headbang.py is a collection of projects related to beat tracking in metal music.

See more at https://sevagh.github.io/headbang.py

headbang has been written and verified with Python 3.8 on an amd64 machine running Fedora 32 Linux. However, there shouldn't be any problems running it on different machines if the requirements can be successfully installed.

The only sticking point is that the BTrack package is not on pip, and needs to be installed manually. My vendored copy supports a Python 3.8 install:

sevagh:~ $ git clone https://github.com/sevagh/headbang.py
sevagh:~ $ cd misc/vendor/BTrack/modules-and-plug-ins/python-module
sevagh:python-module $ python3.8 setup.py build

# install to your system
sevagh:python-module $ sudo python3.8 setup.py install

# OR #

# install for your local user
sevagh:python-module $ pip3.8 install --user -e .

After installing BTrack for headbang, you also need to build and install openpose from source. I followed the instructions for building and installing the Python module successfully on Python 3.7, Fedora 32, amd64 with no issues.

Once the dependencies are installed, you can install headbang with pip or the setup.py file using your Python package manager of choice:

# demo user install to system python
$ pip install --user -r ./requirements.txt
$
$ headbang-
headbang-beats  headbang-hud    headbang-viz

The Python classes are ConsensusBeatTracker and HeadbangBeatTracker. Example usage:

import multiprocessing
import librosa

from headbang.consensus import ConsensusBeatTracker
from headbang import HeadbangBeatTracker

pool = multiprocessing.Pool()

# you must use 44100 (enforced by the inner beat trackers)

x, _ = librosa.load('mysongwithbeats.wav', sr=44100)
cbt = ConsensusBeatTracker(pool)
hbt = HeadbangBeatTracker(pool)

cbt_beats = cbt.beats(x)
hbt_beats = hbt.beats(x)

The installed tool headbang-beats implements all of the parameters of the consensus and headbang beat trackers in command-line arguments, loads an audio file, gets beat locations, and outputs an audio file with clicks on the beat locations.

headbang-hud is a combined MIR beat tracking and 2D pose estimation project for tracking beats and headbanging motion in videos with music (concert footage, covers, music videos, etc.):

headbang-viz creates a visualization of the competing beat trackers and renders the animation in an mp4 video alongside the input audio: