TEMPO AND BEAT ESTIMATION OF MUSICAL SIGNALS

Tempo estimation is fundamental in automatic music processing and in many multimedia applications. This paper presents an automatic tempo tracking system that processes audio recordings and determines the beats per minute and temporal beat location. The concept of spectral energy flux is defined and leads to an efficient note onset detector. The algorithm involves three stages: a frontend analysis that efficiently extracts onsets, a periodicity detection block and the temporal estimation of beat locations. The performance of the proposed method is evaluated using a large database of 489 excerpts from several musical genres. The global recognition rate is 89.7 %. Results are discussed and compared to other tempo estimation systems.