SSL Pitch Estimation

This repository provides a tutorial on a simplified implementation of PESTO, a self-supervised model for pitch estimation [1]. It is mainly intended as a teaching material.

The main tutorial can be accessed in the notebook ssl-pitch-estimation-tutorial.ipynb.

Run in Colab

Simply click the button above and be sure to run the install dependencies code cell provided.

Run Locally

Clone the repository and install the dependencies listed in pyproject.toml

We recommend using uv to create a virtual environment and to install dependencies.

uv venv
source .venv/bin/activate
uv sync

References

We encourage you to check out the original implementation from which this notebook is based and the paper [1].

Note that there is also a pip-installable minimal code implementation for inference only in this fast implementation repository.

[1] A. Riou, “PESTO: Real‑Time Pitch Estimation with Self‑Supervised Transposition‑Equivariant Objective”, Transactions of the International Society for Music Information Retrieval, vol. 8, no. 1, p. 334–352, 2025, doi: 10.5334/tismir.251.