There are a few things I want to add (my first post was done sort of in a rush, I had to pick up my daughter from school )


In doing my tests, I learned a lot about how difficult it is to do controlled comparisons of lenses. I gained a great deal of appreciation for the collection of ISO crops
Bryan has amassed and have become even more grateful for his work. On my first try, I taped the ISO chart to a shed outside and took pictures on a tripod. I tested a number of lenses at different times of the day. I knew different lighting could affect perception of sharpness, but I had no idea how much. Tests done in the morning looked soft compared to those done in the afternoon. To do it right, I decided I had to move my operation inside and use a flash for lighting. (Bryan uses several flashes... I forget how many but I think at least 4).


Another problem I had was getting the optical axis perpendicular to the ISO chart, or in other words, lining up the focal plane with the chart. Bryan describes a very involved method of doing this using expensive equipment that I don't have. I just taped the chart to the wall, leveled the camera with a bubble level, and squared it relative to the wall with a square. This wasn't good enough. Sometimes it worked okay, but sometimes corners would be blurry when I focused on the center, or if I focused on one corner, the others would be oof.


I even found that pictures shot at different ISO appeared to have different IQ. So for my final test, I shot ISO 100, manual mode, and made chose "neutral" picture style (all settings, including sharpness, set to 0).


The final surprise was what appeared to me very different results in different corners. As I said, the lenses didn't look that different in the upper left corner, but in the lower right, I thought the Mk 2 is much closer in quality to the 135mm f/2 than to the Mk 1.


The question "which lens has better IQ" or even "how much better is lens A than lens B" are not simple ones. I restricted myself to 200mm f/2.8 (for the 70-200's... obviously the 135 mm, for comparison sake, was 135mm f/2.8). I can hardly imagine doing all this work at several focal lengths and several f/ numbers for every canon lens made. (!)


All of this begs the question: if it is so difficult to detect differences in controlled conditions, how much is it really worth to have the best possible sharpness? I think question is a valid one. Of course, if one plans to crop heavily, one really does what to get the best sharpness possible. But much of the time, the ISO chart isn't the final word. Bokeh, color, and contrast are differences one does not need to be a pixel peeper to see. I didn't compare bokeh, but I do find color and contrast on the Mk 2 to be an improvement (but then I tend to be very suspicious of such subjective comparisons). The advantage of ISO comparison is that they are objective. It almost seems axiomatic that the more useful the parameter you're comparing, the more subjective the result. One does one's best.