Audio Signal Processing for Howling Suppression in Multi-Talker Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio signal processing methods for reducing howling in multi-talkers scenarios often require complex signal processing and unnecessarily lower the output level of other talkers, making it difficult to clearly hear multiple speakers simultaneously.
Innovation Solution
An audio signal processing method that adjusts the send level of audio signals from microphones to speakers, setting the level to a minimum for the microphone detecting a talker voice and maintaining the reference level for other microphones, thereby reducing howling without complex processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If the output level of the speaker near the selected microphone is reduced to eliminate howling, then howling is suppressed, but the level of other talker voices outputted from the same speaker is also lowered
Solution Approach 1:
The audio signal from each microphone is segmented and routed to different speakers based on spatial relationships. The signal processing is divided into multiple paths: one path reduces the level for the specific microphone-speaker pair causing howling, while another path maintains full level for other microphone-speaker pairs, allowing selective suppression without affecting overall voice output
Solution Approach 2:
Different signal processing treatments are applied to different speaker outputs based on their specific spatial relationships with microphones. The speaker near the selected microphone receives signal level reduction to prevent howling, while other speakers maintain normal signal levels to preserve voice output quality
2Object-affected harmful factors
If complex filter settings and signal processing are applied to reduce howling, then howling suppression is achieved, but the system complexity increases
Solution Approach 1:
The patent extracts and eliminates the need for complex filter settings and adaptive signal processing algorithms by using a simpler approach: detecting talker presence with microphones and selectively adjusting speaker output levels based on predetermined spatial relationships, thereby achieving howling suppression without complex processing
Solution Approach 2:
The invention replaces complex, computationally intensive signal processing systems with a simpler, more economical solution that uses basic level adjustment based on microphone selection and speaker positioning, reducing system complexity while maintaining effectiveness
Data Source
AI summary
An audio signal processing method includes providing a plurality of apparatuses and performing. The step of providing a plurality of apparatuses includes providing a plurality of sound collection apparatuses and a plurality of sound emission apparatuses corresponding to the plurality of sound collection apparatuses. The step of performing includes performing send level adjustment processing that sets a level of an audio signal of a sound collected by the plurality of sound collection apparatuses to a preset send level and sends the audio signal to the plurality of sound emission apparatuses. The send level adjustment processing changes a first send level to a minimum, the first send level being sent from a first sound collection apparatus that detects a talker voice, among the plurality of sound collection apparatuses, to a first sound emission apparatus closest to the first sound collection apparatus.