Music Mood Analysis

This analysis tool extracts tempo and tonality from music using digital signal processing techniques to analyse the mood of a piece of music.

The tool comes with a graphical user interface, which may be used to select the file to be analysed, configure the analysis parameters, run the analysis and finally inspect the analysis results.

Getting started

To run the package, install it using pip, then run it using:

python -m pip install music_mood_analysis
python -m music_mood_analysis

Command line interface

The music analysis tool can also be used directly with its command line interface, bypassing the graphical interface completely:

usage: __main__.py [-h] [--tempo-analysis {fft,lmv}] [--chunk-size CHUNK_SIZE]
                   [--downsample DOWNSAMPLE] [--min-bps MIN_BPS]
                   [--max-bps MAX_BPS] [--lmv-decay LMV_DECAY]
                   filepath

positional arguments:
  filepath              The path to the audio file to be analysed

optional arguments:
  -h, --help            show this help message and exit
  --tempo-analysis {fft,lmv}, -t {fft,lmv}
                        The tempo analysis method to be used
  --chunk-size CHUNK_SIZE, -c CHUNK_SIZE
                        The size (in seconds) of audio data chunk to be
                        analysed at once
  --downsample DOWNSAMPLE, -d DOWNSAMPLE
                        The factor by which to downsample audio data during
                        analysis
  --min-bps MIN_BPS     The minimum amount of beats per second to be
                        considered during tempo analysis
  --max-bps MAX_BPS     The maximum amount of beats per second to be
                        considered during tempo analysis
  --lmv-decay LMV_DECAY
                        The amplitude decay of local maxima during lmv tempo
                        analysis

To bring up this help message, run:

python -m music_mood_analysis -h

Real-time

The tool used to be configured to be used in a real-time analysis setting for analysis of ambient music. In the current configuration of the script, this is currently not possible anymore; however, it should not be too hard to re-instate the real-time functionality if required.

The current configuration uses a .wav file to test the analysis functionality. It streams several data chunks per second from the file, which could be either be read from a file or be provided by a funnel script reading audio data directly from audio hardware. Such a script is not included in this repository, as it is hardware-specific.

Contributions

To contribute to this repository, please read the contribution guidelines.

Python

Written in Python 3.8.3.

License

This repository is open-source software available under the AGPL-3.0 License.

Contact

Please raise an issue for code changes. To reach out, please send an email to richard@baltrusch.net.