Continuous wavelet transform for analysis of speech prosody
Year: 2013
Type: conference-paper
Abstract: Wavelet based time frequency representations of various signals are shown to reliably represent perceptually relevant patterns at various spatial and temporal scales in a noise robust way. Here we present a wavelet based visualization and analysis tool for prosodic patterns, in particular intonation. The suitability of the method is assessed by comparing its predictions for word prominences against manual labels in a corpus of 900 sentences. In addition, the method’s potential for visualization is demonstrated by a few example sentences which are compared to more traditional visualization methods. Finally, some further applications are suggested and the limitations of the method are discussed.
Authors: Martti Vainio, Antti Suni, Daniel Aalto
Language: English
Subfield: Experimental and Cognitive Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: green
Funder: European Commission
OpenAlex ID: W2600829178
