Upload & Sell: Off
| Re: Canon 24-70mm f/2.8L II sample variation tested |
I think the suggestion about real world images someone made is a good one, so I've appended the article with 100% crops of an ISO chart shot with one of the highest and one of the lowest resolving lenses.
I think it helps demonstrate that the number differences, while real, seem a lot greater than what you'd notice in the real world.
Here is another visual demonstration the best 70mm f/2.8 and the worst I saw over a few (many less than you tried of course) copies. I actually was prompted to do the careful test because I noticed, in real world shots, that one didn't seem to be quite popping as much as the others (although this one did seem to have the most even corner performance of them all, oddly enough, so you picks your poison).
These are 200% to help rid jpg artifacts from having any affect on things. Difference is easiest to spot if you save both and flip between them in an image viewer. One should have crisper darks next to brighter whites, better micro-contrast for more image pop.
I think you can definitely see the difference between the best and worst of even just these few copies. If you save them and flip and forth the difference in micro-contrast clarity pop is quite evident. And this was just from a small sampling so it is unlikely that the best I saw was as good as the best RCicala found and unlikely that the worst I found was as bad as the worst he sampled (although you never know for sure), so the extremes should actually be readily apparent IMO, even real world (although the worst did have very even corner performance, perfectly even 100% and despite being the worst it was still as good as a 70-200 f/4 IS at 70mm f/4 wide open, generally considered to be an outstanding lens even if near 70mm is that lenses weak point for wide open performance so still not bad).
I used 10x liveview manual focus, remote trigger release and best of like a dozen or more tries for each.