Wahoowa
Offline
• • •
Upload & Sell: Off
|
 p.3 #10 · Best way to choose the best copy? | |
Just want to follow up on this.
Basically, only a handful of people did answer my question--if you have an opportunity to choose the best out of a bunch, what methodology will you use? The thread turns out to be a debate between one sees it as an important issue to do it.
Let me tell you that I'm one of those who don't pixel peep. Heck, I don't even care if my 24-105L has much more barrel distortion at 24mm than any other L lenses at the same focal length. And that's why the 24-105L is my lens of choice most of the time rather than any other lenses. (And I do own a 24L II.)
Anyhow, I always knew that copy variation exists, but I didn't think it'd be that significant until one of my friends complained his 35L wasn't sharp at all. I original thought he needed to do microadjustment. But, after some experiments, it turned out that it's a lemon that microadjustment wouldn't do any good.
In addition, reading how LensRental did their test on several copies of their lenses, it shows me that copy variation may not be insignificant. For example,
http://www.lensrentals.com/blog/2013/01/canon-24-70-f4-is-resolution-tests
That said, I've heard a lot of members here and on POTN normally look for a 'sharp' copy. So, I'm eager to know how you do it. Going through this thread, I'm kinda surprised that most said they didn't care. On the other hand, I'm sure if my original post was that copy variation is not that important, I would have had responses that said otherwise.
Like I mentioned before, I'd have an opportunity to test five copies and choose to buy one of them if I wanted to. As someone in academia, in engineering field and in related area to photography, I'm interested in knowing how to 'scientifically' measure these. Obviously, MTF measurement is the obvious method. But, after doing some search, Quick MTF is the only available software and it's quite costly, running around $200-$300. I figure I could write the program myself, but that will cost my time. (That's why I said there's a time limit.) Even for that Quick MTF program, it turns out that I'd have to do AFMA for each lens to achieve the best results.
In the end, it seems that Reikan FoCal is a good enough tool that provides some feedback regarding the sharpness of each lens. It gives a 'quality measure' that can imply sharpness of a focus. Also, I figured that I could use it to confirm my AFMA on my three camera bodies.
Running tests through 5 copies of a zoom lens gave the following numbers after AFMA (the first number is the wide end and the second one is the long end):
1. 1713.9, 1785.3
2. 1697.9, 2034.9
3. 1750.9, 2063.1
4. 1783.5, 2169.0
5. 1741.6, 2112.1
Although the lighting at my lab wasn't that great (at about 6 EV), the results were pretty consistent. Note here that the higher the number, the better the quality of focus. I believe these numbers are probably some sort of sum of mean-square of something. So, the difference in tenth is probably not that significant. But it was quite a surprise to see one of the lenses was far inferior than the rest. Pictures that came out of it weren't that bad. I didn't pixel peep them. Unless my friend's 35L that I mentioned before, I wouldn't have noticed anything wrong with the lens from the pictures. But I'd still consider the #1 lens a lemon.
In conclusion, I start to believe that copy variation is there and is greater than what I'd thought for a long long time. Will it change my habit of buying used gear? Probably not. Most of my lenses are used. And I'll continue to buy used lenses, but I'll now be more aware of the 'lemons'.
|
