. Which then get screwed up by production variables and the vagaries of statistics. This lens is only an f/4 and has an objective of 58 mm with 12 elements, making the manufacturing a lot easier than an 80-200 with a 77 mm objective and 16 or 17 elements.

If you start looking at the tolerance stack up problems and statistical variations in such a lens, the surprise is not that there are variations, but that the lenses are as consistent as they are. One of the biggest arguments for prime lenses in my viewpoint is the usually smaller number of elements and the much simpler internal mechanism.

This has benefit not only in tolerance stack up and statistical variation, but also in less flare, higher contrast, and less internal loss. T numbers are not usually reported, but the more complex zooms typically lose from 1/2 to 2/3 stop. TTL metering takes care of the variation, but you paid for an f/2.8 not an f/3.6. Further, if you're using a hand held meter, you have to take this into account. But that's another e mail isn't it.

Regards,

Jerry

Date: Fri, 01 Dec 2000
From: ozetechnology watsondk@my-deja.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Lens Test Reports Valid?

Thankfully I have been able to self test several lens (and have not actually ended up with all of the tested lens) from the pro store that I buy most of my gear from, on the understanding "I break it, I own all the pieces" which is fair.

I am trying to find a good Nikon 17-35AFS so far tested 2 and both failed the tests for different reasons, what looked like manufacturing problems. Will try again in a while, and hope that Nikon has sorted this problem out. No recient posts on this NG with problems, so never know.

You may pay slightly more by going to a pro store rather than the mail order/heavy discounters, but then again are these places going to let you try before buy?

Heard of some horror stories, about discount stores here that do not accept returns at all, you get a problem its to the importer you have to go, which from personal pain is not plesent.

--
site: http://www.ozetechnology.com

pburian@aol.com (PBurian) wrote:

> I'm still not sure what anyone -- including Consumer Reports -- can do
> about this variance.
>
> Test a dozen versions of the same lens, before you decide as to which one to
> buy?
>
> And again: How are people supposed to test several samples before buying one of
> them? Some stores DO allow you to return a lens within x days but many do not.
>
> I think B&H; said (here) that they do for example.
>
> Some charge heavy "restocking" fees, plus you pay all the shipping costs both
> ways.
>
> I suppose the answer is to deal only with stores with a great exchange Policy
> and learn how to test lenses yourself. Whether most people can do the latter is
> questionable.
>
> Peter Burian

Date: Fri, 01 Dec 2000
From: Tony Polson tony.polson@btinternet.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Shutterbug -- Fluff??

w6uv@hotmail.com (Jerry Gardner) wrote:

> PBurian said:
> >Not a single piece of equipment ever comes to Shutterbug. Contributors
> >who do test reports contact the distributor when they want to test a
> >camera or lens.
>
> Unfortunately, this probably results in you testing nothing but cherry-picked
> pieces of equipment. Any review will be much more objective and valuable if
> you buy the equipment retail from someone who does not know you will be
> writing a review.

All this proves is that testing a single sample, whether by an experienced reviewer or by an individual consumer, is not statistically reliable. As a professional engineer I have been involved in the acquisition and use of statistics for longer than I'd really like to say (25 years), and never have I had to make a judgement based on a single piece of data. For the statistics to work a number of samples are required, and the greater that number, the greater the confidence in the final result or recommendation.

There are no photo magazines in the world that routinely test a number of samples of the same item. The way Peter Burian works is the way *every* reviewer works; they get one sample to try.

For an example of what happens when more than one sample is tested, I suggest you should look at the reviews of the Nikon 17-35mm f/2.8 D AF-S on Bj�rn R�rslett's excellent web site at:

http://www.foto.no/nikon/index2_PC.html

Click on "Reviews" in the side bar, then "17-35 mm AFS Nikkor". You should read the whole review, but the following extract is of particular interest:

"The 17-35 mm f/2.8 AFS Nikkor is a remarkable and amazing lens, that well deserves a total ranking of 5 (Excellent). I'm looking forward to enjoyous photo opportunities with this lens and shall post more pictures taken with it when they become available.

"Since this review was incepted, I have had shooting and testing experience with quite a number (more than 10) of 17-35 Nikkors both on my D1 and F5 cameras. The reasons for this are the persistent rumours that this lens is inferior to the 20-35 Nikkor. When I first got my personal sample of the 17-35 , it didn't take me long to detect this sample lens showed severe faults of focus shifts and decentered elements. Since the first two or three of these lenses I had used were superb, I wouldn't accept no less for my own use. So, I demanded a packing case filled with 17-35 lenses for testing from the national Nikon dealer, and surprise, got it! Thorough testing of all these 17-35 Nikkors indicated that mild decentering isn't uncommon, and the same goes for focus shifts. A sample size of 10 is to small to draw any statistical significant conclusions, but finding that 2 out of 10 lenses had severe decentering problems was discouraging. Eventually I located a perfect 17-35 AFS and kept it for my personal use. It has proven itself a highly useful favourite lens on my D1. In fact, my records show I use it each and every day. Must be a favourite, then."

It's impossible to say whether Bj�rn R�rslett's findings are representative of this Nikon lens, the Nikon range or 35mm SLR lenses in general. But they do point to a variability that, ideally, should be considered in a review.

Unfortunately, we do not live in an ideal world. It is unlikely that we will *ever* see manufacturers routinely submitting ten or more examples of lenses for magazine reviews.

One example of a lens that seems to have a large variability is the Cosina 19-35mm lens that's also available as a Soligor or Vivitar, and was available as a Tokina a couple of years ago: The current Tokina 19-35mm and recent 20-35mm are and were very different lenses. The Cosina 19-35 seems to vary from excellent to abysmal depending which person's view you accept. This is a sure sign of variability to an extent that it makes any review of a single example of the lens both unrepresentative and unreliable.

--
Tony Polson, North Yorkshire, UK

Date: Thu, 30 Nov 2000
From: "Wayne D" wdewitt@snip.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Lens Test Reports Valid?

There is variation in any process. Statistical sampling allows for the extrapolation which infers the overall quality of the entire population, nothing more. Today, more than ever, units which would fail an inspection are shipped due to relying on sampling instead of 100% final testing. The last time I looked at Sony's data on their televisions they predicted the shipping of 1%-2% of product which would fail final testing (if tested), an "allowable" percentage considering that consumers don't know what they are looking at to begin with. The process capabilities, and materials in existence, are the best that we've ever had available. The real question is:

"What is the manufacturer's incentive?". If the consumer is uneducated standards can be relaxed until returns approach an economically "uncomfortable' level - then action will be taken. Anybody out there driving on Firestone tires?

....

Date: Thu, 30 Nov 2000
From: OneThumb OneThumb@nospam.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Lens Test Reports Valid?

PBurian wrote:

> <<
> If anything, my suspicion remains that today's lenses are worse in terms
> of tolerance variation than in the past.>>>
>
> Robert: I'm not sure what anyone -- including Consumer Reports -- can do about
> this. Test a dozen versions of the same lens?
>
> Peter Burian

Dozens of lenses which include parts and material that come from changing suppliers and processes and designs that are continually "tweaked" during production. Getting a good sense of product variation probably isn't practical for anyone but the manufacturer. Actually, that's not true. Public forums like this are the best way for us to get a handle on product variation.

Date: Sun, 10 Dec 2000
From: Tony Polson tony.polson@btinternet.com
Newsgroups: uk.rec.photo.misc
Subject: Re: Lens reviews

"CJ" jarvisc@mcmail.com wrote:

> MTF or not MTF that is the question? And I agree whole heartedly that I
> don't want to start a long thread on that sublect. Have you found any  other
> general sites related to lens quality and testing methods?

There's no shortage of such sites. But there is a grave shortage of sites with reliable information. The sites that do exist either offer data culled from testing a single sample of a lens, or are a collection of subjective reviews.

In the first case, you need to make some significant assumptions about variability between samples before you trust the published data. In the second case, you need to understand the very human desire to justify purchasing decisions before you trust the published data.

--
Tony Polson, North Yorkshire, UK

Date: Sat, 13 Jan 2001
From: ee553@freenet.carleton.ca (Harry Liston)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: lens variations Re: Lens Sharpness Review? (Lens Qualities...)

I have a similar theory. I believe that the assembly of lenses varys considerably, even if the components of the lens are within tolerances. Therefore a "tweaked" lens will drastically outperform a consumer lens put together on the assembly line. But hey, the prices on most of these lenses are so low compared to twenty years ago that I wouldn't dare complain. There are some terrific deals in optics these days.

Harry Liston

Date: Sat, 13 Jan 2001
From: bc1959@my-deja.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: lens variations Re: Lens Sharpness Review? (Lens Qualities...)

The main causes of lens-to-lens performance variability are tilt and decentration of the lens elements, not incorrectly polished surfaces, so by and large your theory about lens assembly is correct. There are techniques for lens assembly which eliminate these errors to an arbitrary degree, but its normally not justified in photographic optics. Projection testing or MTF testing across several different image radii will reveal these problems pretty well, but of course it takes a lot of work. I agree with Bob and everyone else who thinks that differences of 0.1 on a scale of 5 are insignificant. Perhaps their only sin is to provide area-weighted average MTF figures with too much precision.

Brian

...

From Rollei Mailing List:
Date: Fri, 26 Jan 2001
From: Martin Jangowski martin@jangowski.de
Subject: [Rollei] Lens quality control at Rollei

Hi!

I just talked to an old Rollei technican, and he told me about quality control at Rollei after WW-II up to about 1965. He said, that Rollei made a test exposure on a glass plate (to insure absolute stability) with every taking lens they got from Schneider or Zeiss and checked the lens resolution with a microscope. They had batches with >50% rejection rates both from Zeiss and Schneider... I wonder, how many duds they were getting after 1965.

Martin

Martin Jangowski E-Mail: Martin@Jangowski.de

Date: Sun, 28 Jan 2001
From: "Daniel H Lauring" dlauring@hotmail.com
Newsgroups: rec.photo.equipment.35mm
Subject: Re: The Cheapest Nikon 400mm?

I agree with you on Photodo's limitation. The biggest issue I have with it is based purely on statistics. The lens to lens variance can, in my experience, be greater than the lens model to lens model variance. Yet, Photodo test just one lens with one basic test (which may or may not be provided by the manufacturer.)

For example, Photodo rates the Canon 100-400 IS a 3.6 while the 300 IS prime gets a 3.4 and the 300 non-IS gets a 4.3. While I'll agree that the non-IS has an advantage over the IS because it has less elements I strongly doubt that your average 100-400 zoom will beat your average 300 prime.

There are enough other cases like this to cloud the reliability of the Photodo numbers.

Of course, that doesn't even begin to address all the other important lens characteristics like resistance to flare and focus speed and reliability.

Photodo is a good resource and I'm glad we have it but when people use it as end-all bible of lenses to proclaim which lens is superior, I cringe.

Danny

Tony Polson tony.polson@btinternet.com wrote

> "Matt Clara" emjaysea@worldnet.att.net wrote:
>
> > I guess I am putting my faith (perhaps unduly) into the reviews found here
> >http://www.photographyreview.com/reviews/35mm_zoom_lenses/product_6040.asp
> > At any rate, it will be awhile before I have the money to spend on the lens.
> > Hopefully there will be some more indepth reviews of it by then.  I think
> > Peter Burian is reviewing it this week in Utah?
>
> As we've discussed on here before, this site is not bad, yet it includes
> reviews by individuals justifying their own buying decisions and by
> others who want to get back at a manufacturer or dealer by unfairly
> criticising the product.  We have discussed to death (and beyond) on
> here the strengths and limitations of the Photodo site, which has at
> least attempted to attain objectivity, but has (in my opinion) largely
> failed.  We've also discussed test bench reviews in magazines and
> reviews by an individual such as Peter Burian which don't include bench
> tests.
>
> The only certainty arising from these discussions is that no single
> review can ever give a definitive answer.  It's best either to get as
> much information as you can, then decide, or to buy the lens from a
> dealer that does returns and test, test, test it until you're sure it's
> up to the standard you need.  Or do both.
>
> The downside risk of buying a lens you don't like is now much reduced
> thanks to eBay.  Instead of losing most of a lens' value when you trade
> it for another, you can now get most of your money back on eBay.
>
> --
> Tony Polson, North Yorkshire, UK

[Ed. note: Mr. Stephen Gandy is a noted camera expert and author of some great web pages at Camera Quest on classic cameras!]
From Rangefinder Mailing List:
Date: Tue, 23 Jan 2001
From: Stephen Gandy Stephen@CameraQuest.com
Subject: Re: [RF List] Lens Tests Reliability

It's not unusual to find people who got very different results than you from the same type of lens. How is this possible if published lens tests are accurate?

The basic problem with lens tests reliability, besides the quality of the test procedures, is that lenses DO vary in performance right off the assembly line, and they can vary even more by the time they are delivered to you right off the shelf.

Don't fool yourself by putting too much faith in lens test results.

These differences show up when different lens testers test the same lenses, but rate them differently when compared to each other. Lens testing procedures do vary. While its difficult if not impossible to directly compare two different lens tests of the same type of lens if different test methods are used, the info does become useful when both testers test the same family of lenses, and rate them differently compared to other lenses within that family. These differences are more likely to show the lens variations of a lens lineup, than bad testing techniques.

If you want truly accurate lens tests of your lens, you must test it yourself, and not trust published lens tests to give you your answers. Your lens can be a dog or a masterpiece, and your friend with the same lens can have the same or opposite results. The higher the priced lens, the better the quality control, the less likely this is to happen, but it still happens.

Besides manufacturing tolerances, failures in assembly, mistakes in assembly by new people, failures in quality control, all of which affect the lenses coming off the assembly line, what happened to your lens AFTER it left the factory ?? How many times was the crate dropped on its way to the camera store ? How many hours was it vibrated during shipment ? Did the previous customer in the store drop it on the carpet an hour before you bought it ? Lens tests themselves have another set of problems.

Why do magazines and lens testers do lens tests, knowing there are problems with the test reliability ? To sell you magazines and books. Why do lens manufacturers publish lens tests and encourage testing ? To sell you lenses.

OK so the average lens test results over many sources will give you the general quality of a lens, it will still not give you the quality of YOUR lens, which may or may not have been dropped numerous times in the crate by a teenager on drugs trying to learn how to use the loader, or who knows what else. To be sure, test your own lenses.

Stephen Gandy

From Rollei Mailing List;
Date: Mon, 19 Feb 2001
From: Eric Goldstein egoldstein@usa.net
Subject: [Rollei] Re: OT Schneider Radionar question

J Patric Dahl�n wrote:

> I compared the Rolleicord 3,5 Zeiss Triotar (uncoated) with the coated 3,5
> Novar on a 6x6 Super Ikonta. The Triotar was sharper and had better contrast
> than the Novar. I think the results of the Novar only looks good because of
> the larger negatives.
>
>
> I once had a Nettar 6x9 camera with an uncoated 4,5 Tessar and it wasn't
> better than the Novar. Maybe a little better at the edges, but the over all
> sharpness was too soft. It didn't improve with stopping down. :-/

The problem with comparing the Rolleicord Triotar to the Super Ikonta Novar and the Nettar Tessar f/4.5 is that there are too many variables from which to draw a reasonable conclusion. First, it is apples and oranges to compare front cell focusers to full focusing optics. Second, older cameras and optics (particularly folders) typically suffer from other ills which preclude sharpness, and it sounds like your Nettar had it's share of these difficulties...

I'd also just note that before the 70's, lens production and fitting was not automated and much hand grinding, matching and fitting was done. Production tolerances were much less stringent, and there was very significant individual variation from lens to lens. This makes it difficult to draw conclusions with regard to each of the lens brands design and executions as a whole from just a few isolated examples.

Eric Goldstein

From Rollei Mailing List;
Date: Sun, 25 Feb 2001
From: "R. Peters" torx@nwrain.net
Subject: [Rollei] Variation in lens samples

Some have pointed out that individual lenses of one type (tessars for instance) can vary in sharpness. This is true, but I think one thing to remember in cameras that are 50 years old is that they have probably been serviced in their lifetime. If an unqualified person has "repaired" them, the sharpness can be lost. (And, that may not be the worst that happens in such a shop.) Some of these cameras with "less sharp" lenses may only need a proper repair to have performance restored.

I once took a TLR to a guy who told me how he was going to correct an infinity focus problem, and he was going about it absolutely backwards. Had he done what he originally proposed, it wouldn't have been focused right at ANY distance. That's one reason I look for a specialist in whatever type camera I need to have repaired. It may not cost any more to have it done right the first time that it does to have it screwed up, then fixed a second time by someone who knows what he's doing.

...

bob

From Rollei Mailing List;
Date: Sun, 25 Feb 2001
From: Richard Knoppow dickburk@ix.netcom.com
Subject: Re: [Rollei] Which camera for Heidi

you wrote:

>ShadCat11@aol.com wrote:
>
>> My Y'mats ranged from good to poor, but even Rollei lenses vary.
>
>
>All lenses of this vintages varied between samples, often dramatically. 
Hand
>grinding, fitting and matching was the order of the day, and tolerances 
were
>much less stringent than today. It was reported in the photographic press 
of
>the day and well into the 60's...
>
>
>Eric Goldstein

What hand grinding? Lenses were made on machines many at a time. Cemented surfaces were lapped to match individually but no hand polishing was done.

Centering is partly hand work. Centering of two accuracies were done. Air spaced lenses were often centered by simply clamping between two tubes and grinding the edges. Cemented elements were centered by sticking them on the end of a tubular tool and rotating slowly while watching the reflections though a telescope. The lens was pushed around on the tube until the images stood still. Then the element was clamped as above, and edge ground so that the edge was concentric and coaxial with the optical axis.

The main variation in old lenses is probably in glass characteristics. The designer assumes a certain value of index and dispersion (or v value) in making the design. If either varies much the corrections go to hell. One of the most important improvements in modern lenses is the improvement in reliability and consistency of glass.

For those interested and willing to do some searching the manufacturing techniques for lenses are covered in a couple of books published in the late thirties and late forties by Adam Hilger & Co. I don't have the specific titles.

Also, the manufacturing process at Zeiss Jena is covered very completely by post war reports by American and Brittish intellegence teams. I have some citations, but would have to really dig for others. These reports are hard to find. AFAIK the only complete (?) set is at the Library of Congress and you need to find a knowlegible librarian to help find them. These teams investigated all German industry. There is some pretty interesting stuff there if you can stand explaining what you want about fifty times. I have a little of the Zeiss information, and much more on various sound recording apparatus that I gathered years ago, in practcally another life, for a microphone history I wrote for the Audio Engineering Society.

----
Richard Knoppow
Los Angeles,Ca.
dickburk@ix.netcom.com

From Rollei Mailing List;
Date: Wed, 28 Feb 2001
From: Richard Knoppow dickburk@ix.netcom.com
Subject: RE: [Rollei] Classic Lenses and Focal Length Variation

you wrote:

>> If there are such differences in FL (focal length) between lenses (even  in
>> the same batch),
>
>My understanding is they do vary.  Also, the aperture they state is not
>always exact either...
>
>The focal length can be 'adjusted', to some degree, by how the lense is
>assembled.  I believe the TLR can be calibrated so that both lenses focus
>the same too.
>
>From the repair manual for the 2.8F it looks like they solid mount the
>taking lense to the carrier, and adjust the viewing lense, which appears  to
>be threaded, to match the taking lense.  Just my guess...

The focal lengths of the two lenses must match very closely, if not you will be in focus at only one distance. I don't know how Rollei handled this. They might simply have depended on the normal variation in focal length of the two types of lenses, or, I would think more likely, resorted to something like the use a graduated spacers to make small adjustments to the FL of the viewing lenses to match them. Such small variations would not upset the corrections of the lens enough to be important in this application. This is the same principle as front element focusing. It does not take much change in spacing between the front and center emement of a Triplet to cause a fairly large change in FL.

AFAIK, all Rollei TLR cameras fix the position of the taking lens and provide a threaded mounting for the viewing lens for adjustment. There is a method of locking it at the right point. The procedure for setting up a Rollei is to adjust the infinity stop, which is actually part of the focusing knob, so that the rearmost movement stops at infinity focus. This is done with either a ground glass at the film plane and a collimator target, or with an autocollimator. Once done the viewing lens is adjusted to match. Actually, the viewing lens can be adjusted at any distance, but its most critical at infinity. It can be checked at other distances, but, unless the lenses have been swapped somehow, it should match at all distances.

I don't know what method Rollei used at the factory to match the lenses, however, its not too difficult to do.

The variations in FL of production lenses is not great but can be significant. It mostly from differences in glass constants.

----
Richard Knoppow
Los Angeles,Ca.
dickburk@ix.netcom.com

From Rollei Mailing List;
Date: Wed, 28 Feb 2001
From: Eric Goldstein egoldstein@usa.net
Subject: Re: [Rollei] Classic Lenses and Focal Length Variation

Richard Knoppow wrote:

> The variations in FL of production lenses is not great but can be
> significant. It mostly from differences in glass constants.

Whether the typical FL tolerance of the day, +/- 5%, is "great" is a matter of opinion. By modern standards (less than +/- 1% for primes) they were monstrous. Two instances have just been raised where these differences were quite significant: in matching viewing and taking lenses in TLRs, and in matching lenses for stereo cameras. This necessitated F&H; and others to resort to additional manual manufacturing steps, which to my knowledge was handled as Bob Shell described: manual measurement of lens FL en masse, then hand pairing of matching lenses.

The differences in FLs and other tolerances were the cumulative result of primitive automation and numerous manual operations in fabrication and fitting. Glass manufacturing was also largely a manual operation, and there, too, production tolerances were wide by modern standards. Differences were significant enough for experienced shooters to test drive individual lenses and cameras to select choice examples, a procedure not generally in practice today.

Eric Goldstein

From Leica Mailing List:
Date: Sun, 8 Apr 2001
From: "Erwin Puts" imxputs@ision.nl
Subject: [Leica] Fw: lens design philosophies (1)

Optical theory and the correction of aberrations is no secret and wellknown at every design department. Most departments even use the same program! A typical design sequence consists of fixing the parameters (aperture, physical dimensions), then creating a design (number of lens elements etc), then a rough calculation and assessment of the basic aberrations, and then the big part: the optimization for a certain predefined merit function (to be translated as image quality or lens performance in practice). Cosina does this as does Leica or Konica or Contax, to list the major RF contenders of now.

What then are the guiding differences between Leica and Cosina, the lens lines of both I have tested fully.

At the start I may mention that no one can do magical tricks and a lens costing USD 2500 should be different from a lens costing USD 1000, even if we allow for economies of scale. Leica lenses are designed for optimum performance at the wider apertures and this choice has implications for all the rest.

The amount of aberrations is several magnitudes higher at wide apertures than at moderate ones. To control these errors one needs to use special glass, and to adhere to very small tolerances of manufacture and mounting, and to use material than can stand all kinds of temperature changes. It is clear that the reduction of optical errors to the micron level implies using manufacturing techniques to the same level of accuracy. As example one may note that a slight amount of decentring will show itself with bigger lens diameters and wider apertures and will go unnoticed in small-diamter glass at small apertures. And fluctuations in surface roughness will cause unwanted reflections which will be seen when the glass is very well coated and extremely transparant, but not so visible when the glass itself is more opaque.

This state of affairs does indicate that the material costs, manufacturing costs and labour costs are important parts of the equation when designing a lens. The design process is also a factor. Now we need to be somewhat high minded.

A lens consists of let us say 6 elements. We have generally 7 diffferent aberrations (spherical aberration, coma etc). Some of these lens elements are more responsible for any one of these aberrations. When correcting a lens one could concentrate per lens element on the related aberration or one can spread the aberration content over all elements. The normal optimization program throws all aberrations and all elements in one basket and uses mathematical techniques (Damped Least Squares) to find a minimum for the sum. Here one reaches a solution quite quickly. Of course one has to assign a weight to every aberration and lens element for this approach to work and here the influence of the desiger or team becomes paramount. Presumably this is the way Cosina works. (most designers do it this way).

Leica has another and more laborious method: they spread the aberration content as anyone does, but in a different way and they optimize per single aberration and do a precise balancing of aberrations to fine tune the residual aberrations. How they do it, is a secret of course, but it takes much time.

The results that can be reached have to fit the possible level of manufacturing accuracy and tolerancing.

An aspherical lens can be pressed in a short time or grinded in a day of work to get at the level of precision required.

End of part 1

Erwin

From Leica Mailing List:
Date: Sun, 8 Apr 2001
From: Erwin Puts imxputs@ision.nl
Subject: [Leica] design 2

It is very sensible for a designer not to try to optimize a lens beyond the level that the mechanical tolerances can allow for.

Some people have objected to my remarks that many Cosina lenses decentred elements as it has been interpreted as Cosina-bashing. It is not. Some older Leica lenses show this decentring too: it is the consequence of the tolerance level that is accepted or defined as the level of precision that is required for a given level of image quality or a given level of manufacturing precision. A designer and the company in which (s)he works knows what is possible or required and engineering is not sorcery: any additional minute of work costs money and more checks and frequent adjustments cost money too.

Generally Cosina lenses are very good and surpass many of the older Leica designs. That is the power of modern computer programs and improved manufacturing technology.

An intriguing question is this: do Cosina lenses show significant differences when stopped down to Leica lenses stopped down to the same aperture.

My simple answer is yes.

The advantages of Leica lenses are its tighter tolerances, use of better quality glass, advanced optimization and this shows at all apertures in higher micro contrast, higher transparency (finer reproduction in high lights and shadow areas), lower level of residuals (finer differentiation of hues and tones in small details), greater fidelity of reproduction etc. All of this over the entire (or most of it) picture area.

There is a diminishing return of course and at f/11 it would be quite difficult to see the difference.

Here one has to insert a few caveats.

If the photographer is not sensitive to subtle differences and does not do his own careful comparisons, these aspects may escape him/her. Listen to an orchestra conducted by two different directors playing the same piece of music!. Some will here the differences immediately and some will not. A trained ear is needed. So is a trained eye.

If the photographer's technique is sloppy or careless or if the material used is not up to the demands of the lens, many differences will be washed away by the generally very high noise level of the imaging chain. My view is this: the Cosina lenses deliver amazingly good image quality and many users will be served very well by this range of lenses, which expands at a remarkable speed.

Generally they offer better imagery than many Leica lenses of the previous generation. Any one who raves about his Summicron DR should try the Nokton 1.5/50 and get an insight in optical progress made possible by current design programs and optical theory.

Current Leica designs have a clear edge and you can appreciate that if you are willing to invest some time in the study of their characteristics: It is easy to gulp down a glass of superb wine and not note the difference in taste when compared to a cheaper wine. As any wine lover will tell you: take your time and chew!

From Rollei Mailing List;
Date: Fri, 27 Apr 2001
From: Richard Urmonas rurmonas@senet.com.au
Subject: Re: [Rollei] Introduction & Question

> Also, while the difference between Planar and Xenotar is
> largely philosophical, there appears to be a bigger difference between  Tessar
> and Xenar (with the Tessar the clear _winner_).

I have shot both Xenar and Tessar lensed TLRs "side by side". I would say the difference between them is no more than sample to sample variations (i.e. Tessar vs Tessar or Xenar vs. Xenar). You may have a particularly good example of a Tessar and so think that Xenars are inferior. Also given the age of these lenses, lens condition is a significant variable.

Richard

From Rollei Mailing List:
Date: Thu, 26 Apr 2001
From: Robert Meier robertmeier@usjet.net
Subject: Re: [Rollei] XA and 35S

There is a lot of variation in the Olympus lenses on the XA as well. I have had three XA's and the first one I got, new in 1980, had the best lens of the three. I have B&W; prints that are 15x23 full-frame prints from that camera that are sharp and stand up well next to any other 35mm camera's prints, including the Rollei 35S. But the second and third XA's I have acquired, are not as sharp and I have stopped using them. The XA-4, with its 28mm f3.5 lens is VERY sharp. Both XA and XA-4 do have noticeable vignetting, which is unfortunate and can be a problem with a lot of pictures (but not with a lot of others).

[Ed. note: even top cost and quality optics can have glitches...]
From Leica User Group Mailing List;
Date: Fri, 18 May 2001
From: Tom.Henson@bakerbotts.com
Subject: RE: [Leica] Re: quality control

Jim,

I have to agree with Seth. While it is indeed true that today's technology makes reporting these problems to a very wide audience almost instantaneous, my experience with Leica Solms has not been the best.

I purchased three new Leica lenses in the last two months: a 28-70mm vario (current version), a 80-200mm vario (current version) and a 100mm 2.8 APO (current version).

I had to send the 28-70mm back three times to get a decent lens; 80-200mm twice, 100mm 2.8 APO three times (never got a good one).

In the end all three lenses were replaced by Leica USA. The 28-70mm and the 80-200mm were hand selected by their quality control people and turned out to be good enough to keep. The 100mm APO was sent to Solms to have the serial number transferred off the bad lens onto a newly manufactured lens that was to be hand selected by the quality control people in Solms, and then returned to me via Leica NJ.

When I received the 100mm APO back it was not only worse than the one I sent in, the back of the lens was all scratched up with screw driver marks where it had been dis-assembled and not even touched up. There were finger prints all over the lens, and the plastic bag that I had shipped it in was not even returned (I had sent box, pouch, plastic bag). I ended up sending this lens back to Leica NJ, and Brenda Olesin was so embarrassed she offered me a full refund on the lens.

I have an email from her confirming that Leica Solms was going to send me a brand new hand selected lens, if anyone would like a copy I will be glad to send it to them off line for a "reference".

Does this sound like a company that occasionally sends out a bad product?

Tom Henson

...

From: ca_creekin@yahoo.com (Bill Tuthill)
Newsgroups: rec.photo.equipment.35mm
Date: 31 May 2001
Subject: Re: Testing barrel/pincushion, distance?

rmonagha@smu.edu (Robert Monaghan) wrote

> yep, sad but true ;-) - as y'all have discovered, it also varies by
> focusing and zoom settings etc. as well as inherent in lens design etc.

For a great example of sample variation, see Jim Tardio's new review of the Nikon 24-85/2.8-4. He bought three (3) lenses before getting a good one!!! The first was unsharp at any aperture, the second better but still inferior to his 28-105/3.5-4.5, and the third one was acceptable.

http://www.jimtardio.com/24-85.html

> rigorous measurements require some rather expensive gear; the usual
> amateur tricks like projecting a slide of flat lines on a lens test   chart
> and measuring divergences from straight lines is problematic - most   slide
> projector lenses are way worse than the typical zoom or fixed lens ;-(

My method for finding barrel or pincushion distortion is to scan a negative or slide. I very much doubt that a progressive patch scanner would introduce linear distortion. It's easy to use photo editing software to draw a line where distortion is evident. I suppose I could measure the number of pixels off straight and calculate percent distortion...

From Leica Mailing List:
Date: Wed, 06 Jun 2001
From: David Prakel dgp@btconnect.com
Subject: Lens tolerances was RE: [Leica] 35mm R Lens recommendation

owner-leica-users@mejac.palo-alto.ca.us wrote:

> In a nutshell, the 35/1.4 - R is the best 35 lens I have used, including
> the 35/1.4-M asph.  But perhaps I was just lucky with the one I got.
>
> This brings up an interesting point regarding sample variations. Years  ago I
> had 3 Nikkor 28/2.8 AIS lenses. One was noticably superior to the other  two;
> so much so that I could pick out photographs I made with that lens.

Flashback to the 60s. There was a good deal of laughter in the camera shop when my father took home their entire stock of new Nikkor 50mm f2s to test and select the one he would keep. There wasn't much laughter when they saw the results of the best of the bunch! I learned an important lesson.

A further baffling observation: I have owned a Canadian M-Summilux 35mm f1.4 since the early 80s and it is probably the one lens I would not part with - even for another 35mm Summilux. On the strength of its performance a good friend cleaned out his piggy bank to buy a lightly used M6 and an original Summilux 35mm f1.4 ASPH. His lens is best described as an occasional performer! He is plagued by vignetting and strangely shallow depth of field as well as adequate but not startling resolution from his theoretically 'better' lens. I can only think I've got a lens from the top of the tolerance range and he's got a dog.

- --
David G Prakel

From Rangefinder Mailing List;
Date: Tue, 23 Jan 2001
From: Stephen Gandy Stephen@CameraQuest.com
Subject: Re: [RF List] Lens Tests Reliability

It's not unusual to find people who got very different results than you from the same type of lens. How is this possible if published lens tests are accurate?

The basic problem with lens tests reliability, besides the quality of the test procedures, is that lenses DO vary in performance right off the assembly line, and they can vary even more by the time they are delivered to you right off the shelf.

Don't fool yourself by putting too much faith in lens test results.

These differences show up when different lens testers test the same lenses, but rate them differently when compared to each other. Lens testing procedures do vary. While its difficult if not impossible to directly compare two different lens tests of the same type of lens if different test methods are used, the info does become useful when both testers test the same family of lenses, and rate them differently compared to other lenses within that family. These differences are more likely to show the lens variations of a lens lineup, than bad testing techniques.

If you want truly accurate lens tests of your lens, you must test it yourself, and not trust published lens tests to give you your answers. Your lens can be a dog or a masterpiece, and your friend with the same lens can have the same or opposite results. The higher the priced lens, the better the quality control, the less likely this is to happen, but it still happens.

Besides manufacturing tolerances, failures in assembly, mistakes in assembly by new people, failures in quality control, all of which affect the lenses coming off the assembly line, what happened to your lens AFTER it left the factory ?? How many times was the crate dropped on its way to the camera store ? How many hours was it vibrated during shipment ? Did the previous customer in the store drop it on the carpet an hour before you bought it ? Lens tests themselves have another set of problems.

Why do magazines and lens testers do lens tests, knowing there are problems with the test reliability ? To sell you magazines and books. Why do lens manufacturers publish lens tests and encourage testing ? To sell you lenses.

OK so the average lens test results over many sources will give you the general quality of a lens, it will still not give you the quality of YOUR lens, which may or may not have been dropped numerous times in the crate by a teenager on drugs trying to learn how to use the loader, or who knows what else. To be sure, test your own lenses.

Stephen Gandy

Joachim Hein wrote:

>  A german photographic magazin `Foto Magazin' reported flare problems  for
> the Ultron 35/1.7 and Color-Heliar 75/2.5 (I didn't check the other
> lenses carefully). If you understand german you can get pieces of their
> tests at www.evita.de. Nonetheless they gave top marks. I have to say I
> don't regard their tests as very reliable. (E.g.: They once gave the
> same top marks to the lens in the Olympus mju-zoom with 35-70. The
> results I got from this camera where really rotten and didn't live up to
> 9.something out of 10.)

Date: Tue, 24 Jul 2001
From: "Art Begun" beguna@mindspring.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Bronica RF645 vs. Leica for street photos

I remember Pentax recalling lenses 30 years ago because of poor performance. They blamed a supplier that had a fitting that was the wrong depth for an element or something like that. But you wonder what percentage of lenses are checked for quality. Pentax did not find the defect. Camera 35 did during testing.

From Rollei Mailing List;
Date: Mon, 02 Jul 2001
From: Eric Goldstein egoldstein@usa.net
Subject: [Rollei] Re: Tessar vs. Planar

Alan Zak wrote:

> Over the years I have used several Rolleiflex/cord lenses, including  f2.8 and
> 3.5 Xenotars, f3.5 Planars, Xenars and a Tessar on about 12 (or more?)
> Rolleis.  My impression was that there were more variations between  samples
> than types.

I second this. There is a lot of individual variation between cameras of this vintage... some of it is the result of poor repairs, but some is the luck of the draw... there was a lot of manual hand work in producing photographic glass and lenses in the up to the 60s...

If you get a "good one," keep it!

Eric Goldstein

Date: 02 Aug 2001
From: shadcat11@aol.com (ShadCat11)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: blind tests again ;-) Re: Some lens tests (Leica, Nikon and CV)

...

I have owned several samples of lenses from reputable makers which performed differently one from another. Those included Nikkor 28 and 50 mm lenses and two samples of 90 mm Summicrons, and several samples of Rollei and Hasselblad lenses. The differences were minor (in some cases, not so), but noticeable.

Lens tests can be useful, but be careful to not place too much value on them.

Allen Zak

Date: Fri, 11 May 2001
From: "Alan Bell" alanb@blk.com
Newsgroups: rec.photo.equipment.35mm
Subject: What Does Lens Sample Variation Mean?

What does lens sample variation mean? I can think of three metaphors.

1. Normal probability curve metaphor. Most lenses fall along a center axis but some are worse and some are better on a continuum.

2. Copying machine metaphor. Most copies that come out of a Xerox machine are identical. But occasionally the copier will jam a piece of paper will come out wrinkled or crooked. Most lenses are identical, but occasionally there will be a bad one.

3. Film batch metaphor. Rolls of film are the same within a batch but vary from batch to batch. The variation from batch to batch isn't considered a mistake, but for critical work photographers want their film to be all from the same batch. Lenses made from certain runs of glass or better than those made from other runs of glass. They all fit the specficiations, but you want one made on a "good day."

Additionally, if you are evaluating a lens, how do you determine whether what you are seeing is characteristic of all lenses of this particular make and model, or whether what you are seeing is peculiar and specific to the particular unit you are evaluating?

Bottom line: I bought a used Nikon AF28-105 f3.5-4.5D from a reputable mail order retailer. It hasn't arrived yet. But when it does, suppose I take a picture and it's a little soft. How do I know if the degree of softness is to be expected in a Nikon 28-105 or is out of bounds.

Date: 11 May 2001
From: ppestis@aol.com (Ppestis)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: What Does Lens Sample Variation Mean?

>Because only a small portion of all lenses are tested for quality control
>purposes, a bad lens can be sold even with reasonable quality control>>

How do you know this is true that only a small portion of lens made are tested for QC, have you ever been in a lens manufacturing facility? I would suppose that lens, like most Japanese high precision manufacturing operations use in -line QA process testing as well as SPCs with upper and lower control limits, I would higly suspect CPK analysis is also used to benchmark the design and production phases, so in fact you won't get a lemon. It would be interesting to hear from an engineer or manufacturing manager that works for Canon or Nikon, relative to what QA (not QC) steps are in place and what test specs are utilized.

Date: Tue, 15 May 2001
From: "bbb_bbb" bbb_bbb@bbb_bbb.ca
Newsgroups: rec.photo.darkroom,rec.photo.equipment.medium-format
Subject: Re: Asking for Help from MF Shooters & Printers

Clint Bailey wrote in message ...

>Especially now,
>that most of the glass is designed by computer and also with the
>majority of the quality control is also done by computer.

Clint,

QC is not "done by computers," at least for the better lenses. Go to the "Virtual Tour of Lens Production" on the Zeiss website to see how modern lenses are made:

http://www.zeiss.de/de/photo/home_e.nsf

As for your original question, I suggest that you simply trust your eyes. Image quality is in the eye of the beholder (or the customer, if you are a commercial photographer).

You claim to be able to see a difference between the two systems that you mention, yet you "would tend to agree with" people who claim that there is no difference. In my case, I would trust my own judgement over hearsay.

Bernard

From: Bob Salomon bob@hpmarketingcorp.com>
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: mfger supplied lenses for tests was Re: sample variation...
Date: Tue, 02 Oct 2001 

"Tom Bloomer" bloomer@snip.net> wrote:

> I guess if I was making a living with photography, I would want to know all
> the strengths and weaknesses of a lens before I decided to keep it.
> However, I would at least want to narrow the field down before buying the
> lens.
> 
> If the manufacturer is supplying lenses to Photodo for testing, I hope they
> *are* sending their best samples.  I would not buy a lens that shows up as a
> dud in the Photodo graphs unless I could find other evidence that the lens
> had merit.  Somewhere I have to get enough information to initially *want*
> the lens.  I do research before I plunk down my money.  There are about a
> dozen web sites that I have found with lens and camera reviews.  I have gone
> and found lenses that scored high on Photodo that got good reviews in all of
> the places that I can find them on the web.  I have also found lenses that
> scored low in Photodo, that have gotten bad reviews in all of the places
> that I can find them.  No matter how much I want that particular lens, if it
> scores low in the MTF and I find 5 people who have used it and don't like
> it - I'm not going to buy it and run a roll of film through it to see if
> they were wrong - DUH!
> --
> Tom Bloomer
> Hartly, DE
> 
> "Robert Monaghan" rmonagha@smu.edu> wrote 
> 
> re: mfgers cherry picking lenses?
> If nothing else, consider the fact that most lenses for magazine tests
> (even big mags like Modern Photo. and Pop Photo) are supplied by the
> importer or manufacturer. Now, if your sales depended heavily on how those
> lens tests came out, would you send a potential dud or dog lens, or would
> you be sure to send the best of the batch? ;-) Duh? ;-) What do you think
> are the real chances that you or I will get as good a lens as that one
> handpicked by the mfgers to be tested? Oops! ;-)
> 
> What do you think happens to all those dud lenses, the 2% or 3% or 5%
> which are returned for serious optical problems, at the mail order stores?
> DO you think they test them to see if they have a problem, and return them
> to the mfgers for realignment and reassembly? No, huh? Think they just
> stuff 'em back in the box and ship 'em out to the next buyer, like you or
> me? Duh! ;-)
> 
> re: MTF charts and me ;-)
> I can't claim to be that consistent; actually, like most folks, I find the
> photodo single number rating useless, but the MTF charts are interesting
> but not sufficient to select a lens for use. For that, I have to have the
> lens in hand and actually test it, first with test charts and then in
> typical shooting situations.
> 
> re: why I lens test:
> The lens test chart shots are mainly to reassure me that the lens has
> not been abused or damaged in shipping or manufacturing and is not
> suffering from certain aberrations and glitches which are easy to see on a
> lens test chart, but less obvious in a stack of slides. I also hope to get
> a feel for how the lens should perform, where aberrations start to clean
> up, and so on. Is the lens best at f/8, or f/16? wide open? The actually
> shooting shots give me a feel for issues like bokeh and flare (eg, sunset
> shots) in situations I often like to shoot with that focal length lens...
> 
> re: humor
> Stephen Gandy has a fun page on cautions at:
> http://www.cameraquest.com/lenstest.htm
> 
> grins bobm
 

Well we are a distributor and we do send lenses out for testing to the 
magazines (but not Modern as they have been gone for over a decade).

And we don't pick special lenses or supply special lenses for testing. 
When we send lenses out for testing they are picked at random by the 
shipping room people and sent out to the magazine.

There is no way you could possibly do otherwise today.

What good would it do to select all of the highest quality lenses for 
tests and then sell lesser quality to dealers for retail?

If a magazine finds that a lens is a dog they will ask for another lens 
to test (the lenses are shipped and all kinds of things can happen in 
shipping. Ever watch a delivery driver at a loading dock?).

If multiple lenses test bad then they will come back and tell the 
supplier there is a problem. This is very rare with today's lenses from 
any major manufacturer.

Also most magazines will not do a comprehensive test of prototype lenses 
as these are not always typical of production lenses.

What you need to look for is if a test is really conducted by the 
magazine or is an "advertorial". An advertorial looks like a magazine 
test or article but is done by the agency for the lens/camera/accessory 
supplier. These will always be marked as an "advertorial" in small type 
at the top or bottom of the page, will only mention that company's 
product and are frequently designed to look like a report from a user 
and are usually followed in the magazine by a full page ad. Frequently 
you will find a series of these in one magazine with each seperated by 
the ad for the preceeding product.

But for a magazines own tests we, as well as everyother company I have 
worked for, Beseler, Rollei, EPOI, etc. did not and do not "cherry pick".

-- 
www.HPMarketingCorp.com 800 735-4373  US Distributor for:
Ansmann; Braun; DF Albums; Ergorest; Gepe; Gepe-Pro; 
Giottos; Heliopan; HP Combi-Plan-T; Kaiser; Kopho; 
Linhof; Novoflex; Rimowa; Rodenstock;Sirostar; 
Tetenal Ink Jet; VR Frames; Wista; ZTS 



From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 01 Oct 2001 
Subject: Re: Un-cementing lens elements

>Colyn wrote:
> 
>> You ever notice when a poster wants to look professional, he/she adds
>> a sig to reflect the particular discussion?
>
>Well, it was kind of funny.  
>-- 
>Anders Svensson, sprell checker
>mail: anders.-.eivor.svensson@swipnet.se


Well let me lay it all out for you. I worked for Hugo Meyer for about 2 years.
I wrote a monthly column for a number of magazines. In camera 35 I wrote Tech
Topics renamed Kramer's Korner.  Then I went to Modern Photography and wrotee a
column called The View From Kramer  for 20  years also doing all their lens
tests during that time. . I then went to Pop and wrote a column called Pro VIew
from Kramer for about three years. Also  my "day" job was as a Senior Vice
President Creative Suprervisor with the J. Walter Thompson Company, the world's
largest ad agency writing  Kodak's advertising during that period . I taught
View Camera and Zone Sytem at the New School for about 10 years. Also I flew in
the war. You might see part or all of the above as suitable where required.

It's been a full life. (g)


Arthur Kramer
Now is the time for all good men to come to the aid of their country
Visit my WW II B-26 website at: http://www.coastcomp.com/artkramer


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 01 Oct 2001 
Subject: Re: Lens Quality

>Source: Herbert Keppler, SLR - Can You See the Difference in Pictures Shot
>with a Super-high-quality Modern Lens and an Inexpensive Old SLR Lens?,
>Popular Photography, May 2001,


Yeah. Bert's tests haven't been too good since I stopped doing them. (grin)

Arthur
Arthur Kramer
Now is the time for all good men to come to the aid of their country
Visit my WW II B-26 website at: http://www.coastcomp.com/artkramer


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 01 Oct 2001 
Subject: Re: Photodo or PhotoDon't ???

>> >Of course, you were just going to tell us that all Leica and Carl Zeiss
>> >lenses have absolutely no sample variation of any kind, and that Zeiss
>> >lenses all perform precisely to the (identical) MTF curves supplied with
>> >every lens.
>> 
>> I never said that. You did.
>
>Hi Art,
>
>I believe you said it on here before, some months ago.  If that's
>incorrect, I apologise unreservedly and withdraw my remarks.
>
>-- 
>Best regards,
>TP
>


No two lenses are absolutely  identical no matter who made them. Lenses are like
fingerprints. You have to examine a few thousand lenses on an optical bench to 
realize that.


Arthur Kramer
Now is the time for all good men to come to the aid of their country
Visit my WW II B-26 website at: http://www.coastcomp.com/artkramer


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 01 Oct 2001 
Subject: Re: Un-cementing lens elements

>One important lens design parameter for easy
>assembly would then be "uncritical" designs, where
>performance isn't dropping radically if tolerances add up
>instead of even out.  

Simply said but not so simply donen. When elements are manufactured  they are
"graded" as to how they fall plus or minus their design aim points. Then they
are grouped  so the the errors are compensating rather than additive. In cheap
lenses  extreme pluses are used to compensate for extreme minuses. But in
better lenses these extreme errors are discard for remelting and
remanufacturing .  There is a saying that how good a lens a company makes is a 
function of how much they ar willing to throw away. Leica throws a lot away. On
another point, not all of Leicas  optical performance is  totally due to lens
quality. The Leica pressure plate design contributes as well. But that is
another subject for another day.


Arthur Kramer
Now is the time for all good men to come to the aid of their country
Visit my WW II B-26 website at: http://www.coastcomp.com/artkramer


From: rmonagha@smu.edu (Robert Monaghan)
Newsgroups: rec.photo.equipment.medium-format
Subject: magazine testing, only good news? Re: mfger supplied test lenses
Date: 3 Oct 2001 

Hmm? For someone who does testing of complex systems for a living, you
don't seem to have a lot of confidence in tests ;-) ;-) 

re: value of multiple test reports
Actually, I was interested to note that climatologists are using a large
number of reported temperatures to derive a highly accurate (0.1 deg. C)
average temperature (e.g., for Europe 100 years ago) despite working with
widely different techniques and instruments (thermometers, freezing lakes) 
not calibrated to a precise standard. 

I think the same process is happening heuristically with us using online
and published lens tests; with enough tests, you begin to approximate the
sampling issues which I am asking about, and can isolate lenses in which
there are some sample variation problems to be wary about ;-) 

So while tests may not be precisely comparable, the net tendency (pun
intended) does reveal aspects of the underlying lens population being
sampled, and that's why we go to the trouble of checking multiple sources
to get a "feel" for the lens quality and sample variation issues ;-)

still, we would prefer more accurate data, but there isn't any interest 
evidently on the magazines (as captives of advertisers?) or mfgers in 
providing such statistically valid data to us. They have it, we don't, and 
that's the way the industry likes it ;-)

my real point is that the issue of sample variation undercuts the validity 
of single point ratings (a la photodo) and even MTF charts from (possibly 
cherry picked) samples. 

I am also arguing that this is why you have to test the lens you select,
but also, as Bob S. pointed out, because abuse in shipping or other events
(demos? returned goods) may have impacted (pun intended) the lens and its
performance. You can't rely on somebody else's tests, but can only use
them to narrow the field of potential lenses to test...

If you were charging $1,000 US more for your lenses than a competitor, 
wouldn't you want to make buyers paranoid about their higher chances of 
getting a lemon lens with your competitors, if only indirectly by praising 
the tight tolerances of your own products? I would ;-) ;-) 

I think the real reason this info isn't more widely available is primarily
that the present system, aimed at the statistically non-savvy public,
GENERATES ONLY GOOD NEWS for the mfgers and magazines. 

We have some mfgers who publish MTF charts based on the ideal, computer
design lens, but not the actual production lenses.  Others publish lens
test data on a production lens, but only one lens, and not a spectrum of
performances within the production run (at +/- 1, 2, 3 s.d.), although I
presume they have such data from Q/C sources readily available. 

At best, a more valid statistical sampling view of actual production would
only bring the average quality expectations downward for buyers with
access to this info. So ideal or (possibly selected cherry samples) actual
MTF charts are likely to be only good news, yes? It also keeps buyers from 
trying to find the better lenses by cherry picking them in the stores, if 
they assume that all lenses are absolutely identical to 2 or 3 decimal 
places as in the magazine ratings ;-) ; -)

And as I have suggested, lower end mfgers would be nuts to not "cherry
pick" their better/best lenses which they provide for magazine tests,
given how critical such reviews are for their sales and profits!  ;-)

And we know that magazines such as MP/PP consult with advertisers about 
anomalous results, such as the bad production (mis-spaced) vivitar 
teleconverters I cited earlier in this thread. Again, I understand that 
magazines are dependent on happy advertisers, but the aura of independent 
lens testing is thereby suspect given such examples. 

Bob S also makes a good point regarding advertorials, which are ads
disguised as editorial commentary, but labeled as such in small type ;-).  
He is too much of a gentleman to point out the editorials and articles in
some magazines which are ghost written by lens mfgers' marketing types
(obviously excluding HPMktg here) and passed off as independent research
and writing in those magazines. Really now, how many really bad lens 
reviews have you seen? ;-) Okay, name just one bad lens review! ;-) 

And there is also the slavish copying of press releases in magazine
"announcements" section, down to glowing text about products, which is
often seen as a benefit for advertising (so many inches of free press for
so many pages of ads?). 

How many lenses have you seen reviewed under import or store labels (e.g.,
Ritz) which don't advertise in that magazine? Why do you suppose that is,
if the magazines are so independent? I mean, aren't there a lot of import
and store label lens sales by Ritz etc., versus the handful of pro sales
for lenses (e.g., fast 600mm f/4 teles) which are of minimal interest to
the average magazine subscriber but get lots of glowing reviews anyway?

And when a glowing report on say Leica M6 is published in a magazine,
doesn't there always seem to be a full page ad for Leica nearby? ;-)

Again, we as subscribers have to realize that we are the PRODUCT, it is 
our eyeballs (and wallets) which are being sold to the advertisers, at 
least in the major photo pubs which are filled with such ads ;-0) yes? ;-)
We only pay the postage and profits, the costs are from the ads. But it is 
important to keeping subscribers happy to promote our delusions that these 
magazines are really serving our interests as consumers, right? ;-) 

As for the magazines, how many of their subscribers resubscribe mainly to
get these lens reviews (given the repetitive nature of annual cycle of
articles)? If the magazines admit their single sample reviews are 
irrelevant, won't that cost them subscribers? and ad revenues based on 
eyeball counts? ;-)

If lens vary, and for 35mm prosumer/consumer lenses that seems to be about
+/- half a grade average (third/variations.html), then doesn't that put
their single test process into question as far as its utility in selecting
lenses? 

The difference in photodo scores between the average fixed lens and zoom
lens is only 0.8 out of 5 units, so a +/- .5 range within a lens batch is
larger than this average zoom vs fixed lens difference! It is also lots
larger than the range usually seen between lenses of different mfgers for
comparable cost similar zoom or fixed optics. 

This is partly why I have become more concerned with issues like film 
flatness, which is more often limiting the results from pro medium format 
optics, and higher resolution limit color films, than the optics 
themselves. These are elements of the final system resolution results 
which are easier to tweak than the individual optics, but which guarantee 
better results with all the lenses I'm using ;-) 

grins bobm



From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 02 Oct 2001 
Subject: Re: severe decentering samples etc. Re: Un-cementing lens elements

>re: missing the point ;-)
>no, I just didn't address that point ;-) I was responding more to the 
>inference that the German optical industry had this technology, but the 
>Japanese by inference did not. They did have automated lens centering 
>measuring and grinding equipment, at least at Nikon from the 1970s per 
>their published materials. That was my limited point in the note.
>
>re: radial vs. centering 
>When I disassemble and remount lenses, I do mark them to maintain radial
>orientation on reassembly. I don't put them on a University optical lab
>bench and try to recenter the elements ;-) 
>
>Maybe I should? http://www.foto.no/nikon/lens_zoom.html  at Bjorn 
>Rorslett's site http://www.foto.no/nikon/17_35_review.html#top I quote:
>
>Since this review was incepted, I have had shooting and testing experience
>with quite a number (more than 10) of 17-35 Nikkors both on my D1 and F5
>cameras. The reasons for this are the persistent rumours that this lens is
>inferior to the 20-35 Nikkor. When I first got my personal sample of the
>17-35 , it didn't take me long to detect this sample lens showed severe
>faults of focus shifts and decentered elements. Since the first two or
>three of these lenses I had used were superb, I wouldn't accept no less
>for my own use. So, I demanded a packing case filled with 17-35 lenses for
>testing from the national Nikon dealer, and surprise, got it! Thorough
>testing of all these 17-35 Nikkors indicated that mild decentering isn't
>uncommon, and the same goes for focus shifts. A sample size of 10 is to
>small to draw any statistical significant conclusions, but finding that 2
>out of 10 lenses had severe decentering problems was discouraging.
>Eventually I located a perfect 17-35 AFS and kept it for my personal use.
>It has proven itself a highly useful all-round lens on my D1. In fact, my
>records show I use it each and every day. Must be a real favourite, then.
>endquote
>
>for those who want to check, quoting the esteemed Mr. Gudzinowicz:
>
>If you want to look for a centering problem directly, there are a couple
>of approaches which require an improvised optical bench. Focus the lens on
>a small bulb filament covered with foil with a pinhole in it (star image),
>and rotate the lens. If the image moves around in a small circle when
>viewed with a microscope, it is decentered. The degree may/may not affect
>performance.
>=====endquote
>
>So I am aware that there seem to be some glitches in this lens
>auto-centering technology, at least at the level of the assembled lens ;-)
>I do find it disheartening that in this sample of high $$ nikon zooms (as 
>a nikon owner/user), a number with severe decentering problems were made 
>and presumably are in user hands. This is one of the few examples of a 
>test of a sample of lenses from the same batch, and shows dramatically my 
>main point that lenses vary, often by a lot, and even the high end high $ 
>pro lenses may have a surprisingly large frequency of dud lenses get 
>out...
>
>personally, I  would be interested in a magazine article by an expert lens 
>tester (Arthur Kramer?) looking into the issue of lens variations similar 
>to the above. So far as I recall, only once did we see a test of 3 lenses 
>(kowa 35mm normal lenses) in Modern Photography in which lenses were shown 
>to vary significantly between these 3 (and an earlier) lens which they 
>tested. To me, this was the most important article on lens testing they 
>had published, since it highlighted that lens sample variations were very 
>large even within the same batch for such optics, and by inference, other 
>lens makers as well. If lenses within the same batch varied circa half a 
>grade, on such a small sample, what must the outlying best and worst 
>optics be like? No wonder magazine reviews don't agree, they are probably 
>testing different lenses which vary significantly as shown by these 
>examples (including Mr. Rorslett's test of high end nikkors above..).
>
>bobm
>-- 
>* Robert Monaghan POB752182 Southern Methodist University, Dallas Tx 75275  *
>* Third Party 35mm Lenses: http://people.smu.edu/rmonagha/third/index.html  *

Excellant post. And of course getting a quantity of lenses and choosing the
best sample  is the best way to go. All of my Nikkors and Leitz lenses were
chosen that way except for  one Leica lens which was chosen for me  by Wetzlar
and proved to be outstanding. Also lenses which undergo 100% inspection at
every stage are quite uniform from one sample to  another. But the 105mm F/5.6
Apo El Nikkor with Quartz element costs $4,000 in a barrel mount. But I have
never seen a better lens anywhere. But as you pointed out. no two lenses are
alike. And a selection of one is no selection at all. Glad you liked my Modern
Photography tests.(s)

Arthur Kramer
Now is the time for all good men to come to the aid of their country
Visit my WW II B-26 website at: http://www.coastcomp.com/artkramer


[Ed. Note: I was very glad to see this note by Mr. Kramer, who is the person who 
performed thousands of the lens tests published by Modern Photography and later 
Popular Photography magazines. Mr. Kramer's posting confirms that many of those 
lens tests were performed on lenses which were "cherry-picked" by the lens 
manufacturers to ensure optimal performance. The chance of you or I getting such 
performance from a random lens off the production line is rather less, yes?...]




From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 02 Oct 2001 
Subject: Re: severe decentering samples etc. Re: Un-cementing lens elements

>No wonder magazine reviews don't agree, they are probably 
>testing different lenses which vary significantly as shown by these 
>examples (including Mr. Rorslett's test of high end nikkors above..).


One more important point.  When a magazine requests a lens for testing, the
maker carefully selects a superior example and sets it aside for magazine
tests. When I was testing for Modern I could immediately spot one of those
especially selected lenses. I would then insist on buying it. The maker
wouldn't admit it was a special lens since he had told me it was a random
sample. I got a lot of outstanding lenses that way. (grin)


Arthur Kramer
Now is the time for all good men to come to the aid of their country
Visit my WW II B-26 website at: http://www.coastcomp.com/artkramer 


From: artkramr@aol.com (ArtKramr)
Newsgroups: rec.photo.equipment.35mm
Date: 04 Oct 2001 
Subject: Re: severe decentering samples etc. Re: Un-cementing lens elements

> Also lenses which undergo 100% inspection at
>>every stage are quite uniform from one sample to  another.
>
>Perhaps that's why I've never had a cavil with even the most modest
>lenses for my Linhof when they came from Linhof.
>
>les clark / edgewater, nj / usa


Actually throughout the  lens production process testing is done on eacn Nth
element then on each Nth cemented pair or gorup then on each Nth finished lens.
The process I am talking abou tis where every element is tested and every pair
tested and every finished lens tested.  This is lens production of the highest
order and lenses made to these standards cost thousand each and in some cases
tens of thousands each. A quality of optics that most of us wil never see. A
case in point is the Apo el-Nikkor and the Carl Zeiss S-Planar.. scientific
application lenses that run about $4,000 each.

Arthur Kramer
Now is the time for all good men to come to the aid of their country
Visit my WW II B-26 website at: http://www.coastcomp.com/artkramer


From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: lens registration distances Re: Photodo or PhotoDon't ???
Date: 5 Oct 2001 

Robert:

rmonagha@smu.edu (Robert Monaghan) wrote 
> In any case, most MTF
> machines have features to set the lens at this distance, as has been
> noted, and testers can check to determine that this value in fact is the
> right point for any given lens, and if not, what its actual registration
> distance is, to precisely set the focal plane. On a $50,000+ MTF machine,
> you get all the extras ;-)

Robert:

The MTF equipment I've worked with does a through-focus MTF
measurement in order to precisely locate the optimum image plane
within about 5 microns.  Focal length tends to fluctuate much more
wildly in lens production than aberrations, so even first rate lenses
will vary by as much as half a percent or more from the design value. 
Any mechanical mount must take this into consideration.

Brian

From: brianc1959@aol.com (brian)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: which diff. limit rule? ;-)
Date: 11 Oct 2001 

Hi Robert:

You're absolutely right about 50mm normal lenses.  Interferometry,
aerial image testing or MTF testing might reveal slight differences at
f/8, but these differences would not show up on film or in electronic
format from a CCD.  This is probably true for most prime lenses in the
50-135mm focal length range, where lens design is a very
straightforward business and has been for about 30 years now. 
530lp/mm might be a stretch, and would definitely require the use of a
narrow band filter to eliminate all chromatic aberration.  I did check
a Nikon design I have for a 50mm f/1.4, and it will resolve just
beyond 500 cycles/mm in monochromatic light (550nm) at f/2.8 near the
optical axis.  This is not possible at f/4 or slower due to
diffraction.  Darn.

You're also right about wideangles, where color fringing usually
prevents really good stopped-down performance.  This is why the Nikon
17-35mm/2.8 zoom is actually better than most of the fixed focal
length wideangles that preceded it.  The zoom has little or no color
fringing except for a small amount at the extreme wide end.

Brian

rmonagha@smu.edu (Robert Monaghan) wrote 
> Thanks, Brian, for sharing these points and info ;-) My impression is that 
> quite a number of 50mm OEM normal lenses reach the high 350 to 530 lpmm 
> aerial resolution ranges at least centrally (not so hot on edges as you 
> noted ;-) per table XVII in Skudrzyk's Photography for the Serious Amateur 
> and calculations from known film resolution limits and observed on-film 
> system resolutions. The slower medium format lenses costing kilobucks may 
> be a bit of a compromise (due to coverage/cost..) but don't seem to leave 
> much room for improvement either, at least for central resolution ;-)
> 
> my original underlying point was that many 50mm normal 35mm SLR lenses
> tend to deliver reasonably similar resolution performance by the time you
> stop them down to f/5.6 or f/8, and that past f/8 the performance is more
> often limited by diffraction than the quality or cost of the lens might
> suggest. Now for many 35mm wide angles, I wouldn't make such a claim ;-) 
> 
> This is why I was not surprised that the pentax 50mm f/1.4 1974 screwmount
> lens when used at f/8 as in Popular Photography's comparison against a new
> Leica Summicron 50mm f/2 performed essentially at the same resolution
> levels (per Keppler's standards and eye). Even at the "sweet spot" of such
> lenses (f/4 to f/5.6), resolution performance is still rather good and
> similar. The real differences are more likely to be seen in other criteria
> than resolution (contrast, distortions...) and more easily seen wide open 
> and in the corners. 
> 
> Many 35mm photographers believe that they would get much sharper (&
> better) photographs by buying the kilobuck lenses instead of the cheapy
> normal lenses they currently have. For general shooting, where you can use
> f/5.6 or f/8 "sweet spots", I doubt many folks could reliably tell the
> difference at f/8 (at least for resolution) between a good cheapy OEM lens
> like the Pentax 50mm and a more pricey 50mm OEM lens by Leica or
> Zeiss/Contax (bokeh, yes/maybe, but resolution?). That's my theory anyway.
> 
> I'm running a blind lens test now in medium format comparing different 
> lenses ranging from $3k US to $40 TLRs. It will be interesting to see just 
> how many (or how few) people can reliably sort which shots were taken by 
> which lens etc. I suspect a similar test for 35mm optics would be quite 
> sobering to many folks who believe a particular brand of lenses is clearly 
> superior to other lenses? ;-) 
> 
> grins bobm



From: ChrisQ lightwork@aerosys.co.uk>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Leica-Konica incompatibility?
Date: Fri, 12 Oct 2001 

Godfrey DiGiorgi wrote:

> 
> Tolerances for back focus and film plane alignment are usually in the +/-
> .0005-.001" range (to account for film curvature) even on relatively
> inexpensive cameras.
> 

Need to be convinced. Any documentary evidence of that or links ?.
Even to provide tolerances of +/- 1 thousandth of an inch, (0.001"),
is difficult in production and would be degraded by the expansion and
contraction of the material with temperature, even more so in the case
of a plastic body. The coefficient of expansion  of aluminium =
0.00001244 per unit length. Assuming 1" between lens mount and film
plane and a temperature range of 32-102 degrees F, we have:

Aluminium: 0.00001244 x 70 = 0.00087"
or
Brass:	0.00001 x 70 = 0.0007"

In each case, about 3/4 of a thou over temp range.

Would think you are probably an order of magnitude out. Would expect
manufacturing tolerances for something like this to be of the order of
+/- 0.005", for mass produced cameras, maybe less for hand made stuff
like Leica and would be one reason why Leica are more expensive, since
adjustment to such fine tolerances is very labour intensive. However,
I doubt if much less than this sort of tolerance could be maintained
over thousands of lenses / bodies and decades of production.

To give an idea of scale, a human hair is around 0.003" thick, as is
80gm laser printer / photocopy paper.

Chris


From: ramarren@bayarea.net (Godfrey DiGiorgi)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Leica-Konica incompatibility?
Date: Fri, 12 Oct 2001 

Again, read Norman Goldberg's book. The 35mm film channel has been around
since the 1930s with little change in dimension, and about .0005"
clearance for film is correct. Backfocus and other specifics determine the
total allowable variation to between .0005 and .0010 inch for best
results. 

The pressure plate being springloaded does not change this.

Godfrey

mceowen@aol.com (McEowen) wrote:

> >> Tolerances for back focus and film plane alignment are usually in the +/-
> .0005-.001" range (to account for film curvature) even on relatively
> inexpensive cameras.  
> 
> 
> That's pretty hard to swallow since most pressure plates are spring loaded . .
> .


From: "Christopher M Perez" christopher.m.perez@tek.com>
Newsgroups: rec.photo.equipment.large-format
Subject: Re: modern 65mm Super Angulon???
Date: Mon, 29 Oct 2001 

Kerry and I tested one 80mm XL (1, uno, just a single example).  It wasn't
all that impressive.  I hear that the factory looked at the lens and said it
shouldn't have been shipped.  It was out of 'spec'.

If there are other 80XL examples being poor wide open, then I don't know of
them.  But if there are a few more, then Schneider probably should be doing
something to correct a production problem.  I've heard that other 80XL
lenses perform brilliantly just like it's longer 110XL brother.

- Chris

"Keith Olivier" keith.olivier_nospam@t-online.de> wrote 
> Considering that the 80mm XL was the last of the aspheric lenses to go in
> production at Schneider, I would find it very hard to believe that it does
> not measure up in terms of quality.  Of course I haven't used it myself, but
> it is certainly a lens that would be on my shortlist for roll film format
> fotography.
>
> One would first suspect a problem with the position of the film plane, but
> it is hard to imagine a situation where someone has a camera with only the
> 80mm XL lens.  Any deviation of the film plane position would reveal itself
> even more sharply at longer focal lengths.  The results should certainly be
> better than the older design Grandagon or Super Angulon systems.
>
> Very puzzling.....   What was the outcome of the complaints ?   Normally
> every lens is accompanied by a signed test report.  Were the lenses
> returned, exchanged or what exactly happened ?
>
> Keith Olivier
> "dg" sacripant@online.fr> schrieb...
> > > Then there is also the 80mm Super Symmar XL which is admittedly somewhat
> > > longer, but very compact & lightweight and very highly corrected.
> >
> > i've read many times (internet forums...) that this lens is not good
> > as his big brother, the 110 xl !
> > with the 80xl you need to stop down to have good result, is not good
> > at all at large aperture.
> > i've never test it by myself...
> > rodenstock grandagon N 75/4.5 and the schneider 72 xl are suppose to
> > be excellent performers (big and heavy !!!)
> > the only lens that can be use on a technika (folded) around 80mm are
> > old ones with a very restrictive image circle !
> > ok for roll film, but not for 4x5.
> > a good compromise seems to be the 90/8 from nikon, sharp, light
> > weight, image cicle 230mm, still too big for your use


From: Steve Bell steve@sb-technical.fsnet.co.uk>
Newsgroups: rec.photo.equipment.35mm,uk.rec.photo.misc
Subject: Re: Sigma 24-70mm DG
Date: Mon, 12 Nov 2001 

"Anthony Polson" acpolson@hotmail.com> wrote:

{snip}

> Like most people who buy Sigma, I was attracted by the claims of similar
> quality to camera brand lenses at a lower price.  In fact, you can't
> ever seem to reproduce (with your own Sigma lens) the amazingly high
> optical standards of the Sigma lenses that are reviewed in magazines.
> 
> I wonder why that might be ...

I suspect everything going for magazine reviews is specially selected or
tweaked. 

I used to tweak monitors for one particular manufacturer before they went
for magazine reviews. For major reviews we would work on 6, then select the
best. Then to avoid the possibility of shipping upsetting the adjustments it
would go door to door by taxi. I even watched a dummy review at PC Mag so we
new what parts of the spec to concentrate on. Never ever trust magazine
reviews.

Steve Bell



From: David Kilpatrick iconmags@btconnect.com>
Newsgroups: rec.photo.equipment.35mm,uk.rec.photo.misc
Subject: Re: magazine test 'fixing' (was Sigma 24-70mm DG)
Date: Tue, 13 Nov 2001 

Skip wrote:
> 
> Ford used to deliver heavily leaned on Mustangs
> to magazines for testing, once, in 1970
> delivering a nearly fully race prepped Boss 302
> with a trick twin inline four barrel set up that
> was never actually in their catalogue.  I never
> trust any magazine test, no matter what the item
> being tested.  At least, not fully.

re:

> > I suspect everything going for magazine reviews is specially selected or
> > tweaked.
> >
> > I used to tweak monitors for one particular manufacturer before they went
> > for magazine reviews. For major reviews we would work on 6, then select the
> > best. Then to avoid the possibility of shipping upsetting the adjustments it
> > would go door to door by taxi. I even watched a dummy review at PC Mag so we
> > new what parts of the spec to concentrate on. Never ever trust magazine
> > reviews.

That's just not true as far as the photo press goes. I've worked on
photo mags since 1974 and I can tell you that the worst instance of
fixing you can get is where a product hits your desk and is so crap you
are obliged to send it back and request another sample. In some cases I
have picked products at random out of the warehouse; in others I have
insisted on having the pre-production sample being shown at photokina or
Focus, or some other defunct trade show, and setting it up outside then
returning it to the stand.

Very often, not only are the products not hand-picked, they are
pre-production samples which are inferior to the final item. This was
the case with our Dimage7 - the example we bought a month later, when
the 'real' cameras arrived, is far better than the test camera was - and
all the press had this same test camera in turn, or its brother. Far
from being able to pick 'specially good' or tweaked examples, Minolta
only had two samples for their entire UK roadshow and promotional
programme including the press loans.

There was an inference that in the early days of CZ Scientific's
distribution, Sigma lenses were hand picked. I was working on PHOTO
TECHNIQUE at the time. The fact is, we got dreadful examples and good
examples. We had to go through three of their first 80-200mm zooms
before finding one which was sharp. Today they are much better and I own
and use shop-bought Sigmas which are as good as anything tested.

When you read most current European photo mags, you are seeing test data
produced under the auspices of TIPA, Tecnical Image Press Association.
TIPA set up a system, working with Zeiss's test labs, that permits a
number of identical lenses to be tested very thoroughly; at the time,
they said five samples of each lens would be needed. You can see the
same TIPA test data in British, Scandinavian, French, German, Spanish
and probably other mags - it is cost-effective to subscribe to this, and
by cherry-picking from the databank, a magazine can do things like
'group tests' without ever needing to get a group of lenses together.
Generally you can tell when this is done - the illustrations show all
the lenses, but there's no big picture with the entire bunch. Amateur
Photographer does not use TIPA but has its own facilities and may at
discretion test one sample, or multiple samples.

The TIPA databank of test results pretty much eliminated the possibility
of 'special quality' test lenses reaching the press. Since all TIPA
members exchange magazines (we are no longer TIPA members, they hold
their annual meeting on my wife's birthday every year and it seemed
pointless to stay in!) any local fixing in one country would soon be noticed.

Modern warehouse systems mean that if I call a distributor to request a
test product, the PR dept doens't even touch it, let alone pick it by
hand. They raise a loan invoice and their warehouse does the despatch,
picking the first box off the shelf. It is not even treated as a 'loan'
and publicity depts do not usually have loan stock the way they did 20
years ago. The item is invoiced out, and the invoice cancelled on
return; the item then usually passes directly to B-stock, and may be
sold at a discount to a dealer as ex-demo. Some companies do have loan
stock pools, or a mechanism which permits loan invoiced product to be
sent on to a further reviewer; example, Fuji with S1s. This means that
every different magazine ends up reviewing the same camera serially.
Others have budgets or limits on how much stock can be loan invoiced,
but as long as they are within budget, six different writers could
request the same product and get it simultaneously, different one for
each mag: example, Nikon. This also means that if we want to buy the
product after testing it (which we have done with several Epson printers
and scanners, two Apple Macs, two Olympus Camedias, one Coolpix 990 -
etc) they just need to leave it with us, and ensure a trade price
invoice completes the cycle.

Silly situation with Minolta, although we do a magazine for their users
- we can't buy the loan stock, ever. Once raised on a loan invoice, the
gear must go back for a full credit. Then if we want we can buy at trade
price a different actual camera. So even if we found a superb, one in a
million lens and called to say 'PLEASE! LET US BUY THIS!' we couldn't -
that lens would end up sold to a dealer as ex-demo, and we would get
another one brand new from stock.

As for computer monitors, the correct procedure is to deliver a monitor
by hand, install it, advising the customer on correct orientation and
ensuring no adverse influences are close to it; leave it in position for
48 hours; then return with the necessary tools or utilities to complete
full set-up and adjustment. Of course, like everyone else we just order
ours by mail, they come by carrier and we plonk them on the desk and the
only adjustments done are the on-screen set up options. Resulting final
score: Mitsubishi Diamondtron 3, Apple Colorsync 0. Four years down the
line all three of our brand new Colorsync had failed totally, needed
repairs worth more than the cost of each monitor. Replacement
Diamondtrons have proved, so far, to be not only 800 per cent sharper
and better to use, but free from sudden death; oldest now about 3 years
old (first Apple monitor went phut after about 14 months).

Now of course, any magazine test of monitors would fail to reveal the
now legendary unreliability of Apple 17 inch Colorsync monitors - and
tests at the time praised them highly. I guess we have the same problem
with digital cameras; how long before pixels start dropping out on the
Coolpix 990 (answer - ex factory in many cases)? When will the
electronic viewfinder on the Dimage7 pack in?

Because of factors like this, I've long ago stopped doing 'technical
benchmark' tests; all I consider today is the end result, and my
reaction to the design, features and handling of the product - and how
it fits in against competition, and into the historical development of
similar products. I don't draw any conclusions which go beyond personal
opinion, since without testing the product for its working lifetime and
in all the conditions likely to be encountered by every user, I will
never even begin the find out whether it has inherent faults.

David Kilpatrick

former - Technical Editor, Photography magazine; Associate Editor, Photo
Technique; Contributing Editor, You and Your Camera; Editor, The
Photographer (twice); Editor, Creative Photography; Editor, The Master
Photographer (twice); publisher and editor, PhotoPro/Photon/Freelance
Photographer; ditto 35mm Photographer; ditto Photo Club News; ditto
PhotoExpert; director and editor, Minolta Image since 1981; test reports
for numerous others ranging from SLR Camera, to Camera User, to AP
(secondhand items only). Currently publishing Freelance Photographer,
Master Photographer, Master Digital and Minolta Image and unaware of ANY
deliberate fixing or prepping or test products we have tested. If
anything to the reverse, we're fed up with being sent beaten up demo
stock or lent a printer which has just done a season of trade shows.

http://www.freelancephotographer.co.uk/


From: Stephe ms_stephe@excite.com>
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 501cm/501cw for landscape? RF? Help pls
Date: Tue, 15 Jan 2002 

John Halliwell wrote:

> Steve Gombosi
> sog@niwot.scd.ucar.edu> writes
>>I'm curious why you haven't considered used equipment.
>>It sounds like you're only thinking about new gear. One of the
>>great advantages of the Hasseblad system is the vast amount of
>>used but perfectly functional equipment that's available.
>>
>>You could save a lot of money and end up with a much more flexible
>>outfit by going used.
> 
> By the same token, there's a lot of beaten up stuff out there as well.
> If you know what you're looking for, maybe you can cherry pick through
> it all. Personally unless I'm happy that I know what I'm looking at and
> respect the dealer, I won't buy used kit.
> 

The bitch is there is a sample to sample variation in everything 
manufactured. I've heard of brand new late schneider "high end" view camera 
lenses that are suposed to be individually tested being so bad they 
couldn't even be brought into focus. At least with used, if it turns out 
the one you bought isn't -cherry- you can probably sell it for close to 
what you paid and try another.  I've had to do this with every camera 
system I've ever owned and frankly some of the more beat up looking stuff 
performed better than other "mint" condition ones did.
-- 
 
  Stephe

From: ralf@free-photons.de (Ralf R. Radermacher)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Pentacon Six === Yashica MAT 124G
Date: Fri, 3 Aug 2001 

mayo kuran.marcin@lumena.com> wrote:

> I decided to replace my Pentacon Six (standard set) with Yashica MAT 124G. I
> just want to have ligher and more compact 6x6 camera especially for mountain
> hiking.

Premise: I'll let my Exakta 66 count as a P6, here. The differences are
insignificant for the purpose of this discussion. 

Now, to someone who's had both cameras but upgraded in the other
direction, this sounds like a rather odd suggestion, at first. On second
thought it, does make sense in your case but let me add a few ideas and
suggestions.

For mountain hiking, you'll certainly appreciate the lower weight of the
Mat. I've been very pleased with mine and was sorry to part with it when
I bought my first Kiev 60. I had to, because I was beginning to have
trouble focussing with the WLF. Old age hitting me at 45....

The finder of the Mat is a little darker than that of the P6 which
shouldn't be much of a problem if you're outside.

> Can you compare optical quality of  Biometar 80mm f/2.8 to Yashinon??

Both lenses are subject to substantial variations. I've had a total of 3
Mat's and one of them was really excellent, definitely better than my
current Biometar. The other two weren't quite as good as the Biometar,
but not by much.

Be aware though, that you're giving up all options to work with lenses
of another focal length. The various tele and wide-angle attachments are
totally unuseable. Their optical quality is rotten and, worse, this
can't be compensated by stopping down because, due to their principle,
this leads to immediate vignetting. That goes for 3rd party lenses as
much as for the original Yashica versions. Don't waste your money
repeating other peoples' experience.

So, if there is a chance that you might want to use other lenses for
other things and you can afford to keep the P6 besides the Mat, I'd
suggest you do so.

> Is my decision reasonable??

If mountain hiking is all you want to do, then yes, I think it is.

Cheers,
Ralf

P.S.: Where on the net are your pictures from these trips? :)

-- 
Ralf R. Radermacher  -  DL9KCG  -  K�ln/Cologne, Germany
private homepage: http://www.free-photons.de
picture galleries - classic and mechanical cameras
Contarex - Kiev 60 - Horizon 202 - P6 mount lenses


Date: Mon, 23 Jul 2001 
To: rollei@mejac.palo-alto.ca.us
From: "John A. Lind" jlind@netusa1.net>
Subject: Re: [Rollei] Subscription and Unsubscription Instructions

Dale wrote (in part):
>Les
>
>My personal experience with the lenses is  no clearer. For example, I have 
>a Rolleicord Va with a very good Xenar lens. I can only tell it apart from 
>my Xenotar lens on a 16x20 because of the brokeh and not the sharpness. If 
>I can experience this close a result with  two difference lens designs, is 
>it any wonder the question between two similar  lens designs is muddy?
>
>Dale

Yes, the topic is severely flogged periodically with passionate opines.  My 
conclusions?  There is more variation in manufacture, and care since 
manufacture, of the Tessar vs. Xenar and Planar vs. Xenotar lenses than the 
true difference between them.  In other words, other factors confound 
making any statistically significant claims about which, overall, is 
"better" than the other, whatever "better" means.  The words "better" or 
"best," as applied to this topic are nebulous.  There are at least a 
half-dozen different, measurable lens performance factors I can think of 
while writing this email.

-- John



From: "Roy L. Jacobs" rljacobs@pipeline.com>
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: P67 Lens Question
Date: Tue, 25 Sep 2001 

If I singled out Pentax, perhaps it was unfair, but lenses do vary in
quality from sample to sample and the variations can be significant. My
worst experience did occur with a Pentax 67 lens. I had a 200(N) which was
bought new from B&H; when I first got my system. I did not use the lens much
when new. When I really starting using it, it became clear that it was not
so hot.  I sold it off. I then bought a used one with a return privilege. It
was like night and day. the 200 I have now is among the sharpest lenses I
have ever used. But since that experience, I test every lens I buy.

"Brian Ellis" bellis60@earthlink.net> wrote...
> I'm curious about the basis for your statement that Pentax lenses vary a lot
> in quality. Which focal length Pentax lenses of the same model have you used
> that have varied widely in quality? I've had quite a few 6x7 lenses and they
> have been uniformly excellent.
> "Roy L. Jacobs" rljacobs@pipeline.com> wrote 
> > At B&H; the price difference is not great, although it had been larger in
> the
> > past. The 45mm is an original design, the 55mm the third time around. Also
> > note that deltainternational.com sells the same lens at far less than B&H;,
> > but they will not take it back if you don't like the image quality.
> (Lenses
> > vary in quality from sample to sample, and Pentax lenses vary a lot.) I
> have
> > both and tend to the 55mm. It is one of the sharpest lens I have used, and
> > fine detail is excellently rendered.  That is not to say the 45 is a bad
> > lens. It is a fine lens also.
> >
> > "DR" rossiter@theboss.net> wrote...
> > > Just curious to find out why a new 45mmf4 lens for the Pentax 67 is
> > cheaper
> > > than the 55mmf4.  In 35mm a 20 is more than a 24, the 24 more than the
> 28
> > > etc (when the max/ap is the same).  I have a 55 and was thinking about
> > > adding the 45 and was surprised to see it advertised slightly below the
> > > 55...
> > >
> > > David


From: "Q.G. de Bakker" qnu@worldonline.nl>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Poor Mans Leica ?
Date: Sun, 2 Dec 2001 

Gordon Moat wrote:

> Erwin's results are based on one (1) sample Konica Hexar. Not exactly what I would
> call definitive statistical analysis. I e-mailed him about that, and asking about
> the standard deviation, and accuracy level of his tests. Any properly done
> scientific analysis should include statistical information, including possible
> error sources. Anyway, tough to draw such a conclusion from just one sample.

I'm sorry, but i do find this a bit silly. When you go out to buy a camera,
would you accept it if the manufacturer would explain to you that though the
particular one you got was a dud, the camera really is a top notch product
because "scientific" statistical testing has shown it is? You would expect
anyone of these to be good wouldn't you? Sure, they can't be all the same,
variation will occur. But it is the manufacturer's task to test and reject
before (!!!) they send the things out to be sold. So if you want a large
enough sample, want to know about standard deviation, accuracy level etc.,
go have a word with the manufacturer. If a single (!) camera that made it
through to the public would appear to be bad, this can only be seen as Q.A.
failing (not a good thing, alarming enough to mention in a review), or, even
worse, as a measure for the common "Q.A.-passed" level of quality (which too
should be reported in a review).



[Ed. Note: Mr. Polson's note highlights that even if you DO get an MTF chart with 
your lens, you can't be sure it is for some other lens, and simply reproduced as 
the standard chart. Even more alarming, it could be an idealized computer projection 
of what the MTF curve should look like if the mfger could produce the lens with 
precisely the positioning and materials parameters used in the computer modeling software. 
Very likely, few if any of the production lenses or glass batches will have such exact 
positioning parameters, and small differences can produce large reductions in some 
designs, so beware!]

From: Anthony Polson acpolson@hotmail.com>
Newsgroups: rec.photo.equipment.35mm
Subject: Re: variations in lenses.. Re: Poor Mans Leica ?
Date: Sun, 02 Dec 2001 

rmonagha@smu.edu (Robert Monaghan) wrote:
> 
> for higher priced lenses, where much of what you pay for includes many 
> tests, including hand inspection and even optimization, chances are pretty 
> good that you can and should expect to have lenses all meet a high minimum 
> standard. For example, rolleiflex tests 100% of its 6k lenses, some mfgers 
> supply MTF charts with their new lenses etc. But not all lenses get such 
> thorough testing. 

Hi Bob,

When you are talking about MTF curves being supplied with new lenses, I
hope that you're not thinking about Zeiss Japan's lenses for the Kyocera
Contax bodies.  They might be supplied with the curves, but all the
curves are the same for that particular optical design.  The curves are
NOT different for each new lens, so they are unlikely to be
representative of the one you just bought.

You may have been thinking about Zeiss Germany's lenses for Hasselblad
medium format SLRs.  I understand they are supplied with MTF curves
which *are* the results of MTF testing of that particular lens, *your*
lens.  So you would have been right, if only it hadn't been off-topic
for this NG.

{big G}

-- 
Best regards,
Anthony Polson


Date: Tue, 11 Dec 2001 
To: rollei@mejac.palo-alto.ca.us
From: Richard Knoppow dickburk@ix.netcom.com>
Subject: Re: [Rollei] Can I read some comparisons - 3.5 Xenotar vs.
  Planar?

you wrote:
>Can I read some comparisons between the Rolleiflex 3.5 Xenotar vs.
>Planar?
>
>Same for the 2.8 versions?
>
>Does anyone have experiences to share, or web sites to direct me to?
>
>Thanks
>Vick
>

  A serious answer. This has been discussed to death many times. A Google
search will turn up probably hundereds of Usenet threads and stuff on web
sites.

  In fact, there is probably minimal difference among these lenses. Testing
lenses is fraught with possible errors plus even the apparent performance
of the cameras in practical use depends on how well the finder lenses are
synchronized to the taking lens. 

  The subject is further complicated because two variations of f/3.5 lenses
were offered by both manufacturers, a five element lens similar to the
f/2.8 and a sixe element lens, pretty much a conventioal Biotar type.

  Schneider's five element f/3.5 lens is the same design as the f/2.8 lens
but the Planar is not. The five element f/2.8 Planar has its cemented
surface in the front component, the f/3.5 Planar in the third component.
This design is probably cheaper to make than the f/2.8 because the cemented
surface is plano and the very thin front element of the f/2.8 lens is
eliminated. 

  Both Xenotar designs use a plano cemented surface in the second component.
 The five element lens is derived from the original six element type by
combining two elements at front or rear and elminating one cemented
surface. Zeiss and Schneider went about this in slightly different ways but
were trying to meet the same specifications. I am not sure why the six
element design was adopted for the f/3.5 lens, its the faster lens one
would think would require the greater complexity. 

  One possibility is cost plus weight and size may have been
considerations.  The Tessar is not a very satisfactory lens beyond f/3.5 (I
am not refering to the defective Tessars used on early Rollei cameras but
to the generic design). Adding an element is a way of improving the
performance at larger apertures, especially away from the center of the
image. It may be that a six element f/2.8 lens would have been too large
and heavy for use in a TLR. The smaller f/3.5 lens may have alowed the use
of the more conventional design. I don't have any idea of relative costs of
the two types but it is not impossible for the sixe element lens to cost
less, despite the need for both an additional element and the cemented
surface (cemented surfaces are expensive to make) if, perhaps, some cheaper
type of glass could be used. 

  Marc migh have some insight or inside information about this. In any
case, all six lenses were intended to deliver deluxe performance, whatever
method was used to get it.  

  All six of these lenses are of suberb performance although people
certainly have preferences. 
  
----
Richard Knoppow
Los Angeles, CA, USA
dickburk@ix.netcom.com


From: dickburk@ix.netcom.com (Richard Knoppow)
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Super-Angulon disapointment ? Opinions anyone?
Date: Mon, 24 Dec 2001 

steadybrook@email.com (Ministry of Truth) wrote:

>I was just browsing through the results of a lens test at:
>
>www.hevanet.com/cperez/testing.html
>
>and happened to see the results of a lens model that I own.
>
>Its a Super-Angulon f8 serial # 5 244 xxx. (Mine that is)
>
>So far I shot a couple of test shots with it but didn't get anything
>worthwhile. The negs were sort of low contrast but I attribute this to
>the fact that I mounted it in a homemade lensboard made of aluminum
>and I neglected to paint the inside black.
> ( OK, so I was in a hurry and sort of newbie!)
>It was a 6" lensboard for a Cambo monorail so there was a lot of light
>bouncing around I suppose.
>
>Anyway, back to the main point....
>
>A NIKON SW f8 90mm was tested at 80 lines/mm center and middle and 
>60 lines/mm at the edge.
>
>The SA was tested at 67lines/mm center and middle and 17lines/mm at
>the edge.
>
>SEVENTEEN!!!!???
>
>How can this lens be so inferior to the NIKON?
>
>I had always thought the Schneider lens were among the best but this
>edge performance is only about 1/4 of the NIKON.
>
>Help! I'm suffering from lens envy! What's going on?
>
>Robert
>

  A couple of things. 

  If you look at the other tests Chris Perez has done you will find a
good deal of variation between the results gotten with lenses of the
same type. Some of this variation is due to actual variation in the
performance of the lenses and some is due to the experimental method.
 
  There can be an astonishing amount of variation among lenses of the
same type due to manufacturing variations and due to the glass
constants not always being what it is expected. 

  Chris is testing under real world conditions using film. Even very
slight variations in the focal plane of the film can make a large
difference in the measured resolution of a lens. 

  There can also be a variation in apparent resolution due to
differences in optical contrast. A clean and coated lens may show
better target resolution because the contrast is higher even though
the actual optical performance is the same as another lens with more
flare.

  Another point is that Nikon makes outstanding lenses. If one looks
at the big three for large format, i.e. Rodenstock, Schneider, Nikon,
there is surprizingly little quality difference among lenses of
similar application. I should add Fuji to this although they certainly
do not push their LF lenses. Fuji makes excellent lenses and has made
some of the best in the world in the past. 

  Real evaluation of a lens requires extensive testing. Some is done
directly on the image (aerial image), some using film. Really
reproducible film testing requires the use of a special camera which
insures absolute film position and flatness. Of course, its also
measuring the performance of the film.

  I think Chris has done a great service to the LF community by
running this series of tests but their limitations should be
understood along with the less than perfect quality control of lens
manufacturers. Its enormously better now than it was in the distant
past, but dogs still escape the factory once in a while. 

  If your Super Angulon isn't sharp there may be a couple causes.
 
  First, it may simply be dirty. Old lenses tend to get a coating of
haze inside which can destroy the contrast. Secondly, its possible the
lens was disassembled at some time and re-assembled with some error,
perhaps the lack of a spacer. I don't know if Schneider offers a
cleaning and repair service for its older lenses but there are several
people who do clean them and could also tell if the lens is
mis-assembled. This is probably not a good lens to start out with
doing it yourself. 

  Haze is easy to check for by using a flashlight. Shine it through
the lens and look for any haze inside. The glass should be sparkling
clear. A magnifying glass may help since the inside lenses are so
small. 

  Black masking tape inside the lens board will do as well as paint
for the time being. Krylon Super Flat Black is the best light
absorbing paing. Better hardware stores should have it, comes in spray
cans. 

  I suspect your low contrast is due to something other than the lens
board. 

 BTW, I have two older, chrome finish, f/8 S.A.s, one 90mm, one 65mm.
Both are quite sharp and good performers. These lenses need to be
stopped down to around f/22 to be sharp in the corners. 

---
Richard Knoppow
Los Angeles, CA, USA.
dickburk@ix.netcom.com


From: Stephe Thayer ms_stephe@excite.com>
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Super-Angulon disapointment ? Opinions anyone?
Date: Tue, 25 Dec 2001 

Ministry of Truth wrote:

> 
> The SA was tested at 67lines/mm center and middle and 17lines/mm at
> the edge.
> 
> SEVENTEEN!!!!???
> 
> How can this lens be so inferior to the NIKON?
> 

I have an old chrome barrel f8 90mm and it's just fine. Just as sharp as 
the 135mm WF ektar I own which pulled much higher numbers on their chart. 
When may be happening is several things as far as the rez tested on that 
site. There may be a little curvature of field with this lens at the 
distance it was tested at, making it look bad. This wouldn't show up in 3D. 
 Also this lens was tested at f11 and f16 while f22 would have yielded 
better numbers. Possibly MUCH better

Last and most important look at the results from the several f6.8's they 
tested. One only yielded 12 lpmm at f22 on the edge while another example 
of the same lens pulled 61 lmm at the same settings! Maybe the lens THEY 
had was this inferior to a nikon they tested? Maybe the lens they had was a 
dud? Look at the test of a brand new 80mm they returned because it was so 
bad they couldn't even focus it? Bottom line, just because a lens tests 
well on someones chart doesn't mean the one you have in your hand is any 
good at all. Conversely just because ONE sample tests bad, doesn't mean 
they all are bad. 

Too many people have used these for too many years for them to all be as 
bad as the one they tested. Mine for one is sharp even when shifted all the 
way to the corners at f16- f22. Go do some testing and I'd suggest ignoring 
lens charts that only have one sample of the lens you are using on it.

-- 
 
  Stephe



From: Stephe ms_stephe@excite.com>
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: 501cm/501cw for landscape? RF? Help pls
Date: Thu, 17 Jan 2002 

John Halliwell wrote:

> Stephe ms_stephe@excite.com writes
>>The bitch is there is a sample to sample variation in everything
>>manufactured. I've heard of brand new late schneider "high end" view
>>camera lenses that are suposed to be individually tested being so bad they
>>couldn't even be brought into focus. At least with used, if it turns out
>>the one you bought isn't -cherry- you can probably sell it for close to
>>what you paid and try another.  I've had to do this with every camera
>>system I've ever owned and frankly some of the more beat up looking stuff
>>performed better than other "mint" condition ones did.
> 
> That's why if I don't know what I'm doing, I buy new with a warranty to
> sort out any problems, rather than pass crap on to some other poor
> unfortunate.
> 

Do you honestly think they will exchange a new lens you feel is just not a 
great one? Sure if it's obvious (like so bad it won't focus) but if it's 
just "OK" instead of great, they aren't going to do a thing, you'll have to 
sell it at a HUGE loss instead of a small one. Read Bob's site about sample 
variations on EVERY manufacturers lenses. It's a fact of life, no brand is 
exempt and no two lenses are EXACTLY the same no matter how much we'd like 
to believe they are. 

One of the lenses I had to "dump" in my 35mm OM glass was a new 28 f2.8 
zuiko that just wasn't very good. I already had a beat up 28 f3.5 that was 
sharper and got really burned selling this new lens used. That was the last 
new lens I've bough unless I make sure I can return or swap it for any 
reason whatsoever. And should I have thrown it in the trash so as not to 
"pass on some crap to some other poor unfortunate" when the store wouldn't 
swap it just because I didn't think it was a great one? Doesn't matter if 
you "know what you are doing" when it comes to finding a good sample of a 
lens, it's just plain luck! Buying new and selling used to buy another new 
etc isn't a smart way to "sift" through the mediocre lenses that are sold.
-- 
 
  Stephe

From nikon mailing list:
From: "Roy L. Jacobs" rljacobs@pipeline.com>
Date: Thu, 31 Jan 2002 
Subject: [Nikon] Sample Variation

I learned this the hard way. When I got into medium format I bought a new
Pentax 67 with the 55mm(N), 105mm and 200mm lenses. I used the 55 and 105
mostly. Both are very fine lenses. When I started using the 200mm after a
year or so, it was a complete turkey. It was ok wide open, then so-so at
F.5.6 and so so beyond that.  That taught me to test each piece of
equipment. About a year ago I bought on approval in an Ebay auction a used
200mm lens. (Both were the redesigned model.)  The new one is tack sharp
wide open and stays that way. It is even tack sharp with the Pentax 1.4 
TC.

The new one was a dud and the used one is great. Test each lens; test each
body. Do not buy expensive used equipment without a right to return.  
Samples vary.




from nikon mailing list:
From: "Roy L. Jacobs" rljacobs@pipeline.com
Date: Thu, 31 Jan 2002
Subject: [Nikon] Sample Variation

I learned this the hard way. When I got into medium format I bought a new
Pentax 67 with the 55mm(N), 105mm and 200mm lenses. I used the 55 and 105
mostly. Both are very fine lenses. When I started using the 200mm after a
year or so, it was a complete turkey. It was ok wide open, then so-so at
F.5.6 and so so beyond that.  That taught me to test each piece of
equipment. About a year ago I bought on approval in an Ebay auction a used
200mm lens. (Both were the redesigned model.)  The new one is tack sharp
wide open and stays that way. It is even tack sharp with the Pentax 1.4
TC.

The new one was a dud and the used one is great. Test each lens; test each
body. Do not buy expensive used equipment without a right to return.
Samples vary.


From rollei mailing list:
Date: Tue, 19 Mar 2002 
From: Richard Knoppow dickburk@ix.netcom.com
Subject: Re: [Rollei] Leica lenses are not always the best.

you wrote:
>     Actually Marc that is not true. You have been listening to the Leitz
>propaganda machine again.
>As test by Photodo using Hasselblad MTF equipment (regarded as among the
>world's best), here are the weighted MTF results from each to compare.
>Canon 50 mm: f1 0,46, f1,4 0,55, f2 0,66, f2,8 0,74, f4 0,78, f8 0,81
>Leica 50 mm: f1 0,53, f1,4 0,61, f2 0,66, f2,8 0,73, f4 0,80, f8 0,85   As
>you can see the Canon holds its own next to the Leica. Very comparable.
>Yes, the Leica does slightly better but not by very much.   Now let's look
>at one more since you brought that up.
>Here are the results for 35mm F1.4 where the Canon is nearly 50% the price
>of the Leica.                        f4 0,82, f8 0,81
>Leica 35 mm: f1,4 0,54, f2 0,65, f2,8 0,75, f4 0,75, f8 0,83   Sorry they
>do not list the F2 and F2,8 for the Canon but the overall given the Canon
>is a 4.0 and the more expensive Leica only a 3.9...ouch! 50% more $$ for
>what?   Peter K   > -----Original Message-----
>> From: Marc James Small [mailto:msmall@infi.net]
>> Sent: Tuesday, March 19, 2002 4:12 PM
>> To: rollei@mejac.palo-alto.ca.us
>> Subject: RE: [Rollei] SLR v. Rangefinders

 MTF data like the above is not too meaningful since it is stated for only
one field angle, which is unstated. It is off axis where lens performance
varies most. Many lenses can deliver near diffraction limit resolution at
or very close to the center of the image, but have much poorer performance
at greater angles.

  Since both Canon and Leica have histories of building outstanding lenses
its entirely possible the performance of the two does not vary much at any
angle, but the above type of evidence is far from sufficient to make that
point.

  Most, if not all, modern lenses are designed with the aid of computers,
which can evaluate performance much more thoroughly than older hand
methods. They can also optimize design. The difference between two lenses
designed this way may be the acceptable cost of glass, which can also limit
performance, or some difference in acceptable tollerances.

  Even large price differences may be a matter of market share and
manufacturing volume  rather than quality.

  Its interesting that these are published _measured_ MTF rather than
calculated from the lens prescriptions, as are nearly all of the
manufacturer's published data.

 I would say the above numbers are probably identical within experimental
error. It would also be intersting to see comparisons of several samples of
each lens type to see what sort of variations occur in manufacture.
----
Richard Knoppow
Los Angeles, CA, USA
dickburk@ix.netcom.com



From Minolta Mailing list:
Date: Mon, 29 Apr 2002 
From: "aranda1984" stephen@aranda4.com
Subject: Re: Minolta lens test?

Chances are that they were all telling the truth, with the samples at
hand, as in one test, one brand came out on the top while the results
were reversed in another test.

This is why lens ratings can be so far our from one sample to the other.
Just imagine a good sample from Minolta against a bad sample from
Tamron. Now reverse the quality from the two different brands.

Today even Leica farms out the work in subcontract to various
companies. / Do not mention this to people who pay thousands more for
those not so Leica lenses. /
Depending on who made a certain sample, your results will vary a lot.

/ If you think you are driving a Ford or a GM car, I got bad or good
news for you. /


Being the Devils advocate, I think whoever is your biggest advertiser
might have something to do with the results also.  You can manipulate
the outcome from a very few samples by choosing better ones from a
certain brand against the poorer samples from another.

I rest my case.

Stephen I. Molnar

...
> Sure ... funny, though, while those magazine tests yield that
> result, photodo rates the Tamron far below either Tokina or Sigma or
> Minolta ... go figure :)
>
> - Dennis


From: drfleming@altavista.com (DFleming)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Lens specifications
Date: 29 Apr 2002

"PhotoMan" photoman@wfeca.net wrote 
> I've often wondered about tolerances in lenses. Knowing that the aperture
> ratio is determined by dividing the actual opening into the focal length,
> why do we see acceptable tolerances of � 5% ? An acceptable 10% range is
> ridiculous!

It is actually worse than you thought. The international standards
allow for as much as 1/2 stop at some apertures.

>From ANSI/ISO 517-1996, "Apertures and Related Properties Pertaining
to Photographic Lenses - Designations and Measurements":

   Tolerances of measured f-numbers:

   Full aperture         +/- 5%
   Smaller than f/5.6    + 12% - 11% (+/- 1/3 stop)
   f/5.6 and larger      + 19% - 16% (+/- 1/2 stop)

> How hard can it be to make the
> opening for a 200mm f4 lens 50mm ?

Probably not difficult, but it would be costly.  As the manufacturing
tolerances get tighter, the cost of production goes up.  You end up
with more rejects.

---
Don Fleming
http://www.dofmaster.com - Depth of field and hyperfocal distance
calculators



From: "Mike" nedsnake@earthlink.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Makinon Lenses
Date: Wed, 15 May 2002


Dan,  Makinon was a low end lens who's quality was a little hit & miss.
When I owned a camera store in 83 I tested a couple of 28-80 zooms with the
idea of selling one with a camera body.  One lens focused perfectly while
the other did not. The one that focused correctly was a very sharp lens. I
have an 11x14 of a light house on the Oregon coast, you can count the
vertical parts of the railing.

$70.00 is a fair price IF you are happy with the photos.  You can always
list it on eBay if it doesn't perform to your satisfaction.

Mike
...
[Ed. note: another reminder of why many older cameras exhibit a lot of variations...]


From: "Shinichi Hayakawa" shaya@plum.plala.or.jp
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Looking for Mamiya Six info
Date: Thu, 16 May 2002


"KFritch" kfritch@aol.com wrote...
> That's a pretty neat idea - sort of mail a lense and shutter to Mamiya over
> there in Occupied Japan and they cusomize it to a camera for you, but I don't
> really think it plays.

Hard to imagine now, but Mamiya actually did that--although not in Occupied
Japan, but in Empire of Japan.

I have read an article on the chronology of Mamiya cameras in a Japanese
camera book written by a Mr. Suzuki who was with Mamiya from 1953 to 1995.
According to Mr. Suzuki, Mamiya custom-fitted lenses and shutters supplied
by customers on request during WW2.  (I have posted a message somewhere in
this thread saying this service was done in " post-WW2 era," but I was
mistaken.  Sorry.)


That was possible because 1)labor was dirt cheap then, and 2)Mamiya Six
rangefinder mechanism was very easy to modify for nonstandard focal-length
lenses.

Shinichi



[Ed. note: another reason for variations in results in Large Format...]
From: dickburk@ix.netcom.com (Richard Knoppow)
Newsgroups: rec.photo.equipment.large-format
Subject: Re: Depth Micrometer Purchase?
Date: Fri, 24 May 2002

...
I've been using a standard combination micrometer and depth gage
with a reference plate. The plate is made of about 1/4" aluminum with
holes in it at the right places. Since I can measure the thickness of
the plate very accurately I can simply subtract it from the readings.
I lay the plate on the holder, avoiding the locating ridge, and
measure the depth with film in the holder. This seems to work well.
I've measured something like 100 4x5 holders and around thirty or more
each of 8x10 and 5x7 holders.

  I commented in another post about the variability of holders. New
ones are fine but many older ones, especially wood holders, are likely
to be off enough to cause noticable defocusing.
---
Richard Knoppow
Los Angeles, CA, USA.
dickburk@ix.netcom.com


Date: Fri, 21 Jun 2002
From: Christopher Perez chrisper@exgate.tek.com
Subject: Re: Just a Cherry Picking Minute
Newsgroups: rec.photo.equipment.medium-format


What you say might be true of MF glass.  And based upon what I've
v.recently seen, you're probably right.  The exception being the newer
glass for Mamiya's 6 and 7-series rangefinders.  Fuji's 690 lens also
appears to be a match for the Germans.  That's recent history.

OTOH, I've inspected literally at 1000's of negatives testing lenses
against USAF resolution charts and can say, with firm conviction, the
following:

   For Large Format work, modern (post 1970's) optics show no visible or
consistant difference in contrast or resolution between German and Japanese
optics.  In fact, the most consistantly fine LF optics have come from Fuji
for the lenses I've seen.

   Prior to that, Kodak appears to have made the finest commercially
available lenses in the world (during the 1950's).

Yes, I'm anal about optics.  And I'm pleasantly surprised by Zeiss lenses.
Though I don't yet know now different it'll be to print from them compared
to Mamiya's 7-series, and rather doubt I'll be able to tell the difference
between them in practice... which is where this whole conversation began,
right?   :-)

- Chris


Mike wrote:

> In general German optics have always had more contrast then Japanese
> optics
> due to the material used in making the glass.  Higher contrast will also
> have the effect of giving the appearance of increased resolution.
>
...


From: fotocord fotocord@yahoo.com
Subject: Re: why TIPA tests 5 lenses, not one etc. Re: cherrypicking
Newsgroups: rec.photo.equipment.medium-format
Date: Thu, 20 Jun 2002


Q.G. de Bakker wrote:

> Robert Monaghan wrote:
>
>> When I test them, against each other on the same camera and film, they
>> sometimes show differences that can be disconcerting ;-) My blind lens
>> tests have shown that one of my backup nikkors for bronica is not so
>> hot vs. the other two lenses.
>>
>> But knowing the used lenses vary is only indirect evidence that the new
>> lenses vary; they could have different histories and abuse or use too.
>> hence, my interest in lens variations in general, including new vs old..
>
> This is intriguing stuff.
> Do i remember right, and do these blind lens tests show that people often
> can't see the difference between two very different (design) lenses?
>

If I recall he said many times people wouldn't choose the lens that should
be the best lens as the best in blind tests.

I know I've had some fairly cheap lenses that performed better than bad
samples of lenses that should be much better i.e. a 3 element 210mm geronar
LF lens that was much sharper than a 8 1/2 comercial ektar that people
constantly rave about being fantastic. Same case with a good sample of a
cheapo 3 element folder lens beating a bad sample of a zeiss tessar.  I
have a 3 element schneider radionar that is better than a 4 element color
skopar, but have another radionar that is a dog.

I realise the QC/QA is better than it used to be but can't believe they no
longer have any variations in lens quality.
--
Stacey


[Ed. note: Thanks to Bob Salomon for sharing these notes on lens selection for
these notable specialty cameras, which relates to our lens variations issues...]

Subject: Re: QC/QA and camera mfgers re-test Re: cherrypicking
From: Bob bobsalomon@mindspring.com
Newsgroups: rec.photo.equipment.medium-format
Date: Fri, 21 Jun 2002

 Robert Monaghan at rmonagha@smu.edu wrote 
> And in another thread in this NG, Bob S. has again blown my mind by noting
> that the "same 90"mm lens will be tested and spec'd and selected for
> different cameras (Aero Technika vs. Technorama etc. ) based on those spec
> testing results.
Not at all.

It means that the lens used for an Aero Technika will only be used at
infinity distances so a lens that might test well at infinity only might be
rejected for a regular camera. But since it must perform at infinity only
they spec this when the lenses are ordered and they check to make sure this
is the case.

As for the Technorama they are only interested in the performance over the
6x17cm film area so lenses are speced and tested to perform best over that
area.

With both edge performance is not a factor.

All of this is tested visually using a Rodenstock Siemens Star remote
controlled projector that fill a wall in a room with Siemen Stars. As the
tester rotates the lens (with the remote) movements of the stars indicate
where in the coverage the lens performs best.

Lenses are then picked for use on specialty cameras, view cameras or are
rejected and returned.

HP Marketing Corp. 800 735-4373 US distributor for: Ansmann, Braun,
CombiPlan, DF Albums, Ergorest, Gepe, Gepe-Pro, Giottos, Heliopan, Kaiser,
Kopho, Linhof, Novoflex, Pro-Release, Rimowa, Sirostar, Tetenal Cloths and
Ink Jet Papers, VR, Vue-All archival negative, slide and print protectors,
Wista, ZTS www.hpmarketingcorp.com


From: John jsphoto@silverworks.com
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Just a Cherry Picking Minute
Date: Fri, 21 Jun 2002

 John Stafford john@stafford.net wrote:
> Cut to the chase, the seat-of-the-eye kinda evaluation. For real life
> photography, how many people can _really_ tell the difference between a good
> Zeiss lens and an average one? (Mean performance lens.)? I submit that
> NOBODY can tell the difference.
>
> There aren't any test charts hanging on gallery walls. Yet.

I can pick out slides shot with my Leica M4 from those shot with my
Nikon's just by differences in saturation and contrast.


From: fotocord fotocord@yahoo.com
Subject: Re: Just a Cherry Picking Minute
Newsgroups: rec.photo.equipment.medium-format
Date: Sat, 22 Jun 2002

> Cut to the chase, the seat-of-the-eye kinda evaluation. For real life
> photography, how many people can _really_ tell the difference between a
> good Zeiss lens and an average one? (Mean performance lens.)? I submit
> that NOBODY can tell the difference.


Have no idea on the modern versions, I do have two 521 folders both with a
75mm f3.5 tessar and one is clearly better even in a 5X7 size print. Unless
someone has several identical modern zeiss lenses, they are just guessing
that they'd never see the difference. One doesn't need to be shooting test
charts to see a lens that has lower resolution or no one would care what
kind of resolving power a len has and we'd all be using single element box
cameras!

Lens performance is a very subjective thing. I have two very different 80mm
lenses for my kiev60, the russian 80mm arsat and a late zeiss jena labeled
(made by schneider for the exacta 66? It came with a schneider lens cap
sealed new in a box...) biometar. The arsat works better close focused, is
better reversed and has a harsher look especially in the out of focus
areas. The biometar works better wide open and has a nice smooth bokeh but
doesn't work nearly as well with teleconverters. I'd say the arsat overall
looks sharper at it's best f-stops but is it a better lens? Depends on what
I'm shooting. I carry both if I have the room!
--
Stacey



From: blades@starband.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: Just a Cherry Picking Minute
Date: Sat, 22 Jun 2002


Anybody with enough experience can tell the difference.  I have no interest
in accepting a challenge because I don't have time to fool around with it,
but I could provide a projected slide demonstration that would allow you to
pick out the differences between a poor lens and a great one.  Your comment
is as it is because you simply don't have that experience.  Good shooting.

Fred
Photo Forums
http://www.photoforums

"John Stafford" john@stafford.net wrote
> Cut to the chase, the seat-of-the-eye kinda evaluation. For real life
> photography, how many people can _really_ tell the difference between a good
> Zeiss lens and an average one? (Mean performance lens.)? I submit that
> NOBODY can tell the difference.
>
> There aren't any test charts hanging on gallery walls. Yet.



From: rpn1@cornell.edu (Neuman - Ruether)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Do all lensmakers cherrypick for magazine tests?
Date: Sat, 22 Jun 2002


bobhickey@webtv.net (Bob Hickey) wrote:

>        I personally don't feel any test is valid if the lens is new.
>It's only new for one shot. Drive it around for a year and then test it,
>if it still starts.        I see some junk sold that doesn't look like
>it'll make the year. Review a few samples of 20 yr old lenses, and
>you've got my attention.
>Bob Hickey
>http://photos.yahoo.com/rollei711

Then go right on down to:

www.ferrario.com/ruether/slemn.html

(where even some 35 year old lenses
are covered...;-). In my lens evaluations,
I indicate the number of samples tried
(sometimes quite a few...) and the range
of variations found...

David Ruether
rpn1@cornell.edu
 http://www.ferrario.com/ruether
Hey, check out www.visitithaca.com too...!



Date: Fri, 21 Jun 2002
From: Christopher Perez chrisper@exgate.tek.com
Subject: Re: Just a Cherry Picking Minute
Newsgroups: rec.photo.equipment.medium-format


GREAT question!  I'd love to see someone actually demonstrate how it would
be possible to tell the difference between a 'outstanding' Ziess lens, and
an 'average' optic from the same manufacturer.  Who's brave enough to take
up this challenge?

------

At the risk of boring people yet again: I was a B&W; print tech working in
Hollywood several decades back.  Printed all formats to a wide variety of
sizes.  Even printed a good number of gallery shows.  In the final print no
one in the lab could tell the difference between lenses in a given format.
Differences were noticable between film formats.  But that was all.

Fast forward beyond the Ice Age to the present: I just checked the
resolution of a couple Zeiss lenses against a USAF resolution chart.  The
thing that is noticable is the level of contrast Zeiss has achieved that's
coupled to outstanding resolution.  I can now tell the difference between
old Mamiya TLR lenses and 25 year old Ziess T* by closely inspecting the
negs!  Never thought it'd be possible, but that's the way it is.


Translation: The negs I work with from Zeiss optics will be contrastier
than Mamiya TLR optics.  The prints might be easier to manipulate into the
desired result.  Only time will tell if this is true.  In any event I'll
have to make very minor adjustments in multicontrast print filter
selections by moving to Zeiss optics.  Many viewers think my Mamiya TLR
images were taken using the more expensive Zeiss lenses... people are so
easily fooled by name dropping.  :-)

Photography should be fun!  Not a contest no one can win!!!

Regards - Chris

John Stafford wrote:

>
> Cut to the chase, the seat-of-the-eye kinda evaluation. For real life
> photography, how many people can _really_ tell the difference between a
> good Zeiss lens and an average one? (Mean performance lens.)? I submit
> that NOBODY can tell the difference.
>
> There aren't any test charts hanging on gallery walls. Yet.




From: fotocord fotocord@yahoo.com
Subject: Re: kiev lenses.. Re: marketing
Newsgroups: rec.photo.equipment.medium-format
Date: Thu, 20 Jun 2002 


Q.G. de Bakker wrote:
> I
>> can't imagine as easily as optics can be made incorrectly, that some
>> don't slip though even from the best manufacturers. It may only be 5 or 10
>> lpmm on
>> a len that can deliver 60+ lpmm so very few people whould say "it's bad"
>> but if you compared it to a good sample, you'd see the difference.
>
> Would you?


Not in everyday use and why this isn't a big issue for most people.

When I first bought my fuji 6X9 I did some controlled tests of all the f
stops to see which were the best. f11-f16 were great and a very slight
degrade at f5.6-f8 and at f22 with noticible loss at f32. In field use it
would be very difficult to see the loss from f16 to f22 and unless you look
at the print with a loupe or are making a very large print, you'd probably
never see it. Unless you shot a negative at both settings and made large
prints of both, you'd think the shot at f22 was fantastic as there is
nothing better to compare it to.

Is this "good enough"? Sure most people would be very happy with the
results of even a "bad" 'blad lens that was at the bottom of their QC/QA
scale. Doesn't mean that another sample out there isn't better.
--
Stacey



From: Alan Browne alan.browne@videotron.ca
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Do all lensmakers cherrypick for magazine tests?
Date: Fri, 21 Jun 2002


This would depend on the integrity of the magazines, wouldn't it?
Any producer of any product will try to put their best measured unit
forward if they can.
The magazines *should* buy, borrow, rent, steal, whatever, lenses from
various sources to get a representative samples.  I doubt this is
happening.  And if a lens maker believes he got slighted, you can bet
that they would appeal for a retest with their own sample (but of course
they would not use their advertising power as leverage, that would not
be ethical would it? ).

Consumer Reports *NEVER* allows the producer to furnish a sample for
evaluation ... they go out and buy from the store like you and me.
Unfortunately, the lens niche is too narrow an appeal for them (yes they
will evaluate cameras from time to time, but it is at the consumer
level.

Cheers,
Alan

Bill Tuthill wrote:

> I'm wondering if all lens manufacturers, including the OEM ones,
> are allowed to turn in their best samples for magazine reviews.
> Or is it only Sigma and other big advertisers?
>
> Pop Photo (June 02) has revised SQF ratings for the Sigma 28-135,
> which were quite good (at 70mm better than the Minolta 24-105).
> Here is a summary:
...



From: emedia01@aol.com (Emedia 01)
Newsgroups: rec.photo.equipment.35mm
Date: 21 Jun 2002 
Subject: Re: Do all lensmakers cherrypick for magazine tests?


Having some experience withe review process of "enthusiast" magazines - not
photo related BTW - If the magazine requests a "review sample" and you don't at
the very least have your tech staff check it out before it goes to a reviewer,
then you are an idiot! I know that in some cases magazines will request to see
YOUR test results of the sample submitted, to cross check their results. I am
not saying this is the policy of the photo mags, but I do know that it goes on
in other industries.

Al


From: "ben brugman" ben@niethier.nl
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Do all lensmakers cherrypick for magazine tests?
Date: Fri, 21 Jun 2002


Sometimes magazines are allowed to come themselves and
select a lens from stock.

So the picking can be done by a magazine employee.



From: bobhickey@webtv.net (Bob Hickey)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Do all lensmakers cherrypick for magazine tests?
Date: Fri, 21 Jun 2002


        I personally don't feel any test is valid if the lens is new.
It's only new for one shot. Drive it around for a year and then test it,
if it still starts.        I see some junk sold that doesn't look like
it'll make the year. Review a few samples of 20 yr old lenses, and
you've got my attention.
Bob Hickey

http://photos.yahoo.com/rollei711


From: Ken kence@idworld.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: why not MTF charts on kilobuck lenses? Re: magazine tests
Date: Tue, 18 Jun 2002 

Robert Monaghan wrote:
>
> So I contend that this test data is fairly compelling evidence for the
> view that new lenses vary considerably in manufacturing, including zeiss
> lenses.


Bob,

Some years ago Ctein and another guy tested a bunch of high end
enlarger lens and wrote it up in one of the darkroom magazines.
They came up with a considerably variation too. IIRC, Ctein
thought poor lens centering was a big problem.

--
 Ken



From: grandguru@uboot.com (grandguru)
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: why not MTF charts on kilobuck lenses? Re: magazine tests
Date: 18 Jun 2002

Bob bobsalomon@mindspring.com wrote 
>  Robert Monaghan at
> rmonagha@smu.edu wrote
>
> > what they don't tell us is that they cherry pick the best lenses for
> > magazine tests
> Who is they?
>
> Us with Linhof, Rodenstock, Wista and formerly Rollei?
>
> Nonsense. We pick a lens from stock at random and ship it to a magazine when
> they want to do a test.
>
> We have neither the time nor the equipment to "cherry" pick a lens.

My experience of the UK market (25+ years) makes me agree with Bob
Salmon. But I would be very interested to know just how much variation
there is between different lenses.

Much more important to my mind is whether that performance
deteriorates over time and use. I know of a hire company that MTF
tests every lens on it's return from loan, thats dedication!

What percentage of photographers can enterpret an MTF chart?  



Date: Mon, 17 Jun 2002 
From: Gordon Moat moat@attglobal.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: cherrypicking 

Robert Monaghan wrote:

> well, the first reason I believe mfgers often "cherry pick" lenses for
> magazine tests is that they would be stupid not to do so, right? ;-)

Actually, I would think it would be in their best interests not to do so. It
is quite a vote of confidence to let the tester choose whatever they want. If
the test reflected really badly, the manufacturer could ask for a return of
that lens, and then get more press by discovering the fault, living up to
their warranty/guarentee, and replacing the lens with another.

There was a similar situation that happened in the car and automotive worlds.
Chevrolet was once caught providing a massaged care (Citation X11 I think).
Also, a few of the Japanese motorcycle companies have been accused of this in
the past. The biggest stink was made about Kawasaki in the early 1990s, when
one magazine pulled apart a motor and found polished intake tracks. They were
accused by the magazine of rigging the results. Kawasaki made good on that by
allowing several journalists the chance to tour the factory where that
motorcycle was made. The journalists found that for that motor on that 
motorcycle, every motor on the assembly line got polished intake tracks. The
results (and a retraction) were printed, and Kawasaki came out of it with
more respect.

So how does this relate to lenses? It could be assumed that if a lens came
direct from a manufacturer, then it might be hand picked . . . but only if
the manufacturer wanted to skew the results. The majority of cars and
motorcycles for tests often come from dealerships. These dealerships prepare
and check out the vehicles, and would not be too likely to release a bad
example, but not exactly "cherry" picking. So if lenses come from
distributors, rather than manufacturers, there is still a possibility of a
bad one getting through.

IN the end, why would it matter? If something purchased does not work as
advertised, it can be returned under warranty. If something tests incredibly
well, then someone buys an example that does not live up to the test data,
then you piss off customers. Really seems like a no win situation to take a 
risk of getting caught supplying magazines with much better equipment than a
consumer can purchase.

By the way, I have seen motorcycles used in tests later sold to the public.
They are often not in too good a shape due to immediate hard use without
break in period. It makes me wonder how well a magazine may treat a FREE
lens, and how much abuse that lens gets prior to a test.

Ciao!

Gordon Moat
Alliance Graphique Studio
http://www.allgstudio.com/gallery.html



From: "Q.G. de Bakker" qnu@worldonline.nl
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: cherrypicking
Date: Tue, 18 Jun 2002

Robert Monaghan wrote:

> well, the first reason I believe mfgers often "cherry pick" lenses for
> magazine tests is that they would be stupid not to do so, right? ;-)

No.
That would only be right if we assume they have so much variance in their
lens production that it would make sense to cherry pick. And that's the
point we need to clear, so it's not acceptable to put it forth as our
assumed starting point.

> And the mfgers aren't idiots, they know lots of buyers are reading
> reviews. I mean, over half a million subscribers (plus readers) read this
> magazine, and a lot of those read the reviews, especially before buying
> stuff; lots of us subscribe to the magazine just to get the reviews ;-)

True. 

> Now given you know that the magazine doesn't go out and buy the lenses off
> the shelf (as I would prefer, like Consumer Reports), but gets these from
> the manufacturers directly, do you think the mfgers are going to risk a
> bad report and loss of income and sales by sending out random lenses for
> magazine review, or get a cherrypicked lens and mail it out to be sure of
> a good report? Duh? ;-)

If the manufacturer is convinced that the money he's spending on QA/QC is
not a complete waste, there is absolutely no reason to cherry pick once
again. After all, that (cherry picking) is what QA/QC is doing for him: they
make sure bad lenses don't get boxed.

And yes, i have seen quite a few tests in which a lens produced unexpected
poor results, and, without fail, in the next issue of the magazine there was
an explanation of why and how a dud landed on the testers desk. ;-)

> We also have notes here and there in the magazine that comment on a
> problem lens which was identified in testing, not uncommon with prototypes
> but also with other cameras, and returned and swapped with the mfgers or
> reps. I am not against this practice, but it does imply a considerable
> degree of communication between magazine and mfgers on lens tests, yes?
>
> We also have examples (e.g., a vivitar zoom with teleconverter test) which
> was re-run with other updated lenses after problems were detected and
> mfger feedback to the magazines that the problems had been fixed etc.

(I really have to start reading the entire post before starting my reply... ;-))
Indeed. These things happen. But it is not necessarily a sign of sample
variation.

The communivcation between lens manufacturer and magazine is no more than
the manufacturer supplying the lens, next reading the magazine in which his
lens gets a bad review, and then begs of the magazine to test another sample
and print the results again, with an apology/explanation for the earlier
poor results.   

> In other words, I suspect based on these examples that if Zeiss had a
> problem with the reports on their lenses, they could have supplied other
> samples or gotten another test run and published - and they didn't. I
> doubt they missed the article in the #1 USA mass magazine, yes?

No.
Zeiss is a company of considerable repute. It would be damaging to this if
they would engage in a dispute with a third party tester over the results
they published, i.e. the quality of their product. It would be far better
marketing to ignore the published test, and let people believe that the
venerable company is above such things, have no need to feel threatened by
magazine reports. Especially since so many magazine reports so often use
every superlative they can muster when talking about Zeiss products, whether
they deserve this or not. You know how it works, it's the same mystique
surrounding brands like Hasselblad, Leitz and a few others.   

> And finally, we have the statement by the magazine lens tester that he
> could identify such cherry-picked lenses, and bought them for himself
> (wouldn't you? I would ;p).

I would too if i knew how to identify them. Of course!
How did he manage to do that?

> I believe that somebody who has tested
> thousands of lenses has a feel for when he is getting a best of the bunch
> or hand optimized or cherry picked lens for review. He has the experience
> to know what to expect from lenses in different categories and price
> classes, and the ability to see when a lens tests surprisingly well,
> right? ;-) This is one of the (few) benefits of doing all those lens
> tests, I'd bet ;-)

What is? The ability to predict that a Holga lens perhaps might not be as
good as a Schneider lens? 

It misses the point of sample variation though. He might think that because
the lens performs better than he expected too the lens might be cherry
picked. And he might still believe that because a test of the same type lens
he did earlier produced different results this is evidence of sample
variation. But it doesn't answer the question where the varition of his
results originate. It could equally well be his fault.

Which brings us back to credibility, integrity, And i don't see how, lacking
evidence, we can decide where to put the blame, Zeiss' manufacture and QA/QC
or the magazine's tester.

> In short, I believe many mfgers provided lenses that they had at least
> checked for proper performance, and in many cases "cherry picked" the best
> available lens for testing, because they would be stupid not to do so and
> risk a less than glowing magazine report. But it is icing on the cake to
> have the lens tester, with thousands of lens tests experience, admit that
> he could identify such "cherry picked" lenses and bought them when he
> could ;-) 

I still don't know how he managed to identify cherry picked lenses. It still
is an assumption.

> Now the ball is in your court. You need to explain to us why you think the
> mfgers would all refraim from "cherry picking" the test lenses they submit
> to magazines for lens testing, knowing that it would heavily impact sales
> with a bad report, and really promote sales if they got a glowing report.

Because they have an efficient QA/QC department, checking at many stages
during and after the production.

The only reason to do this is to have a consistently high quality in all
(!!!) products they eventually box and ship. So they do know that the lens
they are sending to any tester will be as good as any other lens they ship.
They do know what to expect when any of their lenses was picked for testing,
because they have done the tests on these lenses themselves already.
Obvious, isn't it? ;-)   

> You also need to explain to us why a professional lens tester, with
> thousands of published lens tests, believes the mfgers provided him with
> such cherry picked lenses now and again (which he often bought ;-); and
> how such a highly experienced lens testing professional could be misled
> as you seem to believe or unable to make such determinations etc.? ;-0)

No, no.
You need to explain why he would think that. It was proffered by you, as
accepted fact, in support of your assumption that variation indeed exists in
the products, and not in the testing. It is still a mystery though, far from
being accepted fact.

And you still need to explain how he managed to identify these cherry picked
lenses. Did the manufactuer put a sticker on it saying it was the best they
could find? Or what?

> over to you ;-)
See above.



From: John Stafford john@stafford.net
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: cherrypicking
Date: Tue, 18 Jun 2002

 Q.G. de Bakker at qnu@worldonline.nl wrote 

> Because they have an efficient QA/QC department, checking at many stages
> during and after the production.

Does any manufacturers supply a test chart with their consumer lenses? And
why shouldn't a very expensive lens be provided with exactly that?

> They do know what to expect when any of their lenses was picked for testing,
> because they have done the tests on these lenses themselves already.
> Obvious, isn't it? ;-)

Nothing is obvious in QC. Checking at each stage of manufacture is not the
same as testing the final assembly. So, why don't they do that?

Because it's too risky! It's better to let the samples fall where they may:
into the uncritical hands of the consumer!   



From: "Q.G. de Bakker" qnu@worldonline.nl
Newsgroups: rec.photo.equipment.medium-format
Subject: Re: cherrypicking
Date: Tue, 18 Jun 2002

John Stafford wrote:

> Nothing is obvious in QC. Checking at each stage of manufacture is not the
> same as testing the final assembly. So, why don't they do that?

Don't they?

> Because it's too risky! It's better to let the samples fall where they may:
> into the uncritical hands of the consumer!

That would indeed be less expensive. And it would work too. ;-) 



Subject: Re: cherrypicking
From: Bob bobsalomon@mindspring.com
Newsgroups: rec.photo.equipment.medium-format
Date: Tue, 18 Jun 2002

 Robert Monaghan at rmonagha@smu.edu wrote 

> So what do you do if you are a manufacturer or distributor, and a magazine
> calls you and says send us one of your XYZ lenses to test and publish in
> our 3/4ths million circulation magazine? Do you say, hey joe, just go grab
> one out of the box. If you pick a bad one, we'll just ask them to try
> again or we will use the stock as party favors if they don't sell?
>
> Nah, you get a box of lenses, and you get somebody to check 'em and pick
> the best of the breed. Which you either hand deliver (fewer bumps) or
> send in a big box. Why? Because you aren't crazy about having a bad lens
> test report out there when you are trying to sell these puppies, yes? ;-)

You have a vivid imagination but apparently very little practical experience
with most manufacturers / distributors in the photo business.   

While I can't speak for all I can tell you what we do with the products we
sell.

An editor of a publication, photo or otherwise - we are also in the luggage
business - and requests a product - lens, camera, accessory, flash, attachM-i,
suitcase, etc. for a review.

I write an order for that item and give it to the credit manager. She enters
it into the computer and hands it to one of the warehouse people. That
person pulls the item from stock at random and ships it.

NO HAND PICKING AT ANY TIME. P E R I O D.

We expect that our manufacturers, many who have ISO 9001 or 9000 status, are
capable of manufacturing a quality item and delivering a quality item to our
customers.   
                          
We are not in business to select grades of quality to deliver to different
categories of consumers. THEY ARE ALL ENTITLED TO THE SAME QUALITY.

That said there have been times, especially with FedX, where merchandise has
been lost or damaged in shipping. It can and does happen.

But in 30 years of sending out items for testing the number of times when a
product, picked at random, was not able to pass a test is very, very small
in our case.

HP Marketing Corp. 800 735-4373 US distributor for: Ansmann, Braun,
CombiPlan, DF Albums, Ergorest, Gepe, Gepe-Pro, Giottos, Heliopan, Kaiser,
Kopho, Linhof, Novoflex, Pro-Release, Rimowa, Sirostar, Tetenal Cloths and
Ink Jet Papers, VR, Vue-All archival negative, slide and print protectors,
Wista, ZTS www.hpmarketingcorp.com 



Subject: Re: why TIPA tests 5 lenses, not one etc. Re: cherrypicking
From: Bob bobsalomon@mindspring.com
Newsgroups: rec.photo.equipment.medium-format
Date: Wed, 19 Jun 2002

Robert Monaghan at rmonagha@smu.edu wrote 

> And isn't it true that most of the lenses returned to dealers are sent
> back out again, rather than to the factory for a checkup and retest? Yes?
No they are returned ti the factory or sold as a demo.

As for factory tests by 3rd parties they simply ensure that the lenses meet
their specs.

In the case of Linhof a specific lens has to meet different requirements
based on the camera it is destined for.

A 90mm for a Technorama has different specs then a 90 for the Metrika which
again is different then the one for the Aero Technika which has different
specks then the one sold for view cameras. All begin as the same 90.

Why test/ Every part is tested and a lens is just another of the parts. If a
part does not meet spec the product is a second.

The major failure problem in lenses at Linhof (who uses Rodenstock's QC test
and equipment) is not lower performing lenses. It is for lenses that have
dirt inside the system that can not be easily cleaned.

HP Marketing Corp. 800 735-4373 US distributor for: Ansmann, Braun,
CombiPlan, DF Albums, Ergorest, Gepe, Gepe-Pro, Giottos, Heliopan, Kaiser,
Kopho, Linhof, Novoflex, Pro-Release, Rimowa, Sirostar, Tetenal Cloths and
Ink Jet Papers, VR, Vue-All archival negative, slide and print protectors,
Wista, ZTS www.hpmarketingcorp.com 



From: fotocord fotocord@yahoo.com
Subject: Re: why TIPA tests 5 lenses, not one etc. Re: cherrypicking
Newsgroups: rec.photo.equipment.medium-format
Date: Wed, 19 Jun 2002 

Bob wrote:

>  Robert Monaghan at
> rmonagha@smu.edu wrote 
>
>> In short, I think it makes a lot of sense for mfgers to cherry pick
>> lenses, given that so many folks have used these lens tests in magazines
>> (or photodo more recently online) in their lens selection...
> It would be suicide.
>
> How long do you think it would take for consumers to realize that their
> lenses were not performing as expected?     

How many people actually test a lens before they use it? Especially a new
one? I only recently started doing a quick formal test of lenses as I buy
them and have been shocked at the variations I've found in my old gear I
-was- happy with. Ignorance is bliss.

--
Stacey  



From: rpn1@cornell.edu (Neuman - Ruether)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How To Test My Lenses
Date: Sat, 20 Jul 2002

"Meryl Arbing"
 wrote:

>Yes, I guess that is my point...testing is only meaningful relative to other
>tests...to answer the question..Does this particular lens perform better
>than another lens in a specific test?

Yes - that is why I test. The results tell me if the lens
is substandard or not (a simple comparison of the two short
edges of subsequent frames, shot at the same wide stop and
focus, of the same distant detailed subject, tells me if
the lens is optically well-aligned, and if so, it is likely
also to be about as sharp as another good sample of the
same lens), and how it compares with other similar-FL
lenses... In the process of simple alignment, wide vs.
middle aperture sharpness, and corner vs. center sharpness
testing on film, you can also get info on distortion, flare,
etc. by using simple VF tests.
(See: www.ferrario.com/ruether/slemn.html.)

>You may choose to test sharpness or
>contrast or flare or distortion and it will be up to you if you decide to
>keep a lens that, for example, is really sharp but is prone to severe flare.
>Perhaps you don't even test for flare or distortion.
>Even if you only have some arbitrary personal standard that you hope the
>lens lives up to, I still can't see what you do if the lens fails to meet
>those standards? Take it back?

Of course. Why keep a lens you know is sub-standard, either
due to misalignment, or type design?

>Then we return to the "cherry picking"
>approach of trying (and testing) several lenses before deciding on a
>specific individual lens.


It is simpler than this - the defective samples are returned
until one is found that is sufficiently well-aligned for
your standards or purposes (this may take one or two
subsequent tries); the poor designs one gives up on
immediately, unless they are Ok for your purposes...

>Of course, if everybody did that, there would
>eventually reach a point when ALL the 'good' lenses would be picked and the
>stores would be left with the defective stock that they couldn't get rid of.
>What if that has already happened and all you could hope to buy are the
>'substandard' ones...what do you do then? Keep the defective lens? Wait
>until a new shipment comes in? (Remember to keep a list of the serial
>numbers you have already rejected so you don't waste time re-testing!!)

The above, in practice, doesn't happen...

>In all, I just think that you are welcome to try out any lens you buy...but
>don't think that it does anymore than give you some reassurance.

Of course! That is the point...;-)

>Of course,
>you can avoind the whole problem of buyer insecurity by avoiding the cheap
>3rd party offereings and buying the best that you can afford. I doubt that
>Leica owners worry too much about getting a lousy lens...Canon 'L' lenses
>are probably pretty safe too...same with Zeiss.

This is not true - I have seen a bunch of defective Canon
and Leitz lenses in my time (and a lower percentage of
defective Nikkors...;-). The faith that higher price
reduces the defect rate is misplaced...

>If you think that you are going to get the performance of a $1000 lens for
>$100, you are likely to be disappointed more often than surprised.

This is true - but it does happen...


David Ruether
 rpn1@cornell.edu
 http://www.ferrario.com/ruether

From: rpn1@cornell.edu (Neuman - Ruether)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How To Test My Lenses
Date: Sat, 20 Jul 2002

...(quoting above post)

 "Meryl Arbing" marbing@sympatico.ca wrote:
>
>What percentage of defective Zeiss or Leica or Nikkor lenses would be
>reasonable to expect? Are we talking... 25% ...or 2%... or .02%?  I would
>also expect that the percentage of defective lenses would logically be
>higher in Used examples rather than New ones.  So, if a person bought...for
>example...20 used lenses it would be highly unlikely that there would be no
>defects found?

For Nikkors, the range of variation is indicated by
the range in "evaluation numbers" in the list - and if
a particular lens has an unusually high defect rate,
that is covered in the notes with the listing (only a
few Nikkors have a high defect rate [the 35-200 and
35-105MF are about it, though several others show some
variability). 

With Canon, another photographer in town
had to buy four 20mm f2.8's to get a good one (the others
were quite soft), and his 28mm f2 MF had very noticeable
field curvature (likely a design fault). The Cornell
paper was given a 17mm f4 (Canon marketing...), but it
was dreadful (likely defective...). A 24mm f1.4 Canon AF
appears in my Nikon list (!) since it was so good.

In Leitz, several lenses for the CL I tried were soft at
wide stops, with one 90 f4 terrible, and one really
excellent. Of 21mm's with another Leica user, the first
was poor, the second good. An excellent 35mm f1.4 Leitz
appears in my Nikon list... 

For Zeiss, I have run across
few defects in the lenses of the several Rollei 3.5E and
Fs that I have owned, though all showed considerable
field-curvature problems (a design problem). A 50mm
Zeiss for the Rollei 66 was soft (probably defective).
Samples and experience too small to quantify (except
for Nikon), but.....

As far as new vs. used, I have seen no differences - the
lenses tend not to become defective with use except for
obvious wear and damage, or oil on the diaphragm. And,
at least with Nikon, the defect rate is very low (new or
used), but the slight misalignment rate is fairly high
(especially in short zooms and CRC retrofocus
wide-angles - particularly the widest), with the same
expectations for new or used... 

In other words, if you
are particular, test; if not, don't (chances are you
will be OK, except for the few obviously defective lenses,
though you may not have an "optimized" set of lenses
[and may not notice the difference...;-]).

BTW, some specific newer Nikkors are reviewed separately
from www.ferrario.com/ruether/slemn.html at:

www.ferrario.com/ruether/articles.html, and these include
comments about alignment in the samples tried...


David Ruether
rpn1@cornell.edu
 http://www.ferrario.com/ruether
Hey, check out www.visitithaca.com too...!



Date: Sun, 21 Jul 2002
From: Chris Quayle lightwork@aerosys.co.uk
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How To Test My Lenses

...(quoting above post)


Perhaps the basic optical designs of lenses are not be too different in
terms of quality between manufacturers, since they all have access to
the same sort of cad packages for lens design. What is different is the
spread of quality, since stuff like selective fitting and testing is
labour intensive, can't easily be automated and is a fixed manufacturing
cost.

The spread of quality will be greater on a cheap lens for this reason
and because the manufacturing tolerances are likely more relaxed in any
case. Plastic lens housings don't help either, since plastics doesn't
have the dimensional stability of metals. You may get a good example,
but because of the wider spead, you are perhaps just a likely to get a
substandard example. Buy from a quality manufacterer doesn't guarantee a
perfect example either, because there's still a spread, but this is
likely to be much narrower as the target limits etc will be much
tighter. I guess this is what you pay extra for...

Chris


From: rpn1@cornell.edu (Neuman - Ruether)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: How To Test My Lenses
Date: Mon, 22 Jul 2002

...(above post requoted)

>Perhaps the basic optical designs of lenses are not be too different in
>terms of quality between manufacturers, since they all have access to
>the same sort of cad packages for lens design.

I have not found this to be true, though it is logical...
Even within a line of lenses, Nikon (for instance) has
produced several quite different optical-design versions
of the 35mm f2.8, and offers even now two completely
different 35mm f2 designs... There are a lot of other
examples in the Nikon line of various designs being offered
for the same FL/speed lens, often concurrently. Between
brands, the differences in designs and performance are
also obvious, and not all 28mm f2.8s, for instance, are
alike, common as this FL and speed is (Nikon has offered
several, ranging from mediocre to excellent, and the best
Nikkor version of the 28mm is generally better than similar
lenses offered by others, even the "fancy-'spensive"
makers...


>What is different is the
>spread of quality, since stuff like selective fitting and testing is
>labour intensive, can't easily be automated and is a fixed manufacturing
>cost.

I'm not sure I agree with this, either...
An indication: most Nikkor zooms shift focus slightly
when zoomed; most "cheap-brand" zooms do not...

>The spread of quality will be greater on a cheap lens for this reason
>and because the manufacturing tolerances are likely more relaxed in any
>case. Plastic lens housings don't help either, since plastics doesn't
>have the dimensional stability of metals.

I agree with the first part, but not the second - plastic
can be a good material for holding tolerances...


>You may get a good example,
>but because of the wider spead, you are perhaps just a likely to get a
>substandard example. Buy from a quality manufacterer doesn't guarantee a
>perfect example either, because there's still a spread, but this is
>likely to be much narrower as the target limits etc will be much
>tighter. I guess this is what you pay extra for...
>
>Chris

With Nikon, at least, the spread in quality is more related
to lens type, and to a very specific few "problem" lenses,
than to an overall manufacturing permitted tolerance...
I suspect the same is true for "off-brand" manufacturers,
with some lenses in their lines being reasonably consistent,

some not...

David Ruether
 rpn1@cornell.edu
 http://www.ferrario.com/ruether


From: smitbret@hotmail.com (Brett)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Sigma SLRs opinion????????
Date: 20 Jul 2002 


Of course it was from the shared experience of others, that's how must
of this stuff spreads.  Sigma has some crap, but so does Canon.  White
lenses are much better than the non-L lenses, likewise the the EX
lenses tend to be much better than the non-EX lenses.  

If you want to,
I an discuss the Canon 28-135 that had to be replaced twice in two
weeks or the 50mm f1.8 II that did "fall apart in my hands" (a phrase
the anti-Sigmatists use frequently) although my experience is first
person, unlike most of the people who had a friend that heard of
someone that it happened to.  

The only trouble I've ever had with a
Sigma lens was a Quantaray branded 24mm f2.8.  The AF stopped working
after I dropped 20 feet while taking pictures of rock climbers.  It
was only about three weeks old and the store replaced it, never had a
problem since.  

I'll concede the favorable opinions of the L series
70-200 f2.8 and I have yet to try the IS, but to say that the Sigma is
not in the same league as the L is still ignorant.  If Canon is the
Yankees, then maybe Sigma is the Expos, but as we can see this year,
even the Expos can compete on occasion.

-Brett
...


From rollei mailing list:
Date: Sun, 28 Jul 2002 
From: Eric Goldstein egoldste@earthlink.net
Subject: [Rollei] Re: 2.8C Xenotar

> Sam
>
> My spelling error, it is bokeh.
>
> Sacsu@aol.com wrote:
>
>> Fellow Rugers, I still learning about Rollei TLR's and I notice the
>> word "brokeh" used when referring to the quality/characterisitcs of a
>> lens. What's brokeh?


Dale -

I thought you were coining a pretty useful new term... I used more than a
few lenses whose bokeh was brokeh... {g}

I also think you've gotten your arms around a very important variable when
shooting with the classics... Namely how variable the lenses are from camera
to camera. People tend to concentrate more on which model is best... You
really nailed it when you spoke of shooting ten different cameras with ten
different lenses and finding which particular optic is the nicest...


Eric Goldstein


From: rpn1@cornell.edu (Neuman - Ruether)
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Newby Question - Lenses
Date: Sun, 28 Jul 2002

eos10fan@hotmail.com (dan) wrote:

>I am an EOS user
>so you should cconfirm this with some true Nikonians ;)
>
>I have read that the
>Nikon 24-120mm f/3.5-5.6D AF Zoom-Nikkor
>suffers from 'sample variation'.
>This means that some examples of the lens are good,
>and some are not so good.
>
>I have also read that the
>Nikon 28-105mm f/3.5-4.5D AF Zoom-Nikkor
>is a very good lens,
>especially considering it's price.
>
>Any true Nikonians out there that can confirm this?


ALL lenses suffer from "sample variation" - what
is important, for purchasing purposes, is the
likely range of variation. I have checked four
24-120 Nikkors and found remarkably little sample
variation (FAR less than with some Canon lenses,
even non-zooms...;-) I have heard of "so-so" samples
of the 24-120, but I have not seen them (I always
recommend quick film checks for lenses, both new
and used, to spot defective ones while they cam
still be returned/exchanged, though...) The 28-105
Nikkor is a good lens, but the 24-120 easily
outperforms it, I think...

(See www.ferrario.com/ruether/slemn.html,
www.ferrario.com/ruether/articles.html#24-120,
and www.ferrario.com/ruether/articles.html#28-105.)

David Ruether
 rpn1@cornell.edu
 http://www.ferrario.com/ruether



From: fotocord fotocord@yahoo.com
Subject: Re: fast lenses, warning: zeiss-philes don't read Re: Rollei or Hassy
Newsgroups: rec.photo.equipment.medium-format
Date: Sun, 16 Jun 2002


Q.G. de Bakker wrote:

> Robert Monaghan wrote:
>
>> ========
>> WARNING: Do Not Read this if you are a Zeiss Lens Fan - you've been
> warned!
>>
>> Another issue is that zeiss lenses vary, and sometimes by quite a bit. I
>> have some test data from MP on the older hassy lenses, and was surprised
>> to see how large the variations could be; in some cases, with the same
>> test standards, the hassy lenses were rated as "excellent" in one test
>> and sample, as as just "acceptable" on the other test and lens sample - a
>> three step difference! There were a number of two and one step
>> differences in the same lens tests of these zeiss lenses.
>
> May we assume you have established that this variation was indeed in
> Zeiss' manufacture and QA, and not in MP's testing?
> If so, may i ask how you managed to do that?
> ;-)

I don't doubt that zeiss has a closer tolerance and checks their optics but
anything manufactured by humans has production tolerances and as such there
is going to be sample to sample variations. Probably less so on a 'blad
lens compared to a pentacon one but it's still there. I had to buy and sell
three 50mm flektogons to get an outstanding sample. One was just plain bad
and the other was pretty good (I would have been happy) but had already bid
on another that I ended up winning and it was even better!

Another  example, I worked for a honda car dealer which are very precisely
made automobiles and was still amazed at how good some of them performed
compared to other identical cars.

I think the point Bob is trying to make is, if posible, you should try to
test as many samples as you can and then pick the best, sell the rest. For
some camera systems, the cost of the optics makes this almost imposible 
do so one must then hope they are lucky enough to get a good one.
--
Stacey


Subject: Re: why not MTF charts on kilobuck lenses? Re: magazine tests
From: Bob bobsalomon@mindspring.com
Newsgroups: rec.photo.equipment.medium-format
Date: Sun, 16 Jun 2002

Robert Monaghan at rmonagha@smu.edu wrote
> what they don't tell us is that they cherry pick the best lenses for
> magazine tests

Who is they?

Us with Linhof, Rodenstock, Wista and formerly Rollei?

Nonsense. We pick a lens from stock at random and ship it to a magazine when
they want to do a test.

We have neither the time nor the equipment to "cherry" pick a lens.


Additionally why would we want a test report to be so good that the typical
consumer can't achieve that result?

We expect our customers to receive a lens that performs as well as or better
then the test results.

Perhaps some manufacturer or distributor is stupid enough to cherry pick
today but the editors we work with at photo magazines do not ask for picked
lenses.

As for Zeiss and MTF tests they were readily available from us when we
distributed Rollei and their Zeiss lenses as did Hasselblad with their Zeiss
lenses (both sets of tests were identical so they would have appeared to
have been Zeiss tests rather then the camera maker's tests). Those curves
should also be available today without a lot of searching. Just ask
Hasselblad or Rollei for them.

And your link does not open.
HP Marketing Corp. 800 735-4373 US distributor for: Ansmann, Braun,
CombiPlan, DF Albums, Ergorest, Gepe, Gepe-Pro, Giottos, Heliopan, Kaiser,
Kopho, Linhof, Novoflex, Pro-Release, Rimowa, Sirostar, Tetenal Cloths and
Ink Jet Papers, VR, Vue-All archival negative, slide and print protectors,
Wista, ZTS www.hpmarketingcorp.com


Subject: Re: why not MTF charts on kilobuck lenses? Re: magazine tests
From: Bob bobsalomon@mindspring.com
Newsgroups: rec.photo.equipment.medium-format
Date: Sun, 16 Jun 2002 

 Robert Monaghan at rmonagha@smu.edu wrote
> Hi again, Bob!
>
> thanks for the info on your procedures. Since you don't have the
> equipment to test or "cherrypick" any lenses, any test lens is only a
> single point sample, yes? In other words, you don't know if the lens you
> sent off is the worst lens of the bunch, an average lens, or maybe the
> best one of its type in the USA. The assumption is that it is average,
> and that the range of variation is very tiny and that mfger Q/C sets a
> floor for performance below which lenses are not allowed to fall, yes?
> But those are the assumptions I am calling into question here...
>
> Obviously, I prefer a random lens for testing than a cherry picked one,
> given that I aren't likely as a buyer to get a cherry picked performer
> in a random pick out of a box of lenses ;-) So I prefer your methods ;-)
>
> Similarly, the Zeiss lens charts are for a "sample" lens as produced (I'm
> told). We don't know if the reported lens MTF test is an average lens, a
> best of the batch lens, or what. Right?
>
> If we had "guard bands" (say at 1 and 2 sigma std. deviation units) on
> these curves, we would know how far they might be likely to vary in the
> box. Obviously, Zeiss has this Q/C data. As buyers, we don't.
>
> So we don't know even with these MTF graphs how likely we are to get this
> kind of performance as shown, or how widely our lens is likely to vary
> from this performance level, or even what the minimum performance level
> might be that we are guaranteed by zeiss Q/C, yes?
>
> So I don't know what the range of variation really is. But there is
> clearly some range of variation in these lenses and their manufacture. The
> few independent expert test reports we have of new in the box zeiss lenses
> suggest that there is some variation, and in some tests an amazing
> amount of variation is reported for lenses of this caliber, IMHO...
>
> For 2 new 80mm f/2.8 hassy lenses, one sample had 12 "excellent" and 2
> "very good" ratings, while another had only 6 "exc" and 4 "very good"
> ratings, but 4 ratings of "acceptable". As a buyer, I'd like the one that
> had no "acceptable" ratings and 12 excellents instead, yes?  ;-) ;-)
>
> This is more like the kind of variation you might expect between different
> brands (say zeiss vs. Bronica) than within two lenses in the same high $$
> mfger like Zeiss.
>
> I have also done over a dozen comparisons on 60 other normal lenses for
> 35mm (see charts at http://medfmt.8k.com/third/variations.html ) which
> echo this range of variability, but in high $$ brand 35mm lenses too.
>
> So the evidence I have so far is that even new zeiss lenses also vary, and
> per the published MP lens tests, by quite a bit. We don't know if the
> zeiss MTF charts are for a statistically "average" lens, (or a cherry
> picked one ;-) and how much their performance varys, and what the minimum
> performance might be to pass Q/C.
>
> But what data I have seen on multiple lens sample tests suggests that
> a buyer might do well to "cherry pick" the best lens they can get, if
> the range of variation is as great as reported by Modern's tests, yes?
>
> And for used lenses, as Stacey (I think?) noted, I advocate lens testing
> and comparisons, and building a backup kit. One of my hassy zeiss 80mm
> f/2.8 is better than the other wide open by a modest amount. One of my
> bronica 75mm f/2.8 lenses from a kit buy is only so-so, while the other
> two are very good. My two kowa 85mm lenses are very similar though. I
> have 3 Kiron/vivitar zooms of the same type, one is so-so, one stellar,
> the other very good. In other words, I have found in my own real world
> shooting and testing that lots of lenses vary by a noticeable amount,
> including the big names and high dollar optics ;-) So I'm not surprised
> that the new lenses also vary out of the box - and by a noticeable amount.
>
> What does surprise me is how much all these lenses vary - while the
> industry is based on the assumption that these lenses are as alike as
> peas in a pod, so just pick one, yes? ;-) People reject lenses that score
> 0.1 below another lens' photodo score, when the reality may be that the
> lenses vary by 0.3 or more units, or more variation within a brand than
> the average between brands may differ ;-) etc.
>
> Similarly, if lenses vary by the amounts shown by MP's tests (cited at
> third/variations.html) then what is the use of magazine testing? Esp. if
> many of those lenses (not all, as Bob S. has pointed out) were cherry
> picked optics for magazine tests to garner higher scores than the average
> production lens?
>
> in any case, it is a fun subject and appealing to the engineer in me ;-)
>
> regards bobm

I can only speak of real world experience.

In all the years that I have been in the industry I could not count on one
hand the number of lenses that we have sold that have been returned due to
not performing as the owner expected. This includes the lenses we sold to
NASA for the 6008 system, to Edwards for Photogrammetric work, to SI or the
NY Times or any other use.

In 20+ years I can not come up with as many as 5 instances where the lens
did not meet the quality standard of the user.

And any that did not were replaced. Even one enlarging lens that was heavily
used for several years in a lab (300mm Rodagon G).

I can not speak for Hasselblad Zeiss lenses, I never represented them, but I
did represent Rollei and Bronica as well as Fuji over the years and we did
not have the problems with lens variations that you seem to have or seem to
fear.

Of course I am not an engineer so I can only speak from 1st hand experience.
Not from theory.

HP Marketing Corp. 800 735-4373 US distributor for: Ansmann, Braun,
CombiPlan, DF Albums, Ergorest, Gepe, Gepe-Pro, Giottos, Heliopan, Kaiser,
Kopho, Linhof, Novoflex, Pro-Release, Rimowa, Sirostar, Tetenal Cloths and
Ink Jet Papers, VR, Vue-All archival negative, slide and print protectors,
Wista, ZTS www.hpmarketingcorp.com


From manual SLR mailing list:
Date: Mon, 17 Jun 2002 
From: Stephen Gandy LeicaNikon@Earthlink.net
Subject: Re: [SLRMan] lens variations in Pntx, Min, Lei, nik, oly, can, cntx... ;-)

Hi Bob,

yes, it is a very interesting subject.   it's odd that so few people really test
their own lenses, while just about everyone pays attention the test of a lens
that they don't own,  that may in fact perform well above or well below the
lenses of that type.  the matter is made worse by very different test criteria.

Robert Monaghan wrote:

> Hi Stephen!
>
> actually, I used the test data I had access to; the 4 kowa lenses were
> one of the few times four lenses from the same batch were tested and
> published at the same time.
>
> I use it because it refutes the whole idea of one sample mag. tests if the
> lenses vary, why bother to buy the magazines to read the tests if they
> only relate to one lens sample, and your lens will probably vary, and by
> quite a bit, from the one reported on etc. Ditto picking lenses based on 0.3
> units of difference in photodo tests, as it looks like variations within
> a batch of lenses is rather larger than that, even on big name versions,
> as careful reading of the reported tests at third/variations.html will show
>
> even worse, did you see http://medfmt.8k.com/third/variations.html#cherrypick
> this is a quote by one of the major lens testers on how mfgers often
> provide a selected (Cherry picked) lens for magazine testing, so your
> chances of getting similar performance out of the box is much less,
> right? ;-)

yes, I am well aware of that.  with so much money riding on test reports, lens
makers would be fools to do otherwise.  With jobs and profits riding on sales, it
only makes sense for photo executive to do everythign possible to make sure lens
are good a possible.  On the other side, what magazine publisher wants to lose
valuable ad dollars by criticizing a lens maker too much?

That is why  I always trusted the old Modern Photography Test Reports.  while
Modern would accept samples on a preview, tests were only done on randomly
purchased cameras and lenses.  I believe Popular also follows this policy, but I
am not sure.

> however, the web page http://medfmt.8k.com/third/variations.html DOES
> have examples of variations in hasselblad new zeiss lens test data,
> which vary dramatically in some scores (e.g., 2 and even 3 step differences
> in ratings, some lenses "excellent" at one f/stop, the other lens tested
> by the same charts and standards and people was rated "acceptable" etc.
> So the idea that zeiss lenses are immune to sample variation is also wrong
> based on this test data etc.

not all variations may come out of the factory.

suppose you have 5 sea containers going to distributors all over the globe. all
will get different levels of vibrations over time, mostly as a result of how long
the trip is, and how many times they are loaded/unloaded.  but suppose one
container is dropped on the docks.  Do you think those lenses  go back to the
factory  -- or are sold to consumers ?

> I also have examples of multiple tests by Contax, Minolta, Canon, Nikon,
> Chinon, Fujinon, Leica (R vs. SL and M), Pentax, spiratone, Konica,
> Koni Omega, Mamiya, Olympus, and Miranda. Phew! In some cases I provide
> charts of more than one lens (e.g., MC/MD and AF minoltas) type or speed
> (f/1.4 and f/1.8 and series E nikkors..). You need to look at the charts,
> which make it hard to say that I haven't made the case, I think? ;-) ;-)
>
> If anything, I'm fairly surprised by how many examples of major variations
> I was able to document in a few days of research and updates. Moreover, my
> own blind lens testing of multiple lenses of the same type that I own has
> shown that such lenses also vary, and by quite a bit (mf/blindresults.html)
> For example, two of the top 3 med fmt lenses picked by observers have been
> bronica nikkors, but my third bronica nikkor in the same test got low marks.
>
> so the importance of testing the lens you have on your camera, or before you
> finalize your purchase, is re-emphasized by this data and my own tests...
>
> regards bobm



From: fotocord fotocord@yahoo.com
Subject: Re: Fuzzy Hasselblad Distagon 50mm
Newsgroups: rec.photo.equipment.medium-format
Date: Mon, 24 Jun 2002


Anthony Maw wrote:

> Hi all:  I have T* Hasselblad 50 mm Distagon that I bought used some years
> ago.
> I shoot mostly black and white landscapes and I like to make up to 20 x 24
> prints.
>
> I  have always noticed that all four extreme corners of virtually all my
> negatives are fuzzy.
>
> So I'd like to really know if I got a dud lens or is the Hasselblad 50C
> Distagon just a piece of crap?


Sounds like a bad sample. The 50mm zeiss flektogons (east german lens)I've
tried have been:

One was soft in the corners (older single coated)
One was sharp across the frame but didn't have great contrast (again single
coated)
Last one was fantastic (latest multicoated)

My quess is you've got a dud and maybe rent another sample to see what
you're missing?
--
Stacey


Date: Mon, 24 Jun 2002
From: Christopher Perez chrisper@exgate.tek.com
Subject: Re: Just a Cherry Picking Minute
Newsgroups: rec.photo.equipment.medium-format


Everyone except me?  I know, it must be that Ducati thang...   :-)

Can I tell the difference on print?  No.  And that's partly the point.

Unless a lens is unusually bad, a print can be manipulated.  So it's
v.difficult to say with any certainty which print started it's life being
taken with which lens.  As you point out, there are many factors in print
creation.  Original camera optics are only one.  Enlarger optics and
alignment can be another.

So the gent who said they could tell the difference between a 'mediocre'
Zeiss and an 'outstanding' Zeiss lens could help us all by more fully
explaining what he means.


John Stafford wrote:

> Christopher Perez at
> chrisper@exgate.tek.com wrote:
>
>> [...] I can now tell the difference between
>> old Mamiya TLR lenses and 25 year old Ziess T* by closely inspecting the
>> negs!  [...]
>
> Uh, I meant everyone _except_ Chris Perez. Yea. That's what I meant. :)
>
> But to add a small point: once you enlarge your negative, Chris, you've
> introduced another degredation in resolution and accutance via the
> enlarging lens. So... could you tell from the print?


From manual minolta mailing list:
Date: Thu, 12 Sep 2002 
From: "Dave Saalsaa" SaalsD@cni-usa.com
Subject: Re: New versus Old

Are you asking if you should buy an newer MD lens of the same model that you
own in an MC version?  I am the wrong guy to ask about this.  I own
multiples of the "same"lens.   I test one against the other to pick the
better one.  Why you ask?  Because, as my wife would tell you, I'm insane.
;-)   My reason is a bit different that hers, however.  It's all in the name
of research.  ;-)   I think you should do whatever makes you happy.

Dave Saalsaa


From: Martin Trautmann t-use@gmx.net
Newsgroups: rec.photo.equipment.35mm
Subject: Re: Pentax lenses and Photodo tests
Date: 30 Sep 2002


 Joshua Hakin wrote:
>  Well Kostas, I'll tell you what someone else told me: "I've seen so much
>  obviously erroneous data on Photodo (they rate the Pentax
>  F50/1.4 *much* better than the FA50/1.4 - but the lenses are identical except
>  for cosmetics!) that I don't pay attention to them at all any more."
>  I have no proof to back this up, it just fits with your question.

How much do you give for sample variations?

Although I doubt that the results of these tests are very exact and
reproducable, I guess that actual samples may differ here significantly.

Kind regards
Martin



End of Page