Summary
International Symposium on Nonlinear Theory and its Applications
2005
Session Number:3-1-2
Session:
Number:3-1-2-4
Sparse Time-Frequency Analysis of Speech Signals
Albert Kern, Oliver Nagy, Ruedi Stoop,
pp.545-548
Publication Date:2005/10/18
Online ISSN:2188-5079
DOI:10.34385/proc.40.3-1-2-4
PDF download (314.4KB)
Summary:
To obtain a sparse time-frequency representation of a speech signal, which still reflects the evolution of its characteristics in time, consists a great problem. Despite its disadvantages, the spectrogram is traditionally used for this purpose. In this contribution we examine the potential of 1) atomic decomposition methods and 2) time-frequency distributions with respect to speech signals.