Audiolab is a python package to read/write audio files from numpy arrays. Matlab have functions such as wavread, wavwrite, soundsc, etc... The goal of audiolab is to give those capabilities to the scipy environment by wrapping the excellent library libsndfile from Erik de Castro Lopo. Audiolab supports all format supported by libsndfile, including wav, aiff, ircam, flac (an open source lossless compressed format) and ogg vorbis (an open source compressed format - similar to MP3 without the license issues); see here for a complete list.
The main features of audiolab are:
- reading all formats supported by sndfile directly into numpy arrays.
- writing all formats supported by sndfile directly from numpy arrays.
- A matlab-like API (e.g. wavread) for some file formats. Wav, aiff, flac, au, and ogg formats are supported.
- A play function to output data from numpy array into the sound device of your computer (Only ALSA for linux and CoreAudio for Mac OS X is implemented ATM).
Note: The exact features and supported file format depend on your installed version of libsndfile and your platform. Audio playback is only implemented for Linux and Mac OS X.
Note: The library is still in flux: the API can still change before the 1.0 version. That being said, the version 0.10 was a major change, and I don’t expect any more major changes from now on.
Note: The online version of this document is not always up to date. The pdf included in the package is the reference, and always in sync with the package. If something does not work, please refer first to the pdf included in the package.