Efficient Real-Time Acoustic Feedback Cancellation using Adaptive Noise Injection Algorithm.

Area of Science:

• Signal Processing
• Acoustics
• Audio Engineering

Background:

• Acoustic feedback poses a significant challenge in sound reinforcement systems and hearing aids.
• Feedback occurs when the acoustic path creates a loop between microphones and loudspeakers.
• Existing methods often require prior knowledge of the acoustic path or struggle in dynamic environments.

Purpose of the Study:

• To propose an efficient short-duration noise injection algorithm for acoustic feedback path modeling.
• To enable optimal estimation and cancellation of acoustic feedback without prior information.
• To prevent howling and ensure stable system operation in various acoustic conditions.

Main Methods:

• An efficient short-duration noise injection algorithm is developed.
• The algorithm estimates the impulse response of the acoustic feedback path.
• Implementation is demonstrated on a smartphone, emulating close-proximity feedback conditions.

Main Results:

• The algorithm successfully estimates the acoustic feedback path without prior information.
• It effectively cancels acoustic feedback and prevents howling episodes.
• Experimental results demonstrate robust performance in changing acoustic environments.

Conclusions:

• The proposed algorithm offers an efficient and robust solution for acoustic feedback cancellation.
• Its platform-independent nature allows implementation across diverse systems and setups.
• The method shows significant potential for improving sound quality in audio devices and systems.