Our research focuses on Music Information Retrieval (MIR), specifically music similarity and music recognition. This page presents material and data related to our work.
In our paper Neural Network for Chords Recognition (JIM 2012), we presented a database of guitar chords recorded both in studio (anechoic chamber) and noisy environments. The database contains 2000 chords splitted up in 10 classes, giving up to 200 chords per chord type. The files are stored in raw WAV 16 bits mono 44100Hz format. The database is available for download here.