Related Experiment Video
Updated: May 30, 2025
09:44
Neuro-rehabilitation Approach for Sudden Sensorineural Hearing Loss
Published on: January 25, 2016
19.1K
Neural-WDRC: A Deep Learning Wide Dynamic Range Compression Method Combined With Controllable Noise Reduction for
Huiyong Zhang1,2, Brian C J Moore3, Feng Jiang1
1Key Laboratory of Noise and Vibration Research, Institute of Acoustics, Chinese Academy of Sciences, Beijing, China.
Trends in Hearing
|January 27, 2025
Summary
A new deep learning method, Neural-WDRC, improves hearing aid performance by integrating wide dynamic range compression (WDRC) and noise reduction. This approach enhances speech clarity and listening comfort, especially in noisy environments.
Area of Science:
- Audiology
- Signal Processing
- Artificial Intelligence
Background:
- Hearing aids utilize wide dynamic range compression (WDRC) for amplification and noise reduction to improve speech intelligibility and comfort.
- Sequential implementation of WDRC and noise reduction in current hearing aids can distort speech and noise amplitude modulation patterns.
Purpose of the Study:
- To introduce Neural-WDRC, a novel deep learning method for integrated noise reduction and WDRC in hearing aids.
- To evaluate the effectiveness of Neural-WDRC compared to conventional compression techniques.
Main Methods:
- A two-stage, low-complexity deep learning network (Neural-WDRC) was developed to estimate and process speech and noise components separately.
- Fast-acting compression was applied to estimated speech, and slow-acting compression to estimated noise, allowing for controllable residual noise.
- Frame-based processing determined output based on current and preceding frames.
Main Results:
- Objective measures showed Neural-WDRC effectively mitigated negative speech-noise interactions in non-stationary noise conditions.
- Listening tests indicated a preference for Neural-WDRC over conventional and SNR-aware compression methods for speech in non-stationary noise.
Conclusions:
- Neural-WDRC offers a promising approach for simultaneous WDRC and noise reduction in hearing aids.
- The method improves audibility and listening experience, particularly in challenging, dynamic acoustic environments.
Related Concept Videos
01:31Hearing
51.8K
When we hear a sound, our nervous system is detecting sound waves—pressure waves of mechanical energy traveling through a medium. The frequency of the wave is perceived as pitch, while the amplitude is perceived as loudness.
51.8K
01:20Downsampling
126
When considering a sampled sequence with zero values between sampling instants, one can replace it by taking every N-th value of the sequence. At these integer multiples of N, the original and sampled sequences coincide. This process, known as decimation, involves extracting every N-th sample from a sequence, thereby creating a more efficient sequence.
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
The Fourier transform of the decimated sequence reveals a combination of scaled and shifted versions of the original spectrum. This...
126