3 ms·
> All that said, there are definitely cases where you do not want to use a perceptual space. Generally for image filtering, antialiasing, and alpha compositing,
by vanderZwan 6y ago
> All that said, there are definitely cases where you do not want to use a perceptual space. Generally for image filtering, antialiasing, and alpha compositing, you want to use a linear space (though there are subtleties here). And there are even some cases you want to use a device space, as the device gamut is usually nice cube there, while it has quite the complex shape in other color spaces.
Tangent: would it be correct to conclude that if I am trying to convert a high quality image into one with a limited palette that uses dithering (basically, think of gifs here), it would be best to determine the palette in a perceptual color space, and then proceed to handle the image dithering to that palette using a linear color space?
- OscarCunningham 6y agoThat sounds right to me.
- raphlinus 6y agoYes. It turns out there is quite a bit of literature on palette generation and quantization ([1] comes to mind, which uses CIELAB as you describe), but limited palettes are mostly a retro curiosity these days. [1]: https://engineering.purdue.edu/~bouman/publications/pdf/ei94b.pdf https://engineering.purdue.edu/~bouman/publications/pdf/ei94...
- vanderZwan 6y agoOr a tinkering curiosity that is fun to play with, in my case :) I've been meaning to explore if the ideas in Efficient palette-based decomposition and recoloring of images via RGBXY-space geometry[0] could be used for creating better paletted PNGs and GIFs. Learning more about color spaces is quite relevant, I would say. Thanks your for confirming my suspicions and the link! [0] https://cragl.cs.gmu.edu/fastlayers/ https://cragl.cs.gmu.edu/fastlayers/