You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Dithering is an intentionally applied form of noise used to randomize quantization error, preventing large-scale patterns such as color banding in images.
Prior art:
I'm not sure what dithering algorithm tyrquake uses, but the bit depth of the texture source, pbase, and the corresponding output pixel location, pdest, are both 256 colors, so they don't appear to have anything to do with error diffusion and quantization error randomization because the error of colors is negligible. Furthermore, despite the fact that static data used in the process is called dither_kernel, it doesn't use a conventional method of image convolution, so additional diagnostics may be required. I adapted it for the quake-embedded, though, so if it's still required, I can provide it.
Furthermore, despite the fact that static data used in the process is called dither_kernel, it doesn't use a conventional method of image convolution, so additional diagnostics may be required. I adapted it for the quake-embedded, though, so if it's still required, I can provide it.
My intention was to provide a runtime switch to bring visual computing "overheads," so that rv32emu users can be aware of the impacts, especially when JIT compilation is ready for performance boost. Dithering might be a good fit for this purpose. Feel free to propose another approach as a runtime switch.
Dithering is an intentionally applied form of noise used to randomize quantization error, preventing large-scale patterns such as color banding in images.
Prior art:
The upstream work was landed in tyrquake.
The text was updated successfully, but these errors were encountered: