TY - JOUR

T1 - Speaker-invariant suprasegmental temporal features in normal and disguised speech

AU - Leemann, Adrian

AU - Kolly, Marie-José

PY - 2015/12

Y1 - 2015/12

N2 - Cases of blackmailing and kidnapping often involve the perpetrator disguising his voice. In the present contribution, we examined the degree to which one exemplary form of voice disguise – the imitation of a foreign dialect – affects suprasegmental temporal features. Results of two production experiments and one perception test revealed high between-speaker and low within-speaker variability across the disguise condition. Results from the perception test further underlined the difficulty of the imitation task, as the majority of speakers were not accepted as native speakers by listeners of the target dialect. At the same time, the high rate of rejections of imitations as authentic speech indicates a high sensitivity and dialectal awareness on the part of the listeners. Findings suggest that those imitators who were accepted as native speakers may have been accepted partly because they succeeded in adjusting speaking rate and global intensity peak variability to that of the target dialect, amongst other factors.

AB - Cases of blackmailing and kidnapping often involve the perpetrator disguising his voice. In the present contribution, we examined the degree to which one exemplary form of voice disguise – the imitation of a foreign dialect – affects suprasegmental temporal features. Results of two production experiments and one perception test revealed high between-speaker and low within-speaker variability across the disguise condition. Results from the perception test further underlined the difficulty of the imitation task, as the majority of speakers were not accepted as native speakers by listeners of the target dialect. At the same time, the high rate of rejections of imitations as authentic speech indicates a high sensitivity and dialectal awareness on the part of the listeners. Findings suggest that those imitators who were accepted as native speakers may have been accepted partly because they succeeded in adjusting speaking rate and global intensity peak variability to that of the target dialect, amongst other factors.

KW - Voice disguise

KW - Suprasegmental temporal features

KW - Prosody

KW - Forensic phonetics

KW - Speaker invariance

U2 - 10.1016/j.specom.2015.10.002

DO - 10.1016/j.specom.2015.10.002

M3 - Journal article

VL - 75

SP - 97

EP - 122

JO - Speech Communication

JF - Speech Communication

ER -