On Musical Onset Detection via the S-Transform

Musical onset detection is a key component in any beat tracking system. Existing algorithms for onset detection usually rely on short time Fourier transform (STFT). Of numerous existing methods to detect musical onsets, a common limitation is the genre-specific performance. For example, accurate onset locations are not usually identified for genres such as classical music and opera music. In this paper, we propose an algorithm to localize onset components in music by using the classic S-transform. Numerical results are provided to see the behavior of the algorithm under different genres of music. They show that the proposed algorithm can outperform the STFT based algorithms, irrespective of the genres of music.