Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

Python audio coding classes - for dsp and sonification

Project description

Version PyPI GitHub commits License Build Status Travis Build status AppVeyor


What is pyA?

pyA is a package to support creation and manipulation of audio signals with Python. It uses numpy arrays to store and compute audio signals.

It provides:

  • Asig - a versatile audio signal class
    • Ugen - a subclass of Asig, which offers unit generators such as sine, square, swatooth, noise
  • Aserver - an audio server class for queuing and playing Asigs
  • Arecorder - an audio recorder class
  • Aspec - an audio spectrum class, using rfft as real-valued signals are always implied
  • Astft - an audio STFT (short-term Fourier transform) class

pyA can be used for

  • multi-channel audio processing
  • auditory display and sonification
  • sound synthesis experiment
  • audio applications in general such as games or GUI-enhancements
  • signal analysis and plotting
  • at this time more suitable for offline rendering than realtime.

Authors and Contributors

  • Thomas Hermann, Ambient Intelligence Group, Faculty of Technology, Bielefeld University (author and maintainer)
  • Jiajun Yang, Ambient Intelligence Group, Faculty of Technology, Bielefeld University (co-author)
  • Alexander Neumann, Neurocognitions and Action - Biomechanics, Bielefeld University
  • Contributors will be acknowledged here, contributions are welcome.


Note: pya can be installed using pip. But pya uses PyAudio for audio playback and record, and PyAudio 0.2.11 has yet to fully support Python 3.7. So using pip install with Python 3.7 may encounter issues such as portaudio. Solution

Use pip to install pya via

pip install pya

Note: pyA requires PyAudio, for playback and record, which in turn requires portaudio. Since PyAudio 0.2.11 has yet to fully support Python 3.7, you maybe encounter portaudio related error under Python 3.7. Lower versions of Python should be fine. A few solutions are:

  1. Anaconda can install non-python packages, so that the easiest way (if applicable) would be to

    conda install pyaudio

  2. For Mac users, you can brew install portaudio beforehand.

  3. For Linux users, try sudo apt-get install portaudio19-dev or equivalent to your distro.

  4. For Windows users, you can install PyAudio wheel at:

See pyaudio installation

A simple example


from pya import *
s = Aserver(bs=1024)
Aserver.default = s  # to set as default server

Create an Asig signal:

A 1s / 440 Hz sine tone at sampling rate 44100 as channel name 'left':

import numpy as np
signal_array = np.sin(2 * np.pi * 440 * np.linspace(0, 1, 44100))
atone = Asig(signal_array, sr=44100, label='1s sine tone', cn=['left'])

Other ways of creating an Asig object:

asig_int = Asig(44100, sr=44100)  # zero array with 44100 samples
asig_float = Asig(2., sr=44100)  # float argument, 2 seconds of zero array
asig_str = Asig('./song.wav')  # load audio file
asig_ugen = Ugen().square(freq=440, sr=44100, dur=2., amp=0.5)  # using Ugen class to create common waveforms

Audio files are also possible using the file path. WAV should work without issues. MP3 is supported but may raise error if FFmpeg.

Key attributes

  • atone.sig --> The numpy array containing the signal is
  • --> the sampling rate
  • --> the list of custom defined channel names
  • atone.label --> a custom set identifier string

Play signals  

play() uses Aserver.default if server is not specified

Plotting signals

to plot the first 1000 samples:


to plot the magnitude and phase spectrum:


to plot the spectrum via the Aspec class


to plot the spectrogram via the Astft class


Selection of subsets

  • Asigs support multi-channel audio (as columns of the signal array)
    • a1[:100, :3] would select the first 100 samples and the first 3 channels,
    • a1[{1.2:2}, ['left']] would select the channel named 'left' using a time slice from 1

Method chaining

Asig methods usually return an Asig, so methods can be chained, e.g

atone[{0:1.5}].fade_in(0.1).fade_out(0.8).gain(db=-6).plot(lw=0.1).play(rate=0.4, onset=1)

Learning more

  • Please check the examples/pya-examples.ipynb for more examples and details.


  • Please get in touch with us if you wish to contribute. We are happy to be involved in the discussion of new features and to receive pull requests.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for pya, version 0.3.1
Filename, size File type Python version Upload date Hashes
Filename, size pya-0.3.1-py3-none-any.whl (41.8 kB) File type Wheel Python version py3 Upload date Hashes View hashes
Filename, size pya-0.3.1.tar.gz (1.5 MB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page