4 ms·
Yes, but this is really a pity in the specific case of Arithmetic Coding, because, unlike "more recent changes to the format", it's been in the standard since t
by gruturo 3y ago
Yes, but this is really a pity in the specific case of Arithmetic Coding, because, unlike "more recent changes to the format", it's been in the standard since the very beginning - but is not supported by a lot of implementations due to software patents (which meanwhile expired, but their damage remains).
- adgjlsfhk1 3y agoarithmetic encoding is old, but around 2010 (when all the patents expired), there was a ton of really good research on how to to table based ans and vectorized rans to make the performance good. Aside from the patent issues Arithmetic encoding wasn't pursued much because the CPU cost was too high. Now that multiply is cheap and the divisions can be avoided, ANS is a lot better than it used to be.
- JyrkiAlakuijala 3y agoAVIF, AV1, WebP lossy and Opus use traditional arithmetic coding. JPEG XL, Brunsli, Draco 3D and ZStd use table-based arithmetic coding (ANS). JPEG, Deflate, Brotli, MP3, WebP lossless and the fastest mode of JPEG XL lossless use prefix coding.