 |
 |
 |
|---|---|---|
| https://chordify.net/ | http://www.tagtraum.com/ | https://www.audiolabs-erlangen.de/ |
This repository contains supplemental material for the paper "Towards Automatically Correcting Tapped Beat Annotations for Music Recordings" by Jonathan Driedger, Hendrik Schreiber, W. Bas de Haas, and Meinard Müller, presented at ISMIR 2019 in Delft, the Netherlands.
When you use the TapCorrect procedure or the dataset, please cite
@inproceedings{DriedgerSHM19_TapCorrect_ISMIR,
author = {Jonathan Driedger and
Hendrik Schreiber and
W. Bas de Haas and
Meinard M{\"u}ller},
title = {Towards Automatically Correcting Tapped Beat Annotations for Music Recordings},
booktitle = {Proceedings of the 20th International Society for Music Information
Retrieval Conference ({ISMIR})},
year = {2019},
address = {Delft, The Netherlands},
month = {November}
}
A common method to create beat annotations for music recordings is to let a human annotator tap along with them. However, this method is problematic due to the limited human ability to temporally align taps with audio cues for beats accurately. In order to create accurate beat annotations, it is therefore typically necessary to manually correct the recorded taps in a subsequent step, which is a cumbersome task. In this work we aim to automate this correction step by "snapping" the taps to close-by audio cues - a strategy that is often used by beat tracking algorithms to refine their beat estimates. The main contributions of this paper can be summarized as follows. First, we formalize the automated correction procedure mathematically. Second, we introduce a novel visualization method that serves as a tool to analyze the results of the correction procedure for potential errors. Third, we present a new dataset consisting of beat annotations for 101 music recordings. Fourth, we use this dataset to perform a listening experiment as well as a quantitative study to show the effectiveness of our snapping procedure.
The folder python contains our implementation of the automatic tap correction procedure as described in the paper. The file __main__.py includes an example of how to execute the precedure.
The folder dataset contains the original taps, the automatically corrected taps, and the fully corrected taps, as well as metadata and YouTube links to all 101 music recordings used in the paper. Furthermore, downbeats were annotated for all recordings. Annotations are provided in both csv as well as jams format.
In the csv subfolder, you find one folder per item. Each folder then contains four csv-files:
00-meta.csvcontaining information about artist, title, track duration, and YouTube link.01-original_taps.csvcontaining a list of taps as created by the human annotator using Sonic Visualizer.02-automatically_corrected_taps.csvcontaining a list of taps created by automatically correcting the original taps using our proposed TapCorrect procedure.03-fully_corrected_taps.csvcontaining a list of taps created by manually correcting the automatically corrected taps using Sonic Visualizer.
All taps are given in the form timestamp,"beat_count".
The jams subfolder contains 101 JAMS files. Each file contains the meta data as well as the three annotations listed above.
The folder listening_experiment contains a spreadsheet with the detailed responses of our listening experiment's participants.
- The TapCorrect dataset is published under a Creative Commons Attribution-ShareAlike 4.0 International license.
- The sourcecode for the TapCorrect procedure is published under a GNU Lesser General Public License V3.0 license.
In case you have any questions or feedback, please write to jonathan@chordify.net