Related Experiment Videos
The Indiana Speech Training Aid (ISTRA). I: Comparisons between human and computer-based evaluation of speech quality
C S Watson1, D J Reed, D Kewley-Port
1Department of Speech and Hearing Sciences, Indiana University.
Journal of Speech and Hearing Research
|June 1, 1989
Summary
Computer analysis of speech quality is a reliable alternative to human judgment, especially for speech therapy drills. While humans are slightly better judges, computers offer superior consistency in evaluating speech performance.
Area of Science:
- Speech processing
- Human-computer interaction
- Speech pathology
Background:
- Evaluating speech quality is crucial for speech therapy and research.
- Traditional methods rely on subjective human judgments, which can be inconsistent.
- Automated speech analysis offers potential for objective and reliable assessment.
Purpose of the Study:
- To compare the accuracy and reliability of computer-based versus human judgments of speech quality.
- To assess the suitability of computer-based evaluation for speech therapy applications.
Main Methods:
- Speech samples from normal talkers and a hearing-impaired child were recorded.
- Recordings were evaluated by a naive jury, an expert jury, and the Indiana Speech Training Aid (ISTRA) system.
- Correlational analyses were used to compare judgment methods.
Main Results:
- Human judges were slightly more accurate than the ISTRA system in assessing speech quality.
- The ISTRA system demonstrated significantly higher reliability compared to human judges.
- Computer-based evaluation showed strong potential for consistent speech quality assessment.
Conclusions:
- Computer-based speech evaluation can serve as a viable substitute for human judgment in specific contexts, such as speech drills.
- The reliability of automated systems makes them valuable tools for objective speech analysis.
- Further research can explore the integration of computer-based tools in clinical speech therapy settings.