TinySOL: Isolated musical notes audio dataset
A Balanced Audio Dataset for Music Information Retrieval
By [source]
About this dataset
TinySOL is a complete audio dataset of isolated musical notes from 14 instruments, collected from Ircam in Paris and released under the Creative Commons Attribution 4.0 International license. The instruments include Bass Tuba, French Horn, Trombone, Trumpet in C, Accordion, Contrabass, Violin, Viola, Violoncello - a stringed family comprising some of the most iconic sounds in classical music - as well as Bassoon, Clarinet in B-flat; Flute; Oboe; and Alto Saxophone.
Using TinySOL to understand music information retrieval can be valuable on many levels. The application possibilities range from music creation to audio search engines or even AI-driven study of classical compositions. All TinySOL files come with associated metadata that describe the pitch and dynamics of each note as well as its instrument family and performance technique (abbreviations included) - making this dataset an incredibly useful resource for any kind of MIR research or experimental project!
More Datasets
For more datasets, click here.
Featured Notebooks
- 🚨 Your notebook can be here! 🚨!
How to use the dataset
This dataset provides a comprehensive collection of isolated musical notes of 14 instruments recorded at Ircam in Paris, under the Creative Commons Attribution 4.0 International license. It can be used for various music information retrieval applications, such as instrument recognition systems, automatic genre classification algorithms, and audio synthesis.
Research Ideas
- Training and Retraining AI/ML algorithms for Source Separation of Musical Instruments: By using this balanced dataset, researchers can experiment with various source separation techniques and architectures to develop AI and ML models that can robustly isolate different types of instruments from an audio track.
- Generating Isolated Instrument Samples for Synthesizers & Sound Banks: Using the isolated notes in this dataset, developers can generate a large library of samples for use in software synthesisers or hardware samplers, allowing producers to get the most accurate sound possible from their instruments.
- Automatically Tagging Different Labelled Notes in Audio Files: By leveraging the metadata included within each audio file in this dataset, machine learning algorithms could be used to automatically tag different labelled notes within any given audiofile, enabling quicker annotation workflows throughout the music industry
Acknowledgements
If you use this dataset in your research, please credit the original authors.
Data Source
License
License: CC0 1.0 Universal (CC0 1.0) - Public Domain Dedication
No Copyright - You can copy, modify, distribute and perform the work, even for commercial purposes, all without asking permission. See Other Information.
Columns
File: TinySOL_metadata.csv
| Column name |
Description |
| Path |
The file path of the audio sample. (String) |
| Fold |
The fold number of the audio sample. (Integer) |
| Family |
The family of instrument the audio sample belongs to. (String) |
| Instrument (abbr.) |
The abbreviation of the instrument the audio sample belongs to. (String) |
| Instrument (in full) |
The full name of the instrument the audio sample belongs to. (String) |
| Technique (abbr.) |
The abbreviation of the technique used to play the audio sample. (String) |
| Technique (in full) |
The full name of the technique used to play the audio sample. (String) |
| Pitch |
The pitch of the audio sample. (Integer) |
| Dynamics |
The dynamics of the audio sample. (Integer) |
| Needed digital retuning |
The amount of digital retuning needed for the audio sample. (Integer) |
Acknowledgements
If you use this dataset in your research, please credit the original authors.
If you use this dataset in your research, please credit .