Digital Photography Review launches lens reviews

[m_barbu1]

Given how meticulously they review camera bodies, and the void of a similarly

well-done lens review site, I have very high hopes for this new venture: <a

href="http://www.dpreview.com/news/0801/08012913dpreviewlensreviews.asp">http://www.dpreview.com/news/0801/08012913dpreviewlensreviews.asp</a>

[CharlesBecker-Toronto]

that's great-I like their camera reviews and have been waiting for them to do lenses as well;

thanks for posting this info! cb

[bobatkins]

They're still only testing one sample (I assume), and that's always been the downfall of any lens testing scheme. You might get a great sample, you might get a lemon, you just can't tell. Given the number of complaints about "bad" samples that get reported in the forums, QC (consistancy) is almost as important as (possibly more so than) the test numbers on any particular sample of a lens.

[m_barbu1]

While that's certainly true, how do you counter that? If they test 5, they could all be from the same "bad batch".

[bobatkins]

Unless you buy them all from the same store and they have consecutive serial numbers, you'd certainly get a statistically better result by testing 5 than you would by testing one.

 

There's also no real evidence of "bad batches", just sporadic lenses that have problems.

 

It's a problem of course. While testing 10 lenses from 10 stores located in 10 different parts of the country would give you a more statistically valid measure of performance, it's not worth anyone's time or money to do that. Basically all lens reviews on commercial websites are designed to do is draw in readers and get them to buy something.

 

Another thing I'd like to see is repeat tests on the same lens, made 6 months apart. Would they yield exactly the same numbers?

 

I'm saying this because I've done a lot of lens testing myself, and I've also done of lot of scientific optical measurements as part of my job. I know how hard it is to keep systems in calibration and get consistantly repeatable results, even when cost and time pressures aren't an issue.

[alex_lofquist]

I don't get too concerned about lens testing as it may only be a guide as to what you may expect. I used to test my Leica lenses, but they were only indicative of possible problems. A bench test will show spherical aberration, curvature of field, coma, astigmatism, lateral and longitudinal chromatic aberration and distortion. It will not tell you much about how your photos will look by the way you shoot. I had one lens that was praised by some people, but for my purposes it was a dog. Others that only received faint recommendations were excellent for me: "One man's Mede is another man's Persian!"

[jonny_mac]

What is the general consensus of the the-digital-picture.com site, in terms of their reviews. They seem to have every lens imaginable.

[bobatkins]

I think the-digital-picture information is useful. I wouldn't pick a lens based only the information on that site, but the info there might be a factor in making a decision.

 

I wouldn't use any one site, no matter how accurate their data, because if two factors. First sites generally test only one sample of a lens. Second there's more to a lens than simply optical performance. You wouldn't buy a car based only on the 0-60 acceleration time, or the top speed or the maximum lateral grip numbers, and I wouldn't buy a lens based on on the number of lp/mm it resolves when bolted down to an optical bench. Nor would I buy one on the basis of any single review.