Basically, whether the correction is minor or extreme, even the "theoretical" minus at it's worst will be about the same.
The best case is that a pixel of the corrected image more or less exactly aligns with a single pixel in the original image, in which case even in theory there is no degradation.
The worst case that a pixel of the corrected image is composed of up to four pixels in the org image. For instance say, on an R5, we have 8192 pixels in the width, of 36mm, for 227.5 pixels per mm. If we had alternating white and black lines in our image, and had them perfectly lined up with the R5 pixel grid (impossible), we could in theory have 227.5/2=114 lp/mm or so. And if our lens further had 100% contrast in the MTF chart at 114 lp/mm, we'd have actual black and white lines in the org image. And in this case, the corrected image's contrast for 114 lp/mm would fall down to 0, because the corrected pixel would be half-composed of white original pixels, and half of black. The result would be pure grey. But then again, if you shift your camera aim a half-pixel in such a scenario, you'd ALSO have pure grey even in an uncorrected image, because the pixel would be half-illuminated by a white line and half by a black line.
Do the same analysis for 2-pixel wide lines, a 57 lp/mm image, and you'll find your worst-case loss is 25%. Some corrected pixels will still be half of a white original pixel, and half of a black, in the very worst case, but that's only half the pixels. The other half will still be pure white or pure black.
But in the real world, of course, MTF charts only go up to 30. Just look at any MTF chart, try to guess where the 60 lp/mm line would be, and imagine the contrast to be a quarter worse. In practice that means contrast is going from, like 20, to 15, or something.
Can you think of an image you've ever taken, one you've even pixel-peeped for hours and are intimately familiar with, where you'd actually note that 2-pixel-wide features' contrast has gone from 20 to 15?
Upvote
0
