Related Experiment Video
Updated: Jun 16, 2026
08:48
Combining Multiple Data Acquisition Systems to Study Corticospinal Output and Multi-segment Biomechanics
Published on: January 9, 2016
Prediction of acoustic feature parameters using myoelectric signals
1Department of Electronic Engineering, Konkuk University, Seoul 143-701, Korea. kseung@konkuk.ac.kr
IEEE Transactions on Bio-Medical Engineering
|February 23, 2010
Summary
Researchers used mouth myoelectric signals (MES) to predict speech parameters, improving speech synthesis. This method achieved a 30% reduction in acoustic parameter error and high intelligibility scores in listening tests.
Area of Science:
- Biomedical Engineering
- Speech Processing
- Signal Analysis
Background:
- A known correlation exists between human voice production and myoelectric signals (MES) from the oral region.
- Myoelectric signals offer a potential non-acoustic source for speech parameter estimation.
Purpose of the Study:
- To develop a speech synthesis method solely based on myoelectric signals.
- To identify optimal MES features for predicting vocal-tract transfer function parameters.
- To enhance speech synthesis accuracy and intelligibility using MES-derived features.
Main Methods:
- Investigated various MES-derived features to maximize mutual information with acoustic features.
- Developed an acoustic parameter estimation rule using a minimum mean square error (MMSE) criterion.
- Evaluated the system using 60 isolated words for objective and subjective assessments.
Main Results:
- Achieved an approximate 30% reduction in average Euclidean distance between original and predicted acoustic parameters.
- Synthesized speech demonstrated improved intelligibility.
- Listening tests yielded a 65.5% word-level and 73% syllable-level identification ratio.
Conclusions:
- Myoelectric signals alone can effectively predict acoustic parameters for speech synthesis.
- The proposed MMSE-based estimation rule and optimal MES features enhance synthesis accuracy.
- This MES-driven approach shows significant potential for improving speech synthesis technology.
