- Banned
- #21
GlennChan
Banned
Eh, you can always move the monitor away. 
2- Some of my untested ideas to increase sharpness would be:
--Apply enough deconvolution to counteract the effect of OLPF. Beyond that, other approaches are more appropriate. Applying too much deconvolution wouldn't make sense and would cause ringing.
--(Area-based) Tone mapping / HDR companding would likely incidentally improve acutance. A little bit of this would be a good idea anyways.
--Increasing acutance is a cheat that would 'improve' sharpness. Perceived resolution is half the battle though! It's ridiculous how much a difference it makes.
--If the viewer is sitting too far from the display, there may be tricks that can overcome the optical blurring that occurs.
--IMO, the HVS applies some sort of 'vernier scanning'. It's able to detect alignments and details smaller than the size of a photosite by moving the eye. I think that simple sharpen filters and USM works because of this.
--Doing sharpening in the linear light domain may be superior. You can sort of see this in Photoshop, although its levels filter doesn't round properly (which causes big shifts for low values).
--Don't move the camera erratically. Vernier scanning only works on stationary images, or on movement that the eye can track.
--A sharpening algorithm that works hand in hand with noise reduction may be a good idea, as one affects the other.
--If you know that the target display device is a LCD, you might be able to do sub-pixel tricks like Cleartype in Windows XP.
2- Some of my untested ideas to increase sharpness would be:
--Apply enough deconvolution to counteract the effect of OLPF. Beyond that, other approaches are more appropriate. Applying too much deconvolution wouldn't make sense and would cause ringing.
--(Area-based) Tone mapping / HDR companding would likely incidentally improve acutance. A little bit of this would be a good idea anyways.
--Increasing acutance is a cheat that would 'improve' sharpness. Perceived resolution is half the battle though! It's ridiculous how much a difference it makes.
--If the viewer is sitting too far from the display, there may be tricks that can overcome the optical blurring that occurs.
--IMO, the HVS applies some sort of 'vernier scanning'. It's able to detect alignments and details smaller than the size of a photosite by moving the eye. I think that simple sharpen filters and USM works because of this.
--Doing sharpening in the linear light domain may be superior. You can sort of see this in Photoshop, although its levels filter doesn't round properly (which causes big shifts for low values).
--Don't move the camera erratically. Vernier scanning only works on stationary images, or on movement that the eye can track.
--A sharpening algorithm that works hand in hand with noise reduction may be a good idea, as one affects the other.
--If you know that the target display device is a LCD, you might be able to do sub-pixel tricks like Cleartype in Windows XP.
