Opus Audio Codec on YouTube
This article explains what the Opus audio codec is and how YouTube uses it.
What is the Opus Audio Codec?
Opus is an audio codec designed to efficiently compress digital audio data. It was adopted as a standard by the Internet Engineering Task Force (IETF) in 2012 and is supported by numerous applications and platforms such as WebRTC, Skype, WhatsApp, Discord, Mozilla Firefox and Google Chrome. Capable of covering a wide range of audio frequencies, including speech and music, the Opus codec offers high audio quality at low bit rates. It uses a combination of linear and non-linear predictive filters and variable sampling rateto allow efficient compression. This allows high quality audio data to be transmitted at bit rates as low as 6 to 510 kbps. Another advantage of the Opus codec is its adaptability to different network conditions. It supports variable bitrates and can adapt its encoding speed and quality in real time to the available network bandwidth. This ensures that the audio quality remains stable during transmission and that there are no interruptions or delays.
How does YouTube use the Opus audio codec?
Want the best possible, distortion-free playback of your song on every platform?
What bitrates does the Opus audio codec use?
The Opus audio codec can use different bitrates for audio encoding. It was designed to deliver high audio quality at low bitrates, while also supporting higher bitrates when better audio quality is needed. Typical bitrates for the Opus codec range from 6 kbps to 510 kbps. The codec can also support bitrates of up to 768 kbps if even higher audio quality is required. Opus uses variable bitrate encoding (VBR), allowing the bitrate to adapt in real time to the audio content. This means the codec can use higher bitrates for more complex material and lower bitrates for simpler material to balance audio quality and compression. In practice, this makes Opus suitable for a wide range of audio applications, from low-bitrate speech transmission to higher-bitrate music streaming. The YouTube app itself often uses significantly lower bit rates than the browser version on the smartphone.Opus 251 vs. 774: What is the difference?
251 and 774 are format numbers under which YouTube delivers its Opus audio streams. Google itself does not document these numbers – they become visible in download and analysis tools like yt-dlp. The difference lies in the bitrate: Format 251 is the standard Opus stream at around 128 kbps, while Format 774 is the Opus stream with up to 256 kbps, which YouTube Music Premium plays at the "High" quality setting.
| Format 251 | Format 774 | |
|---|---|---|
| Codec | Opus | Opus |
| bitrate | approximately 128 kbit/s | up to 256 kbit/s |
| Available for | all users (standard quality) | Premium only, quality level “High” |
| sampling rate | 48 kHz | 48 kHz |
This changes little for your mastering: Both streams use the same codec. A master with enough true peak headroom sounds clean in both formats – with 774, codec artifacts are just less noticeable thanks to the higher bitrate.
Can the use of the Opus audio codec cause distortion?
Yes, there may be some distortion when using the Opus codec. Distortions can occur if the codec is used at bit rates that are too low, or if the audio material is too complex for the selected bit rate. In these cases, the codec cannot store enough information about the audio, which can result in degraded audio quality and distortion. Another factor that can cause distortion is the way the codec has been implemented. If the codec has not been correctly implemented or is incompatible with the software or hardware used, distortion can also occur. However, it is important to note that the Opus codec is generally optimized for high audio quality at low bitrates. If the codec is used with a sufficient bitrate and the audio is not too complex, the audio quality should be good and distortion can be avoided.How do I avoid codec distortions?
To avoid distortion, the Streaming Services Specifications As a rule of thumb, master your song so that it sounds good. If integrated loudness is higher than -14 LUFS is, better leave 2 dB true peak headroom, um Intersample peak distortion to avoid.Want to know whether your song works on every platform? Our professional mix analysis will tell you.
Opus codec gets an AI update
The new version 1.5 or 1.5.1 of the license-free audio codec Opus has received an AI update. Machine learning (ML) is intended to improve the coding so that the data stream remains compatible with existing decoders. But the decoder also receives artificial intelligence to improve the sound.
AI for better sound quality
A technique called “neural vocoder” is meant to compress speech particularly efficiently. Compared with the LPCNet vocoder, it should put only about one percent load on the CPU cores of laptops or smartphones. The developers call the algorithm Framewise AutoRegressive Generative Adversarial Network (FARGAN). They plan to publish a paper on it later.
They optimize signal processing with the Linear Adaptive Coding Enhancer (LACE) and a nonlinear variant (NoLACE). LACE behaves like a classic postfilter in which a deep neural network (DNN) adjusts the coefficients on the fly using all available data – but the audio signal itself does not pass through the DNN. The result is a small DNN with very low complexity that also works on older phones. The NoLACE variant reportedly requires more computing power but is also significantly more powerful thanks to its nonlinear signal processing. Both improve speech quality significantly.
The Opus codec noticeably shapes how YouTube sounds — the Mastering Guide explains how to prepare your mastering for this in the streaming context.
Opus codec: The bitstream stays standards-compatible
Instead of programming a completely new codec based on ML, Opus remains completely compatible. This ensures that Opus continues to run on older and slower devices while providing an easy upgrade path. While deep learning is often associated with powerful GPU accelerators, the Opus project has optimized everything so that it runs on most processors, including smartphone CPUs.
Most users shouldn't notice the higher load, but those using microprocessors or smartphones that are more than five years old might notice it. The new functions are therefore still deactivated by default and must be activated during compilation and at runtime, for example via command line parameters.
Improve packet loss
Packet loss leads to missing sound fragments. Codecs usually try to prevent this through packet loss concealment (PLC). This is usually a kind of decoder-side interpolation with "plausible audio" inserted at the loss points. Machine learning could be particularly helpful here - the Opus developers are tackling this with a deep neural network (DNN), which increases the codec's binary file by around 1 MByte and leads to one percent more load on a laptop CPU core in the event of severe packet loss.
Tip: Hear how your master sounds after the YouTube Opus codec – free in your browser with the codec preview in the Loudness Meter.
