Hi,
I have repeated the test using a better target and being very careful about focusing and image stabilization. The short answer is that the degradation caused by the filter is negligible in terms of optical quality, which is opposite to my initial, less careful test.
Some observations:
- I could not get best focus using autofocus , even in live view (contrast detect).
- The difference between filter and no filter was higher when using autofocus
- The best result using autofocus without a filter was worse than carefull manual focus with the filter
- It is really difficult to achieve perfect focus with the D800 due to the lousy live view implementation
I have performed test using Bart Van der Wolf target (thank you). Using different apertures and lenses I could always get into the "aliasing" zone (if it can be called as such) which is basically reaching the limit of the sensor (I'm still not convinced that you can get rid of the AA filter, but that's another discussion).
Setup:
Bart's target, located at least 30X the focal length.
Flash lighting to avoid possible camera movement
Camera tethered, using 2 seconds shutter delay
Multiple captures in each combination, selected the best focused image in each case
Process:
All images processed in LR 5.5 from Raw files. No sharpening, or any other modification except for applying a brush with moire reduction (20) in the "aliasing" zone. I found that the color moire caused much distraction when comparing the images.
Thanks for your comments and suggestions.
Here the results: (these are 200% crops)
Left side: no filter, right side with filter
First image: 70-200 VRI f/2.8 @ 200 mm, f/4
Second image: 70-200 VRI f/2.8 @ 200 mm f/5.6
Third image: 70-200 VRI f/2.8 @ 105 mm f/5.6
Fourth image: 17-35 f/2.8 @ 22 mm f/5.6