[quote name='TheChris' timestamp='1304498109' post='8052']
Do you think it has something to do with older vs. newer smaples (regadring the serial number) or is it normal sample variation?
[/quote]
The retested sample was certainly better than the initial one, but I don't think this has any direct relation to the lens' age or serial number. Actually, the retested sample was older than the initially tested one.
However, partly the better results are most likely also due to a much more experienced reviewer <img src='http://forum.photozone.de/public/style_emoticons/<#EMO_DIR#>/wink.gif' class='bbc_emoticon' alt='
' /> Seriously: I initially reviewed the AF-S 50/1.4 quite a while ago and certainly have learned a lot since then. In addition, especially with fast primes, I put a lot more effort in focus stacking now than I did then, which means I am now able to get finer steps (more "slices"), plus I often do two or three runs "through the green dot" (AF confirmation) by default instead of just one (it's less work do to it directly instead of discovering something was not quite right two days later and set up everything again). Imatest has also greatly improved in the recent years regarding some sort of batch mode which, combined with partly automated analysis of the results in Excel, allows for a more detailed analysis with a finer grid of measurement spots.
Don't put too much weight in this part, though: I still have the old RAW files from the first test and sent them to the current (more detailed) Imatest procedure again, the results were only marginally better. So, in this case, the first one was just a not that great sample. Today I would probably be more suspicious about the results and measure a second sample before I publish the results.
Sample variation simply happens occasionally, had it with an AF-D 85/1.8, too recently, as well as an AF-S 14-24 that was not as good as the earlier reviewed samples (but had already seen quite some use, too).
I will also retest the AF-S 24-70 on the D7000 soon and at the same time evaluate that lens on the D3x again. I have a feeling we will see better results with that lens, too.
-- Markus