Audio Collager

Attempt to reconstruct target audio using snippets of source audio.

Background

This tool is inspired by the Scott Johnson album John Somebody (this track in particular) in which samples of speech and laughter are cut up, rearranged and looped to create rhythm and melody.

Setup

Install Python >3.9 and the dependencies with poetry:

poetry install

Run

poetry run audio-collage COMMAND [OPTIONS]

Run poetry run audio-collage --help for more details.

Examples

Creating a collage

poetry run audio-collage collage -t target.wav -s source.wav -o collage.wav

Chopping audio

Chop the given file in to snippets of 250 milliseconds

poetry run audio-collage chop -l 250 -f sample.wav -o sample_slices/

Use Cases

Let's begin with two breakbeats:

• Black Heat Zimba Ku (source)
• The Winstons Amen Brother (source)

We want to recreate the drum pattern of the former using audio from the latter.

poetry run audio-collage collage -t docs/audio/breaks/amen_brother.wav -s docs/audio/breaks/black_heat__zimba_ku.wav -o docs/audio/breaks/out/amen_zimba.wav

You can listen to the output here

You can quickly run this example locally using:

poetry run audio-collage example