Abstract
The pitch perception literature has been largely built on experimental data collected using nonspeech stimuli, which has then been generalized to speech. In the present study, we compare the perceptibility of identical pitch movements in speech and nonspeech that vary in duration and in pitch range. Our nonspeech results closely replicate earlier findings and we show that speech is a significantly more difficult medium for pitch discrimination. Pitch movements in speech have to be larger and longer to achieve the salience of the most common speech analog, pulse trains. The direction of pitch movement also affects one's ability to discern pitch; in particular falling excursions are the most difficult. We found that the perceptual threshold for falling pitch in speech was more than 100 times that of previous estimates with nonspeech stimuli. Our findings show that the perceptual response to nonspeech does not adequately map onto speech, and future work in speech research and its applications should use speech-like stimuli, rather than convenient substitutes like pulse trains, pure tones, or isolated vowels.
Original language | English (US) |
---|---|
Pages (from-to) | 2275-2279 |
Number of pages | 5 |
Journal | Proceedings of the Annual Conference of the International Speech Communication Association, INTERSPEECH |
Volume | 2019-September |
DOIs | |
State | Published - 2019 |
Event | 20th Annual Conference of the International Speech Communication Association: Crossroads of Speech and Language, INTERSPEECH 2019 - Graz, Austria Duration: Sep 15 2019 → Sep 19 2019 |
Keywords
- Just noticeable differences
- Pitch perception
- Speech perception
- Speech resynthesis
- Speech synthesis
ASJC Scopus subject areas
- Language and Linguistics
- Human-Computer Interaction
- Signal Processing
- Software
- Modeling and Simulation