[mst]

[quote name='Brightcolours' timestamp='1329725153' post='15940']

The difference in edge/corner resolution can only be explained by not too accurate testing procedures and/or probably two different lens samples.[/quote]



Allow me to disagree. The D7000 results have been verified by two additional copies of the lens, just to be sure it's not a bad sample.



The D7000 generally shows a tendency to give weaker results towards the borders (which may be a common problem with high density sensors, see the Nex-7 issues Klaus is facing). Compared to the D200 tests, some things have changed, for example we (have to) use a different RAW converter. The higher pixel density means that we measure at a different frequency and different lenses may behave slightly differently here (compared to the lower res D200).



[quote name='Brightcolours' timestamp='1329725153' post='15940']

We can observe another oddity too, but that in my opinion almost certainly has to do with the testing procedure. The "measured" CA is twice as high with the D7000 values than with the D200 values. But this matches the odd calculated values in other tests.. not a reliable method to determine CA amounts.

[/quote]



We believe it's a results of the different RAW converters used for different systems. The different demosaic algorithms may lead to slightly different amounts of CAs by emphasizing the issue or by including a small amount of default CA correction.



In any case: from the very first beginning Klaus has always emphasized that the results are not comparable across different test systems. There are simply too many parameters that have an influence on the measured values to allow direct comparisons.



-- Markus