28.2.17  Joining audio clips together

The splice command joins two audio clips together with an overlap and optionally a crossfade.

• splice takes two mandatory arguments and one or two optional arguments:
  • A, an audio clip.
  • B, an audio clip with the same number of channels, bit depth and sample rate as A.
  • Optionally, len, a positive integer defining the overlap size (by default, len is the minimum of 0.5· s, L_A/2, and L_B/2, where L_A and L_B are the lengths of A and B in samples and s is the sample rate).
  • Optionally, p, a real number defining the crossfade curve (by default, p is computed automatically, see below).
• splice(A,B ⟨,len,p ⟩) returns the audio clip obtained by splicing A and B with an overlap of len samples and a crossfade which is computed using the function f(x)=x^p, x∈[0,1], for p>0.
• For positively correlated audio data on the overlap, it is best to set p=1.0 (constant gain crossfade), while for zero-correlated data p=0.5 (constant power crossfade) works better. If p is not specified, it is computed automatically in [0.5,1.0] such that p−0.5 is proportional to the absolute value of the cosine distance between the overlapping data. Setting p=0 disables crossfading and thus the overlapped data is simply added together.

Example

With the audio file gettysburg.wav downloaded from here:

To remove the portion of the speech between 5.6 and 10.5 seconds, splice the portions on the left and right by using an overlap of size 0.25 seconds. To visualize the audio and the selected portion, you can enter:

Extract the portions from 0 to t₁ and from t₂ to the end and splice them with an automatic crossfade:

Use playsnd (see Section 28.2.14) to hear the result.