(there are MANY posts through this large thread, in almost every page, with lots of excellent detail, and lots has been written in general about Obra Dinn. E.g. choosing different algorithms for different surfaces: https://youtu.be/_3qzyAHMoUU?t=2m24s)
For eink: yeah white mixed in is surprisingly easy to pick out compared to other color combinations. It's fine at enough of a distance, so maybe for a picture frame, but pretty rough at a middle reading distance or closer.
I hope TRMNL will start supporting 6 color e-paper displays more widely. I have one of these 6 color panels running with my own custom firmware on their BYOD license and I'm really happy with it, just wish it could take advantage of the color capabilities.
Yes! I'm in the same boat! My guess is that there's something in the backend (image cache/storage size) that presents as a cost/logistical problem instead of as a technical one.
After seeing a cool demo of using precomputed blue noise with thresholding to approximate transparency in shaders.
I'm quite interested about using the same technique for colour mixing.
Not sure if this is the same thing but I saw this one[0] a while back which has (spatiotemporal) blue noise applied to transparency, dithering, and various other shenanigans.
Obra Dinn's in-game dithering is also a cool entry for clever, tailor-made dithering in general: https://news.ycombinator.com/item?id=42084080
(there are MANY posts through this large thread, in almost every page, with lots of excellent detail, and lots has been written in general about Obra Dinn. E.g. choosing different algorithms for different surfaces: https://youtu.be/_3qzyAHMoUU?t=2m24s)
For eink: yeah white mixed in is surprisingly easy to pick out compared to other color combinations. It's fine at enough of a distance, so maybe for a picture frame, but pretty rough at a middle reading distance or closer.
check out this project https://github.com/Utzel-Butzel/epdoptimize
I hope TRMNL will start supporting 6 color e-paper displays more widely. I have one of these 6 color panels running with my own custom firmware on their BYOD license and I'm really happy with it, just wish it could take advantage of the color capabilities.
Yes! I'm in the same boat! My guess is that there's something in the backend (image cache/storage size) that presents as a cost/logistical problem instead of as a technical one.
After seeing a cool demo of using precomputed blue noise with thresholding to approximate transparency in shaders. I'm quite interested about using the same technique for colour mixing.
Where did you see that? Sounds interesting
Not sure if this is the same thing but I saw this one[0] a while back which has (spatiotemporal) blue noise applied to transparency, dithering, and various other shenanigans.
[0] https://arxiv.org/pdf/2112.09629