pvspeaker 1.0.1

Speaker library for Picovoice.

PvSpeaker Binding for Python

PvSpeaker

PvSpeaker is an easy-to-use, cross-platform audio player designed for real-time speech audio processing. It allows developers to send raw PCM frames to an audio device's output stream.

Compatibility

• Python 3.8+
• Runs on Linux (x86_64), macOS (x86_64 and arm64), Windows (x86_64), and Raspberry Pi (3, 4, 5).

Installation

pip3 install pvspeaker

Usage

Initialize and start PvSpeaker:

from pvspeaker import PvSpeaker

speaker = PvSpeaker(
    sample_rate=22050,
    bits_per_sample=16,
    buffer_size_secs=20,
    device_index=0)

speaker.start()

(or)

Use get_available_devices() to get a list of available devices and then initialize the instance based on the index of a device:

from pvspeaker import PvSpeaker

devices = PvSpeaker.get_available_devices()

speaker = PvSpeaker(
    sample_rate=22050,
    bits_per_sample=16,
    buffer_size_secs=20,
    device_index=0)

speaker.start()

Write frames of audio:

def get_next_audio_frame():
    pass

speaker.write(get_next_audio_frame())

When all frames have been written, run flush() to wait for all buffered pcm data to be played:

speaker.flush()

To stop playing audio, run stop():

speaker.stop()

Note that in order to stop the audio before it finishes playing, stop must be run on a separate thread from flush.

Once you are done (i.e. no longer need PvSpeaker to write and/or play PCM), free the resources acquired by PvSpeaker by calling delete. Be sure to first call stop if the audio is still playing. Otherwise, if the audio has already finished playing, you do not have to call stop before delete:

speaker.delete()

Demos

pvspeakerdemo provides command-line utilities for playing audio from a file.