Speech recognition with dynamic range reduction: (1) deaf and normal subjects in laboratory conditions

Area of Science:

• Audiology
• Speech Processing
• Hearing Science

Background:

• Dynamic range reduction in speech processing aims to enhance intelligibility and prevent auditory overload.
• Automatic Gain Control (AGC) devices face theoretical and practical limitations for speech dynamic range reduction.

Purpose of the Study:

• To evaluate a novel signal processing technique using high-frequency carrier clipping (HFCC) for speech dynamic range reduction.
• To assess the impact of HFCC on speech intelligibility in individuals with normal hearing and hearing impairments.

Main Methods:

• Implementation of signal processing employing high-frequency carrier clipping (HFCC).
• Testing HFCC on normal-hearing subjects with simulated deafness and patients with partial hearing loss.

Main Results:

• HFCC demonstrated increased speech intelligibility in normal subjects with simulated deafness.
• Most partially hearing patients experienced improved speech intelligibility with HFCC.
• Some patients did not show improved intelligibility, potentially due to reliance on lost speech features.

Conclusions:

• High-frequency carrier clipping (HFCC) is an effective method for improving speech intelligibility and protecting hearing.
• Further investigation and potential training may be necessary for patients who do not initially benefit from HFCC.
• HFCC overcomes limitations associated with traditional AGC devices for speech processing.