New Lens Testing Method

There is a lot in your post that I only have a passing understanding of. As my limited understanding goes the Demosaicing issue relates to color rendition. In-camera processing creates the accurate color rendition but in combining the RBG information resolution is lost.

I understand that by using a TC then reducing the resultant image the resolution lost by the demosaicing algorythms is clawed back( by combining information from adjacent sensor pixels). I don't know but guess that it still sees some loss in resolution.

However I have this question. are you really testing the cnr-cnr sharpness of the lens or just cnrs of the resultant Field of view? I ask this as I have a full frame sensor. That may not seem important with the majority of users on APSc. However, I suspect we are going to see increasing numbers of entry level cameras with FF sensors...so I do think there may be an issue of future proofing your work, that deserves some consideration.

That said I don't think I can offer a better alternative.

I do like Mark Goodman's approach but, at times it does make my head hurt; meaning I have to think hard to relate his MTF graphs to likely real life results( He uses MTF 50 I believe). Using actual coins ( as long as the viewer knows them) does make it easier to relate your results to what I might get out of the same lens.

I would suggest that you consider adding in a simple CA measure. The one I have come to understand the most is Mark Goodman's ruler at a 45 degree angle. If you were able to add that measure ( at the optimum F stop only) it would be a great help.

Here are some images to show the new testing method using a 75ARD1 at f4. Each image was rendered from a focus stack of 5 source images.

Overall image, without 2xTC:


Above image, 50% Center Crop:


Overall image, with 2xTC:


The above two images are the ones to be compared, so I will take crops from them, and magnify the crops 200% to make it easier to see differences.

Center crops animation from images with and without 2xTC:


Upper left corner crops animation:


There is some distortion of the images, especially the corner crops, since I was not using a telecentric lens, and turned off scaling in Helicon.