Characterizing temporal organization in presentation-like Japanese speech using the normalized pairwise variability index

Summary

This study examined whether the normalized pairwise variability index calculated from inter-vowel-onset intervals (nPVI-Vo) can characterize the local temporal organization of Japanese speech across different speaking contexts. Recordings from 16 speakers in the Corpus of Spontaneous Japanese were compared across presentation-like speech, dialogue, and rereading.

Presentation-like speech had a higher speaking rate than rereading, whereas dialogue exhibited greater local temporal variability (higher nPVI-Vo) than either presentation-like speech or rereading. Speaking rate and nPVI-Vo were not clearly correlated. In contrast, the conventional nPVI calculated from vowel durations did not clearly distinguish the three contexts in pairwise comparisons.

These findings suggest that vowel-onset-based nPVI provides a candidate acoustic measure of local speech timing that complements overall speaking rate. The study did not directly assess presentation quality, intelligibility, or comprehensibility.

From an HCI perspective, the results offer a measurement foundation for speech analytics and technology-assisted presentation training. Future interfaces could examine speech rate and local timing variability as distinct observable features, while separately evaluating whether these measures are informative for feedback or presentation outcomes.

Research line

• Measuring: operationalizing local temporal organization through vowel-onset intervals and nPVI
• Speech Analysis: comparing presentation-like speech, dialogue, and rereading using corpus-based acoustic measures
• HCI: identifying objective speech features that could be incorporated into interactive presentation-support tools
• Presentation Skills: extending measurement beyond speaking rate without presuming that rhythmic variation indexes presentation quality

Links

• Publisher page
• DOI

Recommended citation

Sakurai, Y., Jan, O., Matsuda, S., & Tachibana, R. O. (2026). Characterizing temporal organization in presentation-like Japanese speech using the normalized pairwise variability index. Acoustical Science and Technology. Advance online publication. https://doi.org/10.1250/ast.e26.99

Recommended citation: Sakurai, Y., Jan, O., Matsuda, S., & Tachibana, R. O. (2026). Characterizing temporal organization in presentation-like Japanese speech using the normalized pairwise variability index. *Acoustical Science and Technology*. Advance online publication. https://doi.org/10.1250/ast.e26.99
Download Paper