Fast Similarity Matrix Profile for Music Analysis and Exploration
Diego F. Silva, Chin-Chia M. Yeh, Yan Zhu, Gustavo E. A. P. A. Batista, Eamonn Keogh
Abstract: Most algorithms for music data mining and retrieval analyze the similarity between feature sets extracted from the raw audio. A conventional approach to assess similarities within or between recordings is to create similarity matrices. However, this method requires quadratic space for each comparison and typically requires costly post-processing of the matrix. We have recently proposed SiMPle, a powerful representation based on subsequence similarity join, which is applicable in several music analysis tasks. In this paper, we propose SiMPle-Fast a highly efficient method for exact computation of SiMPle that is up to one order of magnitude faster than SiMPle. Furthermore, we demonstrate the utility of SiMPle-Fast in cover music recognition and thumbnailing tasks and show that our method is significantly faster and more accurate than the state-of-the-art.
Data
You can find the chroma-based (CENS) features for the YouTube Covers dataset from here. For the other dataset, please contact us.
Source codes
Here, you can find all the (Matlab) code we used in our experiments. If you prefer, you can get the codes separated by task:
In any case, you need the code for either SiMPle-Fast for self-join or for AB-join.
Contact the authors
We appreciate your interest in or work. If you have any question or suggestion, please, be in touch with any of the authors.
Diego Furtado Silva
Webpage: www.dc.ufscar.br/~diego
Email: diegofsilva at icmc dot usp dot br
Chin-Chia Michael Yeh
Webpage: http://www.cs.ucr.edu/~myeh003/
Email: myeh003 at ucr dot edu
Yan Zhu
Webpage: http://www.cs.ucr.edu/~yzhu015/
Email: yzhu015 at ucr dot edu
Gustavo E. A. P. A. Batista
Webpage: www.icmc.usp.br/~gbatista/
Email: gbatista at icmc dot usp dot br
Eamonn Keogh
Webpage: www.cs.ucr.edu/~eamonn/
Email: eamonn at cs dot ucr dot edu