######################################################

What is “Dither”?

“Dither is an intentionally applied form of noise used to randomize quantization error, preventing large-scale patterns such as color banding in images. Dither is routinely used in processing of both digital audio and video data, and is often one of the last stages of mastering audio to a CD.”
More information


Explained with a screen:

color_depth_compare.png


Why using it?

Because “Why not?”. Almost all filters are natively supporting 16bit.
This is giving you some good advatages against banding and can also decrease the file size.


Are there many ways to dither something up/down?

Depends, there are many dither algos. Some of them are implemented in Vapoursynth. Keep in mind that dithering from 16bit to 10bit is almost visible free of dithering. The dither algo doesn’t really matter that much. That’s why the default value is the best to go. But if you dither down to 8bit, you’ve to choose wisely.


Modes you can use with “mvf.Depth(src, dither=X)”:

 

0 Ordered dithering (Bayer matrix).
1 No dither, round to the closest value.
2 Round, may be a bit faster but possibly less accurate.
3 Sierra-2-4A error diffusion, aka “Filter Lite”. Quick and excellent quality, similar to Floyd-Steinberg. (Default)
4 Stucki error diffusion. Preserves delicate edges better but distorts gradients.
5 Atkinson error diffusion. Generates distinct patterns but keeps clean the flat areas.
6 Floyd-Steinberg error diffusion. Classic.
7 Ostromoukhov error diffusion. Slow, available only for integer input at the moment. Avoids usual F-S artefacts.
8 Void and cluster halftone dithering. Better visual aspect than ordered dithering.

Unbenannt.png

 

Advertisements