Speech Prosody 2012

Shanghai, China
May 22-25, 2012

Testing the Perception of Speech Rhythm on Natural and Artificial Stimuli

Paolo Mairano (1), Antonio Romano (2)

(1) GIPSA-Lab, University of Grenoble 3, France
(2) LFSAG, University of Turin, Italy

It has long been assumed that the stress-timed vs. syllabletimed dichotomy is based on perceptual impressions of speech rhythm. However, experimental tests are rare in the literature and not all of them have successfully found perceptual evidence for speech rhythm categorization. Experimental protocols have been very different, sometimes testing naïve vs. non-naïve listeners, sometimes using natural speech stimuli, sometimes preferring synthetic stimuli, sometimes using filters to hide lexical information from speech. In this paper we describe the results of a perceptive test that has been carried out on 43 Italian listeners, who were asked to categorize both natural and artificial stimuli. Results are highly controversial: listeners do not seem to be able to categorize artificial stimuli reproducing prosodic cues of different languages. Yet, there is a mild tendency on the part of speakers to categorize natural speech stimuli of unknown languages in a way that seems to reflect rhythm classes.

Index Terms: speech prosody, speech rhythm, perceptive test, natural stimuli, artificial stimuli.

Full Paper

Bibliographic reference.  Mairano, Paolo / Romano, Antonio (2012): "Testing the perception of speech rhythm on natural and artificial stimuli", In SP-2012, 482-485.