Opus is an audio codec that encodes and decodes interactive speech and audio for transmission over the Internet.
Opus solves the problem of transmitting high-quality audio in real-time over lossy networks. It scales from low bit-rate narrowband speech to high-quality stereo music, handling a wide range of applications including Voice over IP, videoconferencing, in-game chat, and remote live music performances. The codec is specified by IETF RFC 6716 and implements both encoding and decoding as a shared library for raw Opus bitstreams. The implementation includes a deep learning based redundancy encoder called DRED that improves audio quality in lossy network environments by embedding recovery data in packet padding, maintaining full backward compatibility with prior versions.
Opus suits any project requiring interactive audio transmission over the Internet, from real-time communications to stored-file applications like music distribution and game soundtracks. The library provides the core encoding and decoding functionality; for file-based applications, Opus bitstreams should be encapsulated in Ogg format according to RFC 7845, with the separate opus-tools package providing convenient encoding and decoding of Ogg-encapsulated files. The codec is royalty-free under the patent and copyright licenses specified in the repository.
The project maintains a reference implementation in C with active development addressing network challenges. The codebase includes test tools for verifying correct operation of the library, and the team has continued advancing packet loss concealment methodology beyond the original implementation. Development incorporates standardization efforts through the IETF, with the deep learning components having undergone open-source development and standardization processes.
