Opus: The Universal Audio Codec
Opus is an open, royalty-free audio codec developed by the IETF (Internet Engineering Task Force) and published as RFC 6716 in September 2012. It was designed to replace multiple specialized codecs with a single codec that handles everything from ultra-low-bitrate speech to high-fidelity music.
Opus was created by merging two existing technologies: SILK (developed by Skype for voice calls) and CELT (Constrained Energy Lapped Transform, developed by Xiph.Org for music). The result is a hybrid codec that seamlessly switches between speech and music modes depending on the content.
Key point: Opus is not just another audio format — it is the mandatory audio codec for WebRTC, the standard behind all browser-based voice and video calls. Every browser that supports WebRTC (Chrome, Firefox, Safari, Edge) must implement Opus.
How Opus Works: Hybrid Architecture
Opus uses a unique hybrid approach that no other codec offers:
- SILK mode (speech): optimized for human voice at bitrates from 6 to 40 kbps. Uses linear prediction (LP) similar to traditional speech codecs. Ideal for VoIP, voice messages, and podcasts.
- CELT mode (music): optimized for general audio and music at bitrates from 48 to 510 kbps. Uses a modified discrete cosine transform (MDCT) similar to AAC and Vorbis. Ideal for music streaming and high-fidelity audio.
- Hybrid mode: combines both SILK and CELT for bitrates between 40 and 80 kbps. SILK handles the low frequencies (below 8 kHz) while CELT handles the high frequencies. This gives speech naturalness with music quality.
The codec can also switch between modes on the fly — frame by frame — with zero gaps or glitches. If a VoIP call transitions from speech to hold music, Opus seamlessly switches from SILK to CELT mode without any reconfiguration.
| Mode | Bitrate Range | Bandwidth | Best For |
|---|---|---|---|
| SILK (LP) | 6–40 kbps | Up to 8 kHz (narrowband–wideband) | VoIP, voice messages, low-bandwidth speech |
| Hybrid (SILK + CELT) | 40–80 kbps | Up to 20 kHz (super-wideband) | HD voice calls, voice + background music |
| CELT (MDCT) | 48–510 kbps | Up to 20 kHz (fullband) | Music streaming, high-fidelity audio |
Who Uses Opus?
Opus has been adopted by nearly every major internet platform that deals with real-time or compressed audio:
- WhatsApp: all voice messages are encoded as Opus inside OGG containers. When you record a voice message, WhatsApp uses Opus at approximately 16–32 kbps to keep file sizes tiny while maintaining clear speech quality.
- Telegram: voice messages and audio messages use Opus encoding, also in OGG containers.
- Discord: all voice chat in Discord uses Opus for real-time audio. Discord transmits voice at 64–96 kbps Opus, which provides excellent quality at low bandwidth.
- WebRTC: Opus is the mandatory audio codec for WebRTC, the standard behind Google Meet, Zoom (web version), Microsoft Teams (web), and all browser-based calling apps.
- YouTube: audio tracks in YouTube's WebM container use Opus. When you download a YouTube video's audio with tools like
yt-dlp, you often get an Opus file. - Spotify: uses Opus (via Ogg) for streaming on desktop and mobile at up to 320 kbps in the "Very High" quality setting.
- Wikipedia / Wikimedia: all spoken articles and audio files on Wikipedia use Opus as the preferred format.
Scale: WhatsApp alone processes over 7 billion voice messages per day. Combined with Discord, Telegram, WebRTC calls, and YouTube, Opus is decoded billions of times every hour — making it the most frequently used audio codec in the world by a huge margin.
Opus vs Other Audio Codecs
In listening tests (including the official IETF tests), Opus consistently outperforms every other lossy codec at equivalent bitrates:
| Codec | Type | Latency | Quality at 64 kbps | Quality at 128 kbps |
|---|---|---|---|---|
| Opus | Lossy, open | 5–66 ms | Good | Excellent |
| MP3 | Lossy | ~100 ms | Poor | Acceptable |
| AAC-LC | Lossy | ~90 ms | Fair | Good |
| Vorbis | Lossy, open | ~100 ms | Fair | Good |
| HE-AAC v2 | Lossy | ~100 ms | Good (speech) | N/A (designed for low bitrate) |
The key advantage of Opus is not just quality — it is versatility. No other codec handles the full range from 6 kbps narrowband speech to 510 kbps full-bandwidth music. MP3 was designed for music only. Speex was designed for speech only. AAC works for both but has high latency. Opus does everything with ultra-low latency.
Opus File Extensions and Containers
Opus audio data can be stored in several container formats:
.opus— Opus in an Ogg container. This is the standard file extension for Opus audio files. Defined in RFC 7845..ogg— Opus (or Vorbis) in an Ogg container. WhatsApp and Telegram save voice messages as.oggfiles containing Opus audio. This can cause confusion because.oggfiles may contain either Vorbis or Opus..webm— Opus in a WebM (Matroska) container. YouTube uses this for audio-only WebM files..mkv/.mka— Opus in Matroska. Some video files use Opus as the audio track inside MKV containers.
WhatsApp confusion: WhatsApp voice messages are saved as .ogg files, but they contain Opus audio, not Vorbis. If you try to play them in a player that only supports OGG Vorbis, they will fail. Modern players like VLC handle both codecs in OGG containers.
Device and Software Compatibility
Opus support has grown rapidly since its 2012 introduction, but it is still not universal:
| Platform / Device | Opus Support | Notes |
|---|---|---|
| Chrome / Firefox / Edge | Native | Full support since 2013 |
| Safari | Native | Since Safari 15 (2021), web playback only |
| Android | Native | Since Android 5.0 (2014) |
| iOS / iPhone | Partial | Safari web only; Music app does not support .opus files |
| Windows 10/11 | Partial | Edge/Chrome play Opus; Media Player needs codec |
| VLC Player | Native | Full support on all platforms |
| Car stereos | No | No car stereo supports Opus |
| Portable MP3 players | No | Hardware players do not support Opus |
The main reason to convert Opus to MP3 is compatibility with devices that don't support the codec: car stereos, older media players, Apple's Music app, and any device that only understands MP3.