MR20VI

Our dataset for EVAL and DEV used in the Vocal Imitation Challenge hosted at the AES AIMLA conference 2025. You will find the 2 .zip folders and the pdf file describing the dataset. Each csv on the .zip file has: • Label: An identifier for the sample (numerical value). • Class: The sound category or class name (e.g., Birds, CarHorn), indicating the type of sound being imitated. • Items: The filename of the original reference sound (e.g., agapornis chirping 2.wav), representing the sound that participants were asked to vocally imitate. • Query 1, Query 2, Query 3: Filenames of the vocal imitation recordings (e.g., Birds1-01.wav) corresponding to the original reference sound. These are the actual sound files generated as imitations of the Items. Matching Relationship Each row in the CSV file maps one reference sound (Items) to up to three corresponding vocal imitation files (Query 1, Query 2, Query 3). Multiple rows may have the same reference sound but different imitation files, indicating that the same sound may have been imitated multiple times by different participants.