jpeg encoding/decoder native to factor
dct : working but needs to be written again
huffman: working -- seems pretty alright. Apparently not necessary for decoding
quantizier: working
marker: very naive, kind of working. Can mostly do Huffman tables and DQT's now
TODO:
Decode DCT's
Decode scan with huffman tables
Reverse quantization
Run IDCT on the matrices
Test with raylib
The Final Codex needed in explaining what the hell is going on in a JPEG encoding/decoder. For some reason everyone has parts of an explanation but not a good full one.
You know. I usually don't rant in incomplete english all the time. (Although my hero of literature is Ezra Pound). But this, this deserves a special treatment. 10 articles on jpg decoding all incomplete like holes in a bucket. A standard that is filled with irrelevant details for implementation. You know what the best source I've found so far is? A weird russian site containing an article written by a college student from 20 years ago. That's the current status of JPEG documentation. Disgusting. How does the modern world even exist?