Related Experiment Videos
Acoustic parameters measured by a formant-estimating speech processor for a multiple-channel cochlear implant
The Journal of the Acoustical Society of America
|July 1, 1987
Summary
This study evaluated speech processor parameters for cochlear implants. Key acoustic features effectively identified vowels and consonants, even with noise, demonstrating their utility in hearing restoration.
Area of Science:
- Audiology
- Speech Processing
- Biomedical Engineering
Background:
- Cochlear implant (CI) systems rely on wearable speech processors to encode acoustic information.
- Understanding the limitations of these processors is crucial for optimizing speech recognition in CI users.
Purpose of the Study:
- To assess the limitations of a wearable speech processor for cochlear implant systems.
- To evaluate the effectiveness of extracted acoustic parameters for speech sound recognition.
Main Methods:
- Extracted fundamental frequency (EF0), formant frequencies (EF1, EF2), and formant amplitudes (EA1, EA2) from speech signals.
- Assessed parameter degradation with added pink noise at varying signal-to-noise ratios.
- Tested a speech processor encoding these parameters in three cochlear implant patients.
Main Results:
- Extracted parameters provided sufficient information for vowel identification and consonant feature extraction (voicing, manner, place).
- Consonants were more affected by noise than vowels; EF2 degraded more than EF1.
- Patients achieved 76% vowel and 71% consonant recognition scores, with low error rates for voicing and manner.
Conclusions:
- The analyzed acoustic parameters are useful for speech recognition in cochlear implant users.
- The speech processor effectively utilized these parameters for encoding speech information.
- Further research can refine parameter extraction and coding strategies for improved CI performance.