The Beat Machine is a free tool for creating "every other beat is missing" remixes -- and many other kind of beat edits, too.

To use The Beat Machine from your browser, try the official web interface at https://beatmachine.branchpanic.me.

TBM is available on PyPI and can be installed with pip install beatmachine.

TBM is built on effects, which modify on the beats of a song. For example, the swap effect swaps beats and can be used to create the "beats 2 and 4 are swapped" sound.

The easiest way to get started is to download some of the presets from the examples directory. You can use them like so:

$ python -m beatmachine -i in.mp3 -e swap_2_4.json -o out.mp3

The CLI reads effects as an array of JSON objects. Each object represents an effect, and effects are applied sequentially. You can either specify this inline or provide a path to a JSON file. For example, to swap beats 2 and 4:

$ python -m beatmachine -i in.mp3 -e '[{"type": "swap", "x_period": 2, "y_period": 4}]' -o out.mp3

Using python -m beatmachine.dump_schema, you can generate a JSON schema that describes the effects array. This includes definitions of all valid effects.

You can also look at the individual classes in beatmachine.effects to see their parameters. For example, here is the source for the remove effect. The JSON object for an effect must have a key called type with the value of __effect_name__. It can have additional keys for parameters defined in __effect_schema__. So, some examples of valid remove effects are { "type": "remove" } and { "type": "remove", "period": 4 }.

(TODO: add a human-readable list of effects and their parameters.)

(TODO: document the --serialize flag, which can be used to speed up repeated processing. See #54.)

The beatmachine.Beats class lets you modify beats using a Python script.

import beatmachine as bm

beats = bm.Beats.from_song('in.mp3')
beats.apply(bm.effects.RemoveEveryNth(2)).save('out.mp3')

This opens up some interesting possibilities, like turning beats into a NumPy array that you can modify further.

import beatmachine as bm
import numpy as np

beats = bm.Beats.from_song('in.mp3')
y = np.flip(beats.to_ndarray())

Be warned that the API is largely untested outside of the core from_song -> apply -> save path.

(TODO: more detailed docs will eventually live on the wiki.)

The default beat detector is powered by CPJKU/madmom. View its license here.