Related Experiment Videos
Algorithms, hardware, and software for a digital signal processor microcomputer-based speech processor in a
IEEE Transactions on Bio-Medical Engineering
|June 1, 1989
Summary
A new digital signal processing system efficiently models speech in real-time using linear predictive coding (LPC). This compact, low-power technology aids hearing-impaired individuals and cochlear implant development.
Area of Science:
- Digital Signal Processing
- Speech Technology
- Biomedical Engineering
Background:
- Development of advanced signal processing systems for real-time speech analysis.
- Need for compact, low-power hardware for hearing assistance technologies.
Purpose of the Study:
- To develop a powerful, inexpensive, and compact digital signal processing system.
- To extract a low-bit rate linear predictive coding (LPC) speech model in real-time.
- To derive accurate spectral envelope, formant, pitch, and amplitude information.
Main Methods:
- Utilized Texas Instruments TMS320 family processors.
- Designed a compact system with three chips (TMS320E17, A/D-D/A, op-amp).
- Achieved low power consumption (<0.5 W).
Main Results:
- Real-time extraction of LPC speech model parameters.
- Accurate derivation of spectral envelope, formant, pitch, and amplitude.
- Demonstrated a highly compact and energy-efficient processing system.
Conclusions:
- The developed system offers a powerful and efficient solution for real-time speech modeling.
- The technology has significant potential for programmable cochlear implant systems.
- The system is applicable to various assistive devices for the hearing handicapped.