GaryLH
Veteran
Lensrental has started doing something that I was always curious about..lens variance.
http://www.lensrentals.com/blog/2015/06/measuring-lens-variance
This is only an initial presentation so far w/ some samples from the 24mm group. Roger is going to publish more later.
Given manufacturing tolerances and today's QA controls, I have always wondered if paying more from certain lens makers provided better lens variance (less differences) or is really u pay for the name. My personal expectations had been that brands like Leica, Zeiss, Contax, etc..there would be less variance for a given lens because of tighter manufacturing tolerances and better QA control where the higher cost is due to more of the product getting discarded or sent back for rework due to the tighter tolerance values used in the QA inspection.
It will be interesting to c Rogers final result.
Gary
http://www.lensrentals.com/blog/2015/06/measuring-lens-variance
This is only an initial presentation so far w/ some samples from the 24mm group. Roger is going to publish more later.
Given manufacturing tolerances and today's QA controls, I have always wondered if paying more from certain lens makers provided better lens variance (less differences) or is really u pay for the name. My personal expectations had been that brands like Leica, Zeiss, Contax, etc..there would be less variance for a given lens because of tighter manufacturing tolerances and better QA control where the higher cost is due to more of the product getting discarded or sent back for rework due to the tighter tolerance values used in the QA inspection.
It will be interesting to c Rogers final result.
Gary
Last edited:
rscheffler
Well-known
I hope they test Leica and Zeiss lenses in addition to the usual suspects.
I wonder though if they have enough inventory of some of these more specialized lens, perhaps more so the case for Leica, to be able to hit the requirement of 10 tested lens copies? But based on his graph illustrating how average variation decreases as the number of tested copies increases, it seems 6-8 should also be OK.
I'm especially curious how Leica lenses will hold up. Their lens design philosophy apparently is to make each element do as much work in the smallest size possible, which apparently requires tighter tolerances during manufacture and assembly. Therefore slight deviation from an ideal and likely narrow range may have a more dramatic affect on lens performance...
I wonder though if they have enough inventory of some of these more specialized lens, perhaps more so the case for Leica, to be able to hit the requirement of 10 tested lens copies? But based on his graph illustrating how average variation decreases as the number of tested copies increases, it seems 6-8 should also be OK.
I'm especially curious how Leica lenses will hold up. Their lens design philosophy apparently is to make each element do as much work in the smallest size possible, which apparently requires tighter tolerances during manufacture and assembly. Therefore slight deviation from an ideal and likely narrow range may have a more dramatic affect on lens performance...
back alley
IMAGES
i saw that earlier today...very interested in the results as people have such varied experiences with the same lens quite often.
GaryLH
Veteran
I think part of the difference is a combination of expectations vs experience vs how they are evaluating the end product as much as the actual lens variance itself.
Gary
Gary
back alley
IMAGES
yes, those expectations can be killer!
David Hughes
David Hughes
Hi,
" ... it seems 6-8 should also be OK... "
I would have thought that a sample size of about 200 would be needed to get the variation, as a percentage, to single figures, like 5% or so. Perhaps my memory is playing tricks...
As for other lenses, most of the lenses mentioned on RFF are very second-hand and usage and abusage can affect them a lot.
Regards, David
" ... it seems 6-8 should also be OK... "
I would have thought that a sample size of about 200 would be needed to get the variation, as a percentage, to single figures, like 5% or so. Perhaps my memory is playing tricks...
As for other lenses, most of the lenses mentioned on RFF are very second-hand and usage and abusage can affect them a lot.
Regards, David
Brian Atherton
Well-known
Back in the 70s, RG Lewis of London used to shoot and provide a set of negatives of a resolution chart taken with each lens they sold, giving the resolution figures. If there was more than one lens, one had the choice of which to buy.
I remember when I bought my Pentax SL and 50mm Takumar that there were four or five examples to choose from, though frankly, supplying resolution negatives was probably more hype than real use photographically, but at the very least it was nice to know I had purchased a lens that had been checked optically.
I remember when I bought my Pentax SL and 50mm Takumar that there were four or five examples to choose from, though frankly, supplying resolution negatives was probably more hype than real use photographically, but at the very least it was nice to know I had purchased a lens that had been checked optically.
GaryLH
Veteran
Reality check of a business w/ limited monetary resources, I would suspect.
Though I agree that out there are enough questionable used lenses, I remember also reading about pro photographers that were allowed to cherry pick a lens from a camera store stock to get the best lens on blogs in the past...
Gary
sevo
Fokutorendaburando
It is quite likely, from a marketing perspective, that shops make that kind of offer. But it is just a colourfully phrased return policy - regular pros have neither the means nor the time to do test runs. And those with studios (or rental businesses) big enough to be technically able to cherry pick a lens (by virtue of employing a in-house optician with test bench) probably do not deal with shops but with the manufacturers or importers themselves.
daveleo
what?
This is a very interesting project and I hope they follow through on it.
Rigorous statistics (if I recall correctly) say that a sample of 25 is needed to accurately calculate the std dev of a Normally distributed characteristic. But I think the sample of 6-10 will be fine for non-scientific purposes.
Publicizing this kind of data is long overdue. Technically (geekly) very fun stuff too.
Rigorous statistics (if I recall correctly) say that a sample of 25 is needed to accurately calculate the std dev of a Normally distributed characteristic. But I think the sample of 6-10 will be fine for non-scientific purposes.
Publicizing this kind of data is long overdue. Technically (geekly) very fun stuff too.
GaryLH
Veteran
I agree.. I doubt the average pro has the time to test and cherry pick a lens. But I remember reading blogs from two photographers that mentioned they were given the opportunity to cherry pick and one said he went though a shipment worth of a particular zoom lens before he found one that match his expectations.
Lensrentals seems to be an exception to the rental business since they not only have a test bench, they have a custom one. I don't think they use it to cherry pick, more than check out lenses and bring into reasonable specs or return them to the manufacturer if they cannot bring into their own expectations.
Gary