Related Experiment Video
Updated: Dec 10, 2025
09:09
Foreign Accent and Forensic Speaker Identification in Voice Lineups: The Influence of Acoustic Features Based on Prosody
Published on: September 27, 2024
715
A tool for efficient and accurate segmentation of speech data: announcing POnSS
Joe Rodd1,2, Caitlin Decuyper3, Hans Rutger Bosker3
1Max Planck Institute for Psycholinguistics, Radboud University, Postbus 310 6500AH, Nijmegen, The Netherlands. joe.rodd@mpi.nl.
Behavior Research Methods
|September 2, 2020
Summary
POnSS streamlines speech data segmentation, reducing annotator time by 23% while maintaining reliability. This novel system enhances the efficiency of creating research-grade speech segmentations.
Area of Science:
- Speech processing
- Computational linguistics
- Human-computer interaction
Background:
- Manual speech segmentation is crucial for research but labor-intensive.
- Advances in automatic speech recognition (ASR) have not eliminated the need for human input.
- Existing tools like Praat require significant annotator time.
Purpose of the Study:
- To introduce POnSS, a novel browser-based system for efficient speech segmentation.
- To combine automatic speech recognition (ASR) with minimal human input for improved segmentation.
- To streamline the process of segmenting word onsets and offsets in speech data.
Main Methods:
- Developed POnSS, a specialized browser-based system for segmenting speech data.
- Implemented distinct interfaces for sub-tasks of segmentation to optimize annotator workflow.
- Evaluated POnSS performance against conventional Praat-based segmentation methods.
Main Results:
- POnSS achieved comparable reliability to traditional Praat-based segmentation.
- The POnSS system reduced annotator time investment by 23%.
- The system demonstrated greater efficiency without sacrificing segmentation reliability.
Conclusions:
- POnSS offers a significant methodological advancement for speech data segmentation.
- The system enhances efficiency in producing research-grade speech segmentations.
- POnSS reduces the labor-intensive nature of manual speech segmentation.
