really need small-dev/OSS projects to stop stepping into the April Fools pitfall of making common user feature requests their punchline
I guess dodging things like LUTs might make things tricky but PNG replacement seems like the easiest usecase to mess with
re:
https://www.ludicon.com/castano/blog/2026/03/hardware-image-compression/, I wonder if any console emulators inject texture compression into uploaded textures to save bandwidth, that seems like a lowball Android/PNG texture replacement optimization
Hardware Image Compression
One of the things I've always lamented about hardware image formats is the slow pace of innovation. Developers were usually unwilling to ship textures in a new format unless that format was widely available. That is, the format had to be supported in the majority of the hardware they were targeting,
Phenac immediately giving me the average Times Square experience
#NintendoSwitchmareep backwards is pee ram
they made essential oils for Pokemon
#NintendoSwitchtbh I suspect the blotchiness and nose-smushing here is more an artifact of the dataset they used versus the architecture, Apple likely used JPEGs and way more faces in their dataset since users use JPEGs and a lot of faces
The actual depth maps get things right though, sharp edges but generally tends towards blotchiness when it's uncertain
(InfiniDepth has its own splat generation but it creates more holes, needs more vertices, and doesn't wrap edges well)
Dark
