Music tonality features for speech/music discrimination
Authors:
- 1 May 2014
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- No. 15206149,p. 2489-2493
Abstract
We introduce a novel set of features for speech/music discrimination derived from chroma vectors, a feature that represents musical tonality. These features are shown to outperform other commonly used features in multiple conditions and corpora. Even when trained on mismatched data, the new features perform well on their own and also combine with existing features for further improvement. We report 97.1% precision on speech and 93.0% precision on music for the Broadcast News corpus using a simple classifier trained on a mismatched corpus.Keywords
This publication has 9 references indexed in Scilit:
- A Decision-Tree-Based Algorithm for Speech/Music Classification and SegmentationEURASIP Journal on Audio, Speech, and Music Processing, 2009
- A speech/music discriminator based on RMS and zero-crossingsIEEE Transactions on Multimedia, 2005
- Speech/music segmentation using entropy and dynamism features in a HMM classification frameworkSpeech Communication, 2003
- Real-time discrimination of broadcast speech/music2002
- Construction and evaluation of a robust multifeature speech/music discriminator2002
- Speech/music discrimination based on posterior probability features1999
- A comparison of features for speech, music discrimination1999
- A framework for audio analysis based on classification and temporal segmentation1999
- A multivariate speech activity detector based on the syllable rate1999