Octave provides a few functions for dealing with audio data. An audio `sample' is a single output value from an A/D converter, i.e., a small integer number (usually 8 or 16 bits), and audio data is just a series of such samples. It can be characterized by three parameters: the sampling rate (measured in samples per second or Hz, e.g., 8000 or 44100), the number of bits per sample (e.g., 8 or 16), and the number of channels (1 for mono, 2 for stereo, etc.).
There are many different formats for representing such data. Currently,
only the two most popular, linear encoding and mu-law
encoding, are supported by Octave. There is an excellent FAQ on audio
formats by Guido van Rossum email@example.com which can be found at any
FAQ ftp site, in particular in the directory
/pub/usenet/news.answers/audio-fmts of the archive site
Octave simply treats audio data as vectors of samples (non-mono data are not supported yet). It is assumed that audio files using linear encoding have one of the extensions lin or raw, and that files holding data in mu-law encoding end in au, mu, or snd.
Convert audio data from linear to mu-law. Mu-law values use 8-bit unsigned integers. Linear values use n-bit signed integers or floating point values in the range -1 ≤ x ≤ 1 if n is 0.
If n is not specified it defaults to 0, 8, or 16 depending on the range of values in x.
Convert audio data from mu-law to linear. Mu-law values are 8-bit unsigned integers. Linear values use n-bit signed integers or floating point values in the range -1≤y≤1 if n is 0.
If n is not specified it defaults to 0.
Load audio data from the file name.ext into the vector x.
The extension ext determines how the data in the audio file is interpreted; the extensions lin (default) and raw correspond to linear, the extensions au, mu, or snd to mu-law encoding.
The argument bps can be either 8 (default) or 16, and specifies the number of bits per sample used in the audio file.
Save a vector x of audio data to the file name.ext. The optional parameters ext and bps determine the encoding and the number of bits per sample used in the audio file (see
loadaudio); defaults are lin and 8, respectively.
The following functions for audio I/O require special A/D hardware and operating system support. It is assumed that audio data in linear encoding can be played and recorded by reading from and writing to /dev/dsp, and that similarly /dev/audio is used for mu-law encoding. These file names are system-dependent. Improvements so that these functions will work without modification on a wide variety of hardware are welcome.
Play the audio file name.ext or the audio data stored in the vector x.
Record sec seconds of audio input into the vector x. The default value for sampling_rate is 8000 samples per second, or 8kHz. The program waits until the user types <RET> and then immediately starts to record.
Execute the shell command ‘mixer’, possibly with optional arguments w_type and value.
Load the RIFF/WAVE sound file filename, and return the samples in vector y. If the file contains multichannel data, then y is a matrix with the channels represented as columns. — Function File: [y, Fs, bps] = wavread (filename)
Return the number of samples (n) and channels (ch) instead of the audio data.
See also: wavwrite.
Write y to the canonical RIFF/WAVE sound file filename with sample rate Fs and bits per sample bps. The default sample rate is 8000 Hz with 16-bits per sample. Each column of the data represents a separate channel.
See also: wavread.