This repository contains an annotated dataset of guitar solo segments present in 60 popular rock songs and the code used to develop a machine learning model to detect guitar solos automatically. This work titled A Dataset and Method for Guitar Solo Detection in Rock Music, was published as a Conference Paper at the 2017 AES (Audio Engineering Society) International Conference on Semactic Audio
Pati, Kumar Ashis, and Alexander Lerch. "A Dataset and Method for Guitar Solo Detection in Rock Music." Audio Engineering Society Conference: 2017 AES International Conference on Semantic Audio. Audio Engineering Society, 2017.
@inproceedings{pati2017dataset,
title={A Dataset and Method for Guitar Solo Detection in Rock Music},
author={Pati, Kumar Ashis and Lerch, Alexander},
booktitle={Audio Engineering Society Conference: 2017 AES International Conference on Semantic Audio},
year={2017},
organization={Audio Engineering Society}
}
Please cite the publication if you are using the dataset and/or the code in this repository.
A blog post summarizing the above paper can be found here and the full paper is available here.
The folder structure is as follows:
- Dataset: which contains
song_lists.txt
which lists the names and discog information of the songs present in the dataset.Annotations
folder which contains the individual.txt
files containing the start-time and durations (in seconds) of guitar solo segments present in those songs.
- Code: which contains the scripts and functions to extract different features and train and evaluate an SVM (Support Vector Machine) based model.
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.