The goal of this component is to perform the synchronization of the all audio files in an existing dataset.
On top of the txt files listed bellow, both options cut the audio files in the dataset according to the segments and save them in “sounds_cut/”.
Example: “sounds_cut/1/10.0” has all cut audio files that exist in cluster 1 starting on 10 seconds until the time stamp where the next segments starts.
For testing and validation purposes, a dataset with recordings crawled from YouTube was used. Each song had one professional recording (BBC) and a set of user-recordings.
Number of songs: 10 Number of samples: 90