Audio Service

Note

The Audio Service is considered to be in Beta phase on Linux systems.

The Audio Service allows enabling and disabling audio for the local and remote participants, as well as configuring audio capture modes.

Audio Interface

#include <dolbyio/comms/audio.h>

Provides access to the local and remote audio services.

class audio

The audio service.

Public Functions

virtual local_audio &local() = 0

Gets the local audio service instance.

Returns:: the local audio service.

virtual remote_audio &remote() = 0

Gets the remote audio service instance.

Returns:: the remote audio service.

Local Audio

class local_audio

The local audio service.

This service is used to control local participant’s input audio properties, and enabling or disabling audio capture and sending into conference.

The start and stop audio methods of the local audio service should only be used when the conference is active.

Attention: The local_audio interface contains methods that return async_result. Each function that returns async_result is asynchronous and the operation is executed during the SDK event loop. The caller can block the calling thread until the operation completes using the wait helper. The caller can also chain consecutive operations, which are dependent on the completion of this method, using the async_result::then calls. Each async_result chain needs to be terminated with an async_result::on_error.

Public Functions

virtual async_result<void> start() = 0

Opens the input audio device and starts processing audio.

This method creates a WebRTC AudioTrack and attaches the AudioTrack to an active Peer Connection. This method also connects the Audio Sink of the media_source_interface with the WebRTC Audio Source, creating the audio delivery pipeline. When the async_result of this method resolves, the application can start an audio stream from any media source it chooses, pass the raw audio PCM data to the injector module from this source, and inject the audio data into a conference.

Attention: This method does not start the capturing from a media source; this action needs to be performed separately, when the async_result of this method is resolved.

Returns:: The result object producing the operation status asynchronously.

virtual async_result<void> stop() = 0

Closes the input audio device and stops processing audio.

This method destructs a WebRTC AudioTrack and detaches it from an active Peer Connection. This method also disconnects the Audio Sink of the media_source_interface from the WebRTC Audio Source, which deconstructs the audio delivery pipeline. All frames captured after the return of stop_audio are discarded, so an application should stop capturing audio from the local media source when async_result of this method resolves.

Attention: This method does not stop the capturing from a media source; this action needs to be performed separately by an application before or after calling stop_audio.

Returns:: The result object producing the operation status asynchronously.

virtual async_result<void> set_capture_mode(const audio_capture_mode::unprocessed &mode) = 0

Sets the unprocessed audio capture mode.

Parameters:: mode – The audio capture mode.
Returns:: The result object producing the operation status asynchronously.

virtual async_result<void> set_capture_mode(const audio_capture_mode::standard &mode) = 0

Sets the standard audio capture mode.

Attention: The Standard capture mode with low noise reduction is not supported in server-side use-cases or non-Dolby Voice conferences. Trying to set standard capture mode in the aforementioned cases will result in actual setting of Default capture mode.

Parameters:: mode – The audio capture mode.
Returns:: The result object producing the operation status asynchronously.

virtual async_result<audio_capture_mode::variant> get_capture_mode() = 0

Reads the current audio capture mode.

Attention: This method is only available for Dolby Voice Conferences.
Attention: This method is not supported in server-side use-cases.

Returns:: The result object producing the variant holding audio capture mode asynchronously.

Remote Audio

class remote_audio

The remote audio service.

This service is used to control the local properties of the remote participants’ audio, that is the audio which is played in the conference on the current client’s endpoint.

All methods of the remote audio service should only be used when the conference is active.

Attention: The remote_audio interface contains methods that return async_result. Each function that returns async_result is asynchronous and the operation is executed during the SDK event loop. The caller can block the calling thread until the operation completes using the wait helper. The caller can also chain consecutive operations, which are dependent on the completion of this method, using the async_result::then calls. Each async_result chain needs to be terminated with an async_result::on_error.

Public Functions

virtual async_result<void> start(const std::string &participant_id) = 0

Enables the participant’s audio to be played locally.

This method resumes the reception of audio coming from the remote participant, whose audio has been previously stopped locally. By default, remote participants’ audio tracks are enabled, and attempting to start their audio locally will do nothing. This method allows an audio track from the desired remote participant to be mixed into the Dolby Voice audio stream received by the SDK. If the participant does not have their audio enabled, this method does not enable their audio track.

Attention: This method is only available for Dolby Voice Conferences.

Parameters:: participant_id – The ID of the participant whose audio track you would like to hear.
Returns:: The result object producing the operation status asynchronously.

virtual async_result<void> stop(const std::string &participant_id) = 0

Disables the local playback of the remote participant’s audio.

This method allows the local participant to not receive an audio track from a selected remote participant. This method only impacts the local participant; the rest of conference participants will still hear the participant’s audio. This method does not allow the audio track of the selected remote participant to be mixed into the Dolby Voice audio stream that the SDK receives.

Attention: This method is only available for Dolby Voice Conferences.

Parameters:: participant_id – The ID of the participant whose audio track you would like to stop.
Returns:: The result object producing the operation status asynchronously.

Audio Processing

enum class dolbyio::comms::noise_reduction

The audio noise reduction level.

Values:

enumerator high: Reduce all noise and isolate voice.

enumerator low: Remove steady/stationary background noise.

struct unprocessed

Capture background noise and non-voice sounds.

This mode uses echo suppression. No noise suppression is applied, and so a low end microphone or a noisy environment may introduce the noise into the conference.

struct standard

Capture the voice and reduce non-voice sounds.

This mode focuses on best voice quality. Echo suppression is always enabled when using this mode. Additionally, it offers a tunable setting for the level of voice isolation and non-voice noise reduction, and supports the following settings:

when using noise_reduction::high, provides a strong voice isolation and attempts to remove all non-voice sound from input.
when using noise_reduction::low, removes only steady background noise from the input.

Public Functions

inline constexpr standard(noise_reduction reduce_noise)

The constructor.

Parameters:: reduce_noise – the noise reduction level.

Public Members

noise_reduction reduce_noise: The noise reduction level.

using dolbyio::comms::audio_capture_mode::variant = std::variant<unprocessed, standard>: A type-safe tagged union capable of holding any of the supported audio capture modes.