PySide6.QtMultimedia.QAudioSink¶

class QAudioSink¶

The QAudioSink class provides an interface for sending audio data to an audio output device. More…

Synopsis¶

Methods¶

def __init__()
def bufferFrameCount()
def bufferSize()
def bytesFree()
def elapsedUSecs()
def error()
def format()
def framesFree()
def isNull()
def processedUSecs()
def reset()
def resume()
def setBufferFrameCount()
def setBufferSize()
def setVolume()
def start()
def state()
def stop()
def suspend()
def volume()

Signals¶

def stateChanged()

Note

This documentation may contain snippets that were automatically translated from C++ to Python. We always welcome contributions to the snippet translation. If you see an issue with the translation, you can also let us know by creating a ticket on https:/bugreports.qt.io/projects/PYSIDE

Detailed Description¶

Warning

This section contains snippets that were automatically translated from C++ to Python and may contain errors.

You can construct an audio output with the system’s default audio output device. It is also possible to create QAudioSink with a specific QAudioDevice . When you create the audio output, you should also send in the QAudioFormat to be used for the playback (see the QAudioFormat class description for details).

QAudioSink can be used in two different modes:

Using a QIODevice from an application thread

Using a callback-based interface from the audio thread

Starting to play an audio stream is simply a matter of calling start() with a QIODevice. QAudioSink will then fetch the data it needs from the io device. So playing back an audio file is as simple as:

QFile sourceFile # class member.
QAudioSink* audio # class member.
    sourceFile.setFileName("/tmp/test.raw")
    sourceFile.open(QIODevice.OpenModeFlag.ReadOnly)
    format = QAudioFormat()
# Set up the format, eg.
    format.setSampleRate(44100)
    format.setChannelCount(1)
    format.setSampleFormat(QAudioFormat.Int16)
    info = QAudioDevice(QMediaDevices.defaultAudioOutput())
    if not info.isFormatSupported(format):
        qWarning() << "Raw audio format not supported by backend, cannot play audio."
        return

    audio = QAudioSink(format, self)
audio.connect(QAudioSink::stateChanged, self.handleStateChanged)
    audio.start(sourceFile)

The file will start playing assuming that the audio system and output device support it. If you run out of luck, check what’s up with the error() function.

After the file has finished playing, we need to stop the device:

def stopAudioOutput(self):

    audio.stop()
    sourceFile.close()
    del audio

At any given time, the QAudioSink will be in one of four states: active, suspended, stopped, or idle. These states are described by the State enum.

The QIODevice interface is designed to be used from the application thread. A wait-free ringbuffer is used to communicate to the audio thread. The size of this ringbuffer can be configured with setBufferSize() and defaults to 250ms. The state of this buffer can be queried with bytesFree() . If the ringbuffer runs out of data, the audio thread will send silence to the audio device and the state will change to IdleState and resume to ActiveState when more data is available from the QIODevice.

The preferred way to achieve low audio latency is to use the callback-based interface. It allows you to write audio data directly to the audio device without having to go through a QIODevice. This is done by calling start() with a callback function that will be called from the audio thread. This callback function will be called with a QSpan<SampleType> whenever the audio backend requires data.

QAudioSink* audio # class member.
float phase # class member.
    format = QAudioFormat()
# Set up the format, eg.
    format.setSampleRate(44100)
    format.setChannelCount(2)
    format.setSampleFormat(QAudioFormat.Float)
    info = QAudioDevice(QMediaDevices.defaultAudioOutput())
    if not info.isFormatSupported(format):
        qWarning() << "Raw audio format not supported by backend, cannot play audio."
        return

    audio = QAudioSink(format, self)
    phaseIncrement = 2 * M_PI * 220.0 / format.sampleRate() # 220 Hz sine wave
    audio.start([phase, phaseIncrement] (QSpan<float> interleavedAudioBuffer) {
    # The audio callback should not call any functions that may potentially be blocking
    # Fill the audio buffer with a sine wave
        sampleCount = interleavedAudioBuffer.size() / 2 # Stereo, so divide by 2
        for i in range(0, sampleCount):
            sample = std::sin(phase)
            interleavedAudioBuffer[i * 2] = sample # Left channel
            interleavedAudioBuffer[i * 2 + 1] = sample # Right channel
            phase += phaseIncrement # Increment phase for next sample

})
if not audio.error() == QtAudio.Error.NoError:
    # in addition to the other start() signatures, starting the audio callback will fail if
    # * the backend does not implement callback-based IO (the API is available on all major
    #   platforms)
    # * the signature of the audio callback does not match format.sampleFormat()
    qWarning() << "Error starting audio output:" << audio.errorString()

Unlike the QIODevice-based interface, the QAudioSink can only be in the states active, suspendend and stopped. The setBufferSize() API is not available when using the callback, the size of the callback argument is determined by the audio backend.

Note

This API is only available on platforms that support the callback API: Apple’s CoreAudio (macOS, iOS, etc), Windows, Linux (using the PulseAudio or PipeWire backend) and Android.

Note

The callback will be called on a soft-realtime audio thread. It is important to ensure that the callback does not block, as this can cause audio glitches or dropouts. This includes performing blocking IO, locking mutexes, allocating memories or any other operations that may block. For best practices consult Ross Bencina’s article Real-time audio programming 101: time waits for nothing . Also consider using clang’s Realtime sanitizer to validate the audio callback.

State changes are reported through the stateChanged() signal. You can use this signal to, for instance, update the GUI of the application; the mundane example here being changing the state of a play/pause button. You request a state change directly with suspend() , stop() , reset() , resume() , and start() .

The QAudioSink will enter the StoppedState when an error is encountered. The error type can be retrieved with the error() function. Please see the Error enum for a description of the possible errors that are reported. Calling stop() or reset() will reset the error state to NoError .

You can check for errors by connecting to the stateChanged() signal:

def handleStateChanged(self, newState):

match newState:
        case QtAudio.IdleState:
        # Finished playing (no more data)
        AudioOutputExample::stopAudioOutput()

        case QtAudio.StoppedState:
        # Stopped for other reasons
            if audio.error() != QtAudio.NoError:
            # Error handling


        case _:
        # ... other cases as appropriate

See also

QAudioSource QAudioDevice

__init__([format=QAudioFormat()[, parent=None]])¶

Parameters:

format – QAudioFormat
parent – QObject

Construct a new audio output and attach it to parent. The default audio output device is used with the output format parameters. If format is default-initialized, the format will be set to the preferred format of the audio device.

__init__(audioDeviceInfo[, format=QAudioFormat()[, parent=None]])

Parameters:

audioDeviceInfo – QAudioDevice
format – QAudioFormat
parent – QObject

Construct a new audio output and attach it to parent. The device referenced by audioDevice is used with the output format parameters. If format is default-initialized, the format will be set to the preferred format of audioDevice.

bufferFrameCount()¶

Return type:: int

Returns the audio buffer size in frames.

If called before start() , returns platform default value. If called before start() but setBufferSize() or setBufferFrameCount() was called prior, returns value set by setBufferSize() or setBufferFrameCount(). If called after start(), returns the actual buffer size being used. This may not be what was set previously by setBufferSize() or setBufferFrameCount().

See also

setBufferFrameCount() bufferSize

bufferSize()¶

Return type:: int

Returns the audio buffer size in bytes.

See also

setBufferSize() bufferFrameCount

bytesFree()¶

Return type:: int

Returns the number of free bytes available in the audio buffer.

Note

The returned value is only valid while in ActiveState or IdleState state, otherwise returns zero.

See also

framesFree

elapsedUSecs()¶

Return type:: int

Returns the microseconds since start() was called, including time in Idle and Suspend states.

error()¶

Return type:: Error

Returns the error state.

format()¶

Return type:: QAudioFormat

Returns the QAudioFormat being used.

framesFree()¶

Return type:: int

Returns the number of free frames available in the audio buffer.

Note

The returned value is only valid while in ActiveState or IdleState state, otherwise returns zero.

See also

bytesFree

isNull()¶

Return type:: bool

Returns true is the QAudioSink instance is null, otherwise returns false.

processedUSecs()¶

Return type:: int

Returns the amount of audio data processed since start() was called (in microseconds).

reset()¶

Immediately halts audio output and discards any audio data currently in the buffers. All pending audio data pushed to QIODevice is ignored.

See also

stop()

resume()¶

Resumes processing audio data after a suspend() .

Sets state() to the state the sink had when suspend() was called, and sets error() to QAudioError::NoError. This function does nothing if the audio sink’s state is not SuspendedState .

setBufferFrameCount(framesCount)¶

Parameters:: framesCount – int

Sets the audio buffer size to value in frame count.

Note

This function can be called anytime before start() . Calls to this are ignored after start() . It should not be assumed that the buffer size set is the actual buffer size used - call bufferFrameCount() anytime after start() to return the actual buffer size being used.

See also

bufferFrameCount() setBufferSize

setBufferSize(bytes)¶

Parameters:: bytes – int

Sets the audio buffer size to value in bytes.

Note

This function can be called anytime before start() . Calls to this are ignored after start() . It should not be assumed that the buffer size set is the actual buffer size used - call bufferSize() anytime after start() to return the actual buffer size being used.

See also

bufferSize() setBufferFrameCount

setVolume(volume)¶

Parameters:: volume – float

Sets the output volume to volume.

The volume is scaled linearly from 0.0 (silence) to 1.0 (full volume). Values outside this range will be clamped.

The default volume is 1.0.

Note

Adjustments to the volume will change the volume of this audio stream, not the global volume.

UI volume controls should usually be scaled non-linearly. For example, using a logarithmic scale will produce linear changes in perceived loudness, which is what a user would normally expect from a volume control. See convertVolume() for more details.

See also

volume()

start()¶

Return type:: QIODevice

Returns a pointer to the internal QIODevice being used to transfer data to the system’s audio output. The device will already be open and write() can write data directly to it.

Note

The pointer will become invalid after the stream is stopped or if you start another stream.

If the QAudioSink is able to access the system’s audio device, state() returns IdleState , error() returns NoError and the stateChanged() signal is emitted.

If a problem occurs during this process, error() returns OpenError , state() returns StoppedState and the stateChanged() signal is emitted.

See also

QIODevice interface

start(device)

Parameters:: device – QIODevice

Starts transferring audio data from the device to the system’s audio output. The device must have been opened in the ReadOnly or ReadWrite modes.

If the QAudioSink is able to successfully output audio data, state() returns ActiveState , error() returns NoError and the stateChanged() signal is emitted.

If a problem occurs during this process, error() returns OpenError , state() returns StoppedState and the stateChanged() signal is emitted.

See also

QIODevice interface

state()¶

Return type:: State

Returns the state of audio processing.

stateChanged(state)¶

Parameters:: state – State

This signal is emitted when the device state has changed. This is the current state of the audio output.

Note

The QtAudio namespace was named QAudio up to and including Qt 6.6. String-based connections to this signal have to use QAudio::State as the parameter type: connect(source, SIGNAL(stateChanged(QAudio::State)), ...);

stop()¶

Stops the audio output, detaching from the system resource.

Sets error() to NoError , state() to StoppedState and emit stateChanged() signal.

Note

On Linux, and Darwin, this operation synchronously drains the underlying audio buffer, which may cause delays accordingly to the buffer payload. To reset all the buffers immediately, use the method reset instead.

See also

reset()

suspend()¶

Stops processing audio data, preserving buffered audio data.

Sets error() to NoError , state() to SuspendedState and emits stateChanged() signal.

volume()¶

Return type:: float

Returns the volume between 0.0 and 1.0 inclusive.

See also

setVolume()