In-Car Audio Feedback Suppression via Localized Frequency Bin Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In-car communication systems face challenges in mitigating audio feedback, which can cause ringing due to the feedback path, despite efforts to manage closed-loop gain, as existing solutions are not always effective in preventing this issue.
Innovation Solution
A method and system that calculate a smoothed frequency spectrum of an audio signal, detect candidate feedback tones, modify feedback coefficients based on signal characteristics, and determine actionable feedback tones to generate suppression coefficients for mitigating audio feedback, using a combination of an acoustic echo canceller and feedback suppressor module.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If audio feedback suppression is increased to reduce ringing, then feedback mitigation is improved, but audio quality and naturalness of conversation deteriorate
Solution Approach 1:
The patent applies different suppression coefficients to different frequency bins based on local characteristics. The feedback suppressor calculates suppression coefficients selectively for frequency bins that exhibit feedback characteristics (high energy, sustained over time, matching echo path), rather than applying uniform suppression across all frequencies. This localized approach maintains audio quality in non-feedback frequencies while effectively suppressing feedback tones.
Solution Approach 2:
The suppression coefficients are dynamically adjusted based on real-time analysis of audio signal characteristics. The system continuously monitors frequency spectrum changes, updates suppression coefficients frame-by-frame, and adapts to varying feedback conditions. This dynamic adaptation allows the system to maintain high feedback suppression while preserving natural audio quality as conditions change.
2Reliability
If feedback suppression is applied continuously, then feedback mitigation is improved, but distortion of speech and music signals increases
Solution Approach 1:
The patent applies feedback suppression partially rather than continuously across all frequency components. Suppression is applied only to frequency bins that are identified as feedback tones based on specific criteria (energy threshold, temporal persistence, correlation with echo path). By applying suppression only where necessary and at appropriate levels, the system achieves effective feedback mitigation while minimizing distortion of speech and music signals.
Solution Approach 2:
The system uses feedback from the audio signal itself to control the suppression process. By continuously analyzing the microphone input and comparing it with the echo path characteristics, the system identifies feedback tones and adjusts suppression coefficients accordingly. This feedback mechanism ensures suppression is applied only when and where feedback is actually present, preserving signal fidelity.
3Reliability
If aggressive feedback suppression is used to eliminate ringing, then feedback mitigation is improved, but loss of useful audio information increases
Solution Approach 1:
The patent carefully controls the suppression parameter (suppression coefficient) to balance feedback elimination and information preservation. The coefficient is calculated based on feedback tone characteristics and applied with varying strength across different frequency bins. By adjusting this parameter dynamically and selectively, the system eliminates feedback while preserving useful audio information such as speech and music content.
Solution Approach 2:
Different suppression strengths are applied to different frequency regions based on local feedback characteristics. Frequency bins identified as feedback tones receive suppression, while other frequency bins maintain their original signal. This localized differentiation preserves audio information in non-feedback regions while effectively suppressing feedback in identified regions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and method for mitigating audio feedback may calculate a smoothed frequency spectrum of an audio signal. Previously detected candidate feedback tones may be obtained. Candidate feedback tones may be determined responsive to a frequency spectrum of the audio signal, the smoothed frequency spectrum and the previously detected candidate feedback tones. One or more signal characteristics associated with the audio signal and feedback coefficients associated with the candidate feedback tones may be obtained. The feedback coefficients may be modified responsive to the one or more signal characteristics. Actionable feedback tones may be determined responsive to the associated feedback coefficients exceeding a respective feedback threshold. Feedback suppression coefficients associated with each of the determined actionable feedback tones may be generated and may be utilized to suppress the actionable feedback tones.