SSM Tools

This repository contains Jupyter notebooks for creating self-similarity matrix (SSM) representations from audio and symbolic music files.

Environment Setup

The code and notebooks in this repository are optimized for Python 3.11. We recommend setting up a virtual environment using Conda to manage the dependencies. You can create a new environment named ssm_tools by running:

conda create -n ssm_tools python=3.11

Once the environment is created and activated, install the required packages using:

pip install -r requirements.txt

Repository Structure

The repository contains the following files:

• music21_df_ssm.ipynb: A notebook for creating SSM representations based on symbolic music inputs. It can process all formats compatible with music21, using pitch class distribution as the basis for SSM calculation.
• music21_df_ssm_plotly.ipynb: A version of the music21_df_ssm notebook, but with a Plotly implementation for convenient zooming and panning.
• ssm_audio.ipynb: A notebook for creating SSM representations based on audio files. It uses librosa for audio processing and works best with WAV files, featuring MFCC and chroma features for SSM calculation.
• ssm_audio_plotly.ipynb: A version of the ssm_audio.ipynb notebook, but with a Plotly implementation for convenient zooming and panning.
• ssm_detail_vector.ipynb: A tutorial notebook that goes into detail about how MFCC is calculated and how feature vectors are defined.

Getting Started

To get started, clone this repository and navigate to the cloned directory. Activate your ssm_tools environment and install the required packages as described in the Environment Setup section above.