User input audio low distortion processing method in multi-source noise environment
By calculating the sound source steering vector and frequency domain observation column vector of the microphone array, updating the noise covariance matrix, and adopting the minimum variance distortionless response criterion, the problem of target echo being misjudged as interference noise in a strong reverberation environment in an in-vehicle intelligent cockpit is solved, and low-distortion audio processing is achieved.
Patent Information
- Application Number
- CN202610147571.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-03
- Publication Date
- 2026-05-01
- Estimated Expiration
- 2046-02-03
AI Technical Summary
In the voice interaction system of in-vehicle intelligent cockpit, existing technologies are prone to misidentifying the target echo as interference noise in strong reverberation environments, resulting in audio distortion. Existing algorithms such as MVDR and GSC perform poorly in strong reverberation environments and cannot effectively suppress interference and maintain the quality of the target voice.
By acquiring the sound source steering vector and frequency domain observation column vector of the microphone array, the power spectra of the superimposed reference signal and the differential reference signal are calculated. Combined with the preset acoustic transmission upper limit energy ratio and interference presence probability weight, the noise covariance matrix is updated. Beamforming is performed using the minimum variance distortionless response criterion to accurately distinguish target echoes from external interference.
It effectively suppresses external interference, reduces audio distortion, improves speech enhancement, and ensures low-distortion audio processing capabilities in multi-source noise environments.