Well, we all know USM are bad, old, and ugly. My point was that ugliness should be lighter in linear.
That’s still output (as opposed to processing).
Quantization is a non-issue since any modern image processing has an internal pipeline using 32 bits floats. You only care about that stuff while saving a file in integer format, not while processing filters.
- @Carmelo_DrRaw went for an edge-aware guided USM in linear,
- @heckflosse went for a lens blur deconvolution in linear,
- I’m doing something similar as @Carmelo_DrRaw in linear,
- Anat Levin does blind deconvolution in linear (see the Matlab code)
That makes a lot of people with image processing experience converging toward the same solution…