I am currently performing some tests with my 50mm lenses and comparing them to the Voigtlander Nocton 58mm.
After doing very careful focussing tests I have come to the conclusion that these lenses reach the sensor limit of the K10D already at f1.4!
Below are 400% crops from the inner part of the USAF sheet (best out of 20).
What one can see is that the lens has less contrast at f1.4.
But the resolution is extraordinary high.
The curious thing is that at the number 4 it appears that the orientation of the bars is changing. Here the resolution is so high that the sensor cannot cope with it anymore correctly and artificial, "spurious" patterns appear. I think this has to do with the
Nyquist theorem or something similar.