Opus WebCodecs Registration

1. Fully qualified codec strings

The codec string is "opus".

2. EncodedAudioChunk data

An EncodedAudioChunk containing Opus can be in two different formats.

If the bitstream is in opus format, EncodedAudioChunk have to be Opus packets, as described in section 3 of [OPUS]

If the bitstream is in ogg format, EncodedAudioChunk have to be audio data packets, as described in section 3 of [OPUS-IN-OGG].

3. AudioDecoderConfig description

description can be optionally set to an Identification Header, described in section 5.1 of [OPUS-IN-OGG].

If a description has been set, the bitstream is assumed to be in ogg format.

If a description has not been set, the bitstream is assumed to be in opus format.

4. EncodedAudioChunk type

The [[type]] for an EncodedAudioChunk containing Opus is always "key".

NOTE: Once the initialization has succeeded, any packet can be decoded at any time without error, but this might not result in the expected audio output.

5. AudioEncoderConfig extensions

partial dictionary AudioEncoderConfig {
  OpusEncoderConfig opus;
};

opus, of type OpusEncoderConfig

Contains codec specific configuration options for the Opus codec.

5.1. OpusEncoderConfig

dictionary OpusEncoderConfig {
  OpusBitstreamFormat format = "opus";
  OpusSignal signal = "auto";
  OpusApplication application = "audio";
  [EnforceRange] unsigned long long frameDuration = 20000;
  [EnforceRange] unsigned long complexity;
  [EnforceRange] unsigned long packetlossperc = 0;
  boolean useinbandfec = false;
  boolean usedtx = false;
};

To check if an OpusEncoderConfig is valid, run these steps:

1. If frameDuration is not a valid frame duration, which is described section 2.1.4 of [RFC6716], return false.

2. If complexity is specified and not within the range of 0 and 10 inclusively, return false.

3. If packetlossperc is specified and not within the range of 0 and 100 inclusively, return false.

4. Return true.

format, of type OpusBitstreamFormat, defaulting to "opus"

Configures the format of output EncodedAudioChunks. See OpusBitstreamFormat.

signal, of type OpusSignal, defaulting to "auto"

Specificies the type of audio signal being encoded. See OpusSignal.

application, of type OpusApplication, defaulting to "audio"

Specificies the encoder’s intended application. See OpusApplication.

frameDuration, of type unsigned long long, defaulting to 20000

Configures the frame duration, in microseconds, of output EncodedAudioChunks.

complexity, of type unsigned long

Configures the encoder’s computational complexity, as described in section 2.1.9. of [RFC6716]. The valid range is 0 to 10, with 10 representing the highest complexity. If no value is specificied, the default value is platform-specific: User Agents SHOULD set a default of 5 for mobile platforms, and a default of 9 for all other platforms.

packetlossperc, of type unsigned long, defaulting to 0

Configures the encoder’s expected packet loss percentage. The valid range is 0 to 100.

NOTE: The packet loss percentage might be updated over the course of an encoding, and it is recommended for User Agents to support these reconfigurations.

useinbandfec, of type boolean, defaulting to false

Specifies whether the encoder provides Opus in-band Forward Error Correction (FEC), as described by section 2.1.7. of [RFC6716].

usedtx, of type boolean, defaulting to false

Specifies if the encoder uses Discontinuous Transmission (DTX), as described by section 2.1.9. of [RFC6716].

5.2. OpusBitstreamFormat

enum OpusBitstreamFormat {
  "opus",
  "ogg",
};

The OpusBitstreamFormat determines if extra-data are necessary to decode the encoded audio stream.

opus

No metadata are necessary to decode the encoded audio stream.

ogg

The metadata of the encoded audio stream are provided at configuration via AudioDecoderConfig.description.

5.3. OpusSignal

enum OpusSignal {
  "auto",
  "music",
  "voice",
};

The OpusSignal indicates the default value for the type of signal being encoded.

auto

The audio signal is not specified to be of a particular type.

music

The audio signal is music.

voice

The audio signal is voice or speech.

5.4. OpusApplication

enum OpusApplication {
  "voip",
  "audio",
  "lowdelay",
};

The OpusApplication indicates the default value for the encoder’s intended application.

voip

Process signal for improved speech intelligibility.

audio

Favor faithfulness to the original input.

lowdelay

Configure the minimum possible coding delay by disabling certain modes of operation.

6. Privacy Considerations

Please refer to the section Privacy Considerations in [WEBCODECS].

7. Security Considerations

Please refer to the section Security Considerations in [WEBCODECS].