28.2.8  Extracting samples from audio clips

The channel_data command gets samples from an audio clip.

• channel_data takes one mandatory argument and up to two optional argument:
  • A, an audio clip.
  • Optionally options, which can be the following (the order is unimportant):
    • n, a postive integer (channel number, zero-based) or matrix, the symbol.
    • range=[offset,length] for nonnegative integer offset and positive integer length, or range=a..b for floating point numbers a and b.
    Instead of channel number n, the symbol left resp. right may be used to specify the left resp. right channel.
• channel_data(A ⟨,options⟩) returns data from the channels as a sequence of lists. The returned sample values are all within the interval [−1.0,1.0], i.e. the amplitude of the returned signal is relative. The maximum possible amplitude is represented by the value 1.0.
  • With no options, data from all channels are returned as a sequence of lists.
  • With the option n (the channel number) or if there is only one channel, only data from this channel is returned in a single list.
  • With the option matrix, the lists representing the channel data are returned as the rows of a matrix.
  • With the option range=[offset,length], the block of length samples is extracted starting at offset.
  • With the option range=a..b, with floating point numbers a and b, then a and b are bounds in seconds.
• Alternatively, channel data can be extracted by using the [] operator on an audio clip A with one to three arguments:
  • n, a channel number (left and right are allowed).
  • Optionally, offset, an integer specifying the offset in samples (by default, offset=0).
  • Optionally, len, a positive integer (by default, len=L−offset, where L is the length of A in samples).
  Instead of specifying offset and len parameters, you can enter an interval t₁..t₂ as the second argument. In that case, the channel data between timestamps t₁ and t₂ (in seconds) is returned. The floor command is used for rounding the corresponding offsets during conversion to sample units.

Example

Assuming that the directory sounds contains example.wav, a wav file with three seconds of stereo sound, input:

Output: A list L resp. R containing data between 1.2 and 1.5 seconds in the left resp. right channel of the original file.

Output: A list of 100 samples from the left channel, starting from the sample with index 27400.

Output: A list of samples from the left channel falling between 1.2 and 1.5 seconds.