I'd argue compressed file formats make up a lot less of the broader problem space than most people realize. If you're doing anything high performance (where th

I'd argue compressed file formats make up a lot less of the broader problem space than most people realize. If you're doing anything high performance (where the difference matters) but not immediately consumer facing you're spending a lot of time working with lossless/uncompressed imagery in ram/disk. Stuff like MIPI->CUDA for classification, this question can go a number of different ways very quickly. Keep in mind, at least in my experience, if you're even remotely thinking about doing anything with grayscale and this question comes up you're probably doing something...interesting (read: multi band remote sensing shenanigans).