More innovations in H.264/AVC software decoding

This talk will present a range of architectural and optimization techniques that were used in the development of a state-of-the-art H.264 software decoder (https://github.com/tvlabs/edge264), to drastically reduce code and binary size and improve speed. The techniques are applicable to other audio/video codecs, and will be presented as HOWTOs to help participants use them in their projects. It complements my talk from last year at FOSDEM'24, and will focus this time on (i) efficient parsing of input bitstream, (ii) writing C that yields near-optimal SIMD for both SSE and NEON, and (iii) additional architectural patterns to reduce code complexity .