Basic Pitch is a Python library for Automatic Music Transcription (AMT), using lightweight neural network developed by Spotify's Audio Intelligence Lab. It's small, easy-to-use, and pip install-able.
Basic Pitch may be simple, but it's is far from "basic"! basic-pitch is efficient and easy to use, and its multipitch support, its ability to generalize across instruments, and its note accuracy competes with much larger and more resource-hungry AMT systems.
Provide a compatible audio file and basic-pitch will generate a MIDI file, complete with pitch bends. Basic pitch is instrument-agnostic and supports polyphonic instruments, so you can freely enjoy transcription of all your favorite music, no matter what instrument is used. Basic pitch works best on one instrument at a time.
Research Paper
This library was released in conjunction with Spotify's publication at ICASSP 2022. You can read more about this research in the paper, A Lightweight Instrument-Agnostic Model for Polyphonic Note Transcription and Multipitch Estimation.
If you use this library in academic research, consider citing it:
@inproceedings{2022_BittnerBRME_LightweightNoteTranscription_ICASSP,
author= {Bittner, Rachel M. and Bosch, Juan Jos\'e and Rubinstein, David and Meseguer-Brocal, Gabriel and Ewert, Sebastian},
title= {A Lightweight Instrument-Agnostic Model for Polyphonic Note Transcription and Multipitch Estimation},
booktitle= {Proceedings of the IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)},
address= {Singapore},
year= 2022,
}
Demo
If, for whatever reason, you're not yet completely inspired, or you're just like so totally over the general vibe and stuff, checkout our snappy demo website, basicpitch.io, to experiment with our model on whatever music audio you provide!
Installation
basic-pitch is available via PyPI. To install the current release:
pip install basic-pitch
To update Basic Pitch to the latest version, add --upgrade to the above command.
Compatible Environments:
- MacOS, Windows and Ubuntu operating systems
- Python versions 3.7, 3.8, 3.9
Usage
Model Prediction
Command Line Tool
This library offers a command line tool interface. A basic prediction command will generate and save a MIDI file transcription of audio at the <input-audio-path> to the <output-directory>:
basic-pitch <output-directory> <input-audio-path>
To process more than one audio file at a time:
basic-pitch <output-directory> <input-audio-path-1> <input-audio-path-2> <input-audio-path-3>
Optionally, you may append any of the following flags to your prediction command to save additional formats of the prediction output to the <output-directory>:
--sonify-midito additionally save a.wavaudio rendering of the MIDI file--save-model-outputsto additionally save raw model outputs as an NPZ file--save-note-eventsto additionally save the predicted note events as a CSV file
To discover more parameter control, run:
basic-pitch --help
Programmatic
predict()
Import basic-pitch into your own Python code and run the predict functions directly, providing an <input-audio-path> and returning the model's prediction results:
from basic_pitch.inference import predict
from basic_pitch import ICASSP_2022_MODEL_PATH
model_output, midi_data, note_activations = predict(<input-audio-path>)
<minimum-frequency>&<maximum-frequency>(floats) set the maximum and minimum allowed note frequency, in Hz, returned by the model. Pitch events with frequencies outside of this range will be excluded from the prediction results.model_outputis the raw model inference outputmidi_datais the transcribed MIDI data derived from themodel_outputnote_eventsis a list of note events derived from themodel_output
predict() in a loop
To run prediction within a loop, you'll want to load the model yourself and provide predict() with the loaded model object itself to be used for repeated prediction calls, in order to avoid redundant and sluggish model loading.
import tensorflow as tf
from basic_pitch.inference import predict
from basic_pitch import ICASSP_2022_MODEL_PATH
basic_pitch_model = tf.saved_model.load(str(ICASSP_2022_MODEL_PATH))
for x in range():
...
model_output, midi_data, note_activations = predict(
<loop-x-input-audio-path>,
basic_pitch_model,
)
...
predict_and_save()
If you would like basic-pitch orchestrate the generation and saving of our various supported output file types, you may use predict_and_save instead of using predict directly:
from basic_pitch.inference import predict_and_save
predict_and_save(
<input-audio-path-list>,
<output-directory>,
<save-midi>,
<sonify-midi>,
<save-model-outputs>,
<save-note-events>,
)
where:
<input-audio-path-list>&<output-directory>- directory paths for
basic-pitchto read from/write to.
- directory paths for
<save-midi>- bool to control generating and saving a MIDI file to the
<output-directory>
- bool to control generating and saving a MIDI file to the
<sonify-midi>- bool to control saving a WAV audio rendering of the MIDI file to the
<output-directory>
- bool to control saving a WAV audio rendering of the MIDI file to the
<save-model-outputs>- bool to control saving the raw model output as a NPZ file to the
<output-directory>
- bool to control saving the raw model output as a NPZ file to the
<save-note-events>- bool to control saving predicted note events as a CSV file
<output-directory>
- bool to control saving predicted note events as a CSV file
Model Input
Supported Audio Codecs
basic-pitch accepts all sound files that are compatible with its version of librosa, including:
.mp3.ogg.wav.flac.m4a
Mono Channel Audio Only
While you may use stereo audio as an input to our model, at prediction time, the channels of the input will be down-mixed to mono, and then analyzed and transcribed.
File Size/Audio Length
This model can process any size or length of audio, but processing of larger/longer audio files could be limited by your machine's available disk space. To process these files, we recommend streaming the audio of the file, processing windows of audio at a time.
Sample Rate
Input audio maybe be of any sample rate, however, all audio will be resampled to 22050 Hz before processing.
Contributing
Contributions to basic-pitch are welcomed! See CONTRIBUTING.md for details.
Copyright and License
basic-pitch is Copyright 2022 Spotify AB.
This software is licensed under the Apache License, Version 2.0 (the "Apache License"). You may choose either license to govern your use of this software only upon the condition that you accept all of the terms of either the Apache License.
You may obtain a copy of the Apache License at:
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software distributed under the Apache License or the GPL License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the Apache License for the specific language governing permissions and limitations under the Apache License.
Essentially, there were a few notes played by 62 different pianos, despite the original performance being played on 1.
![[feature request] get f0 pitch as 1D array instead of MIDI notes](https://avatars.githubusercontent.com/u/5549677?v=4)
13 Oct 3, 2022
4 Dec 15, 2022
190 Dec 17, 2022
3 Oct 20, 2021
921 Jan 5, 2023
4 Sep 2, 2022
8 Dec 30, 2022
4 Mar 10, 2022
1 Nov 19, 2021
6 Nov 23, 2022
2 Dec 20, 2022
239 Jan 3, 2023
341 Dec 26, 2022
215 Nov 16, 2022
419 Dec 26, 2022
34 Jun 29, 2022
166 Jan 8, 2023
105 Dec 26, 2022