Abstract
Sensorimotor synchronisation with metronome and music have been extensively studied, while synchronisation with speech is still relatively poorly understood. The present study looks into the question how to define the best anchor of synchronised movement (finger tapping) in speech, and compares manually identified vowel onsets with four acoustic landmarks that were derived by different signal processing algorithms. Participants listened to repetitions of natural English sentences and were instructed to tap in synchrony with what they perceived to be the sentence beat. The time course of the sentences was tagged for a number of rhythmically relevant events, including vowel onsets, fastest energy increase (maxD), a combination of high local pitch and periodic energy (PPP), and the largest amplitude of intersyllabic and interstress timescales (IMF1 and IMF2). Vowel onsets and maxD showed consistent tapping patterns, while other landmarks performed worse than vowel onsets. These findings suggest that local energy changes shape sensorimotor synchronisation with speech and that energy contours might serve as anchors of rhythmic attention in spoken language.
Original language | English |
---|---|
Title of host publication | Proceedings of the International Conference on Speech Prosody |
Pages | 1-5 |
Number of pages | 5 |
Volume | 2020 |
DOIs | |
Publication status | Published - 25 May 2020 |
Externally published | Yes |
Event | 10th International Conference on Speech Prosody 2020 - Virtual due to COVID-19 (Tokyo), Virtual, Japan Duration: 25 May 2020 → 28 May 2020 Conference number: 10 https://sp2020.jpn.org/ |
Publication series
Name | Proceedings of the International Conference on Speech Prosody |
---|---|
Publisher | International Speech Communication Association |
Volume | 2020 |
ISSN (Print) | 2333-2042 |
Conference
Conference | 10th International Conference on Speech Prosody 2020 |
---|---|
Abbreviated title | SP2020 |
Country/Territory | Japan |
City | Virtual |
Period | 25/05/20 → 28/05/20 |
Internet address |