Vorbis as a lossy audio codec provides excellent quality audio compared to other codecs at a similar bit rate, including MP3, and even comparable to AAC. High-quality audio combined with a fully open license should be valued by the open source community as an alternative to codecs that require a patent license for encoding or payments in order to legally stream.
However, FFmpeg’s native vorbis encoder is currently labeled as experimental or broken due to its relatively poor performance and audio quality in comparison to competing audio codecs. It serves only to output a valid bitstream, with little concern about quality. This project aims to improve the general performance of the encoder, bringing it closer to a viable status. This will be done by integrating features of the existing AAC psychoacoustic model system within the native Vorbis encoder and adapting the FFmpeg Vorbis decoder to utilize that new information. Finally, additional documentation and test coverage will be added to enable long-term development in this area.