Feature-Based Timbral Characterization of Historical and Modern Violins

Violin timbre is a very complex case of study. The sound properties that distinguish an historical violin from a modern one are still not clear. The purpose of this study is to understand what are these properties, by means of feature-based analysis. We extract audio features related to timbre and we exploit feature selection techniques in order to investigate what are the most characterizing ones. We compare different feature selection algorithms and we illustrate how we applied their outcome to a classification task with historical and modern instruments. Results show that the classification performance improves when using the selected features.

[1]  Ewa Łukasik.  Matching violins in terms of timbral features , 2014 .

[2]  Petri Toiviainen,et al.  MIR in Matlab (II): A Toolbox for Musical Feature Extraction from Audio , 2007, ISMIR.

[3]  Gilles Louppe,et al.  Independent consultant , 2013 .

[4]  Larry A. Rendell,et al.  The Feature Selection Problem: Traditional Methods and a New Algorithm , 1992, AAAI.

[5]  M. Slaney,et al.  PERCEPTUAL DISTANCE IN TIMBRE SPACE , 2005 .

[6]  György Fazekas,et al.  Training-Based Semantic Descriptors Modeling for Violin Quality Sound Characterization , 2015 .

[7]  J. Poitevineau,et al.  Soloist evaluations of six Old Italian and six new violins , 2014, Proceedings of the National Academy of Sciences.

[8]  P. Juslin,et al.  Cue Utilization in Communication of Emotion in Music Performance: Relating Performance to Perception Studies of Music Performance , 2022 .

[9]  R. Plomp,et al.  Tonal consonance and critical bandwidth. , 1965, The Journal of the Acoustical Society of America.

[10]  Thomas Sikora,et al.  MPEG-7 Audio and Beyond: Audio Content Indexing and Retrieval , 2005 .

[11]  Bob L. Sturm,et al.  Musical instrument identification using multiscale Mel-frequency cepstral coefficients , 2010, 2010 18th European Signal Processing Conference.

[12]  Ewa Lukasik,et al.  Long Term Cepstral Coefficients for Violin Identification , 2010 .

[13]  Chang Dong Yoo,et al.  Music genre classification using novel features and a weighted voting method , 2008, 2008 IEEE International Conference on Multimedia and Expo.

[14]  Johan A. K. Suykens,et al.  Least Squares Support Vector Machine Classifiers , 1999, Neural Processing Letters.

[15]  Keld K. Jensen,et al.  Timbre Models of Musical Sounds , 1999 .

[16]  Lie Lu,et al.  Music type classification by spectral contrast feature , 2002, Proceedings. IEEE International Conference on Multimedia and Expo.

[17]  Pierre Geurts,et al.  Extremely randomized trees , 2006, Machine Learning.