Perception modeling of native and foreign-accented Japanese speech based on prosodic features of pitch accent
This study investigates the influence acoustic measures of pitch accent have on L1 and Australian English (AusE) L2 Japanese speech perception, expanding Tsurutani (2010) and Ishihara, Tsurutani, and Tsukada (2011), and motivated by Munro and Derwing (2001), which studies the role of speaking rate o...
Gespeichert in:
Hauptverfasser: | , , |
---|---|
Format: | Tagungsbericht |
Sprache: | eng |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | This study investigates the influence acoustic measures of pitch accent have on L1 and Australian English (AusE) L2 Japanese speech perception, expanding Tsurutani (2010) and Ishihara, Tsurutani, and Tsukada (2011), and motivated by Munro and Derwing (2001), which studies the role of speaking rate on judgments of L2 speech. We establish native and advanced AusE listeners of Japanese differ in their judgments of foreign accent in terms of accentedness and comprehensibility (Munro and Derwing 1995, 1999) through a listening task. Selected acoustic measures of pitch accent from the speech stimuli, which displayed significant variance across listener groups -- delta-pitch, max and mean max delta-intensity, and duration per mora -- are correlated with L1 and L2 listener data. Testing for a relationship between each of the acoustic measures and listener judgments, the regression analyses show a considerable relationship between comprehensibility judgments and duration and intensity features, ranging from adjusted R^2 = 14.3% to 24.6% across listeners, and indicating the degree of variance between judgments can be attributed to these acoustic measures. We can interpret that comprehensibility is linked to intensity and duration, which supports the authors' prior findings that timing is considered more important than pitch in the detection of foreign-accented speech. |
---|---|
ISSN: | 1939-800X |
DOI: | 10.1121/1.4799501 |