In this code we implement the IEEE 1857.2 lossless audio coding extension using 5 different source coding techniques:
- Arithmetic Coding using the actual probability mass function of the source (benchmark).
- Arithmetic Coding using an estimated probability mass function of the source.
- Golomb Coding choosing m.
- Adaptive Golomb-Rice coding with exhaustive search for the best k where m=2^k -> slow encoding fast decoding.
- Adaptive Exponential Golomb-coding with exhaustive search for the best k -> slow encoding fast decoding.
- Adaptive Golomb-Rice coding -> fast encoding fast decoding.
Clone repository to your local directory and execute: IEEE_1857dot2.m
General implementation
Linear predictive modeling (Analysis/Synthesis)
IEEE-1857.2 Encoder
IEEE-1857.2 Decoder
More information
This thesis provides a comprehensive guide to lossless audio compression and coding. It includes a MATLAB-based implementation that emulates the IEEE 1857.2 lossless audio coding extension offering capabilites of experimenting with new source encoders/decoders. Also, to improve compression efficiency we may consider adding ATC-ABS method.