Multichannel Audio Noise Estimation Using SRP Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current noise estimation technologies are ineffective for processing multichannel audio signals acquired by multiple microphones, limiting their accuracy and robustness in noise suppression and interference suppression in voice recognition systems.
Innovation Solution
An audio signal noise estimation method that calculates delay differences between microphones and determines SRP values at preset sampling points, using these values to update a noise SRP multidimensional vector, allowing for accurate noise recognition in multichannel audio signals through the comparison of present frame SRP and noise SRP vectors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If traditional noise estimation technology is used, then processing of single-channel audio signals is accurate, but processing of multichannel audio signals is ineffective
Solution Approach 1:
The patent segments the multichannel audio signal processing into multiple independent SRP calculations for different sampling points. Each sampling point undergoes separate delay difference calculation and SRP value determination, allowing the system to handle multichannel signals by breaking them down into manageable discrete components that can be processed individually and then aggregated.
Solution Approach 2:
The patent transitions from single-channel to multichannel processing by adding spatial dimensionality through multiple sampling points. The SRP multidimensional vector incorporates spatial information from multiple microphones and sampling points, effectively moving the problem from one-dimensional (single channel) to multi-dimensional (spatial distribution) processing.
2Measurement precision
If noise estimation is performed without SRP multidimensional vectors, then processing is simpler, but noise recognition accuracy in multichannel signals is poor
Solution Approach 1:
The SRP multidimensional vector serves as an intermediary that bridges the gap between raw multichannel audio signals and noise recognition decisions. Instead of directly comparing complex multichannel signals, the patent uses SRP vectors as an intermediate representation that captures spatial energy distribution, making noise detection more accurate while managing complexity through structured intermediate data.
Solution Approach 2:
The patent replaces traditional mechanical noise estimation approaches with a signal processing-based SRP calculation system. Instead of using simple energy comparison or spectral subtraction methods, the system uses spatial correlation and delay-based SRP calculations to achieve more accurate noise recognition in multichannel environments.
3Measurement precision
If delay difference calculation is performed for every two MICs at each sampling point, then spatial resolution is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary delay difference calculations based on known microphone positions and sampling point positions. By pre-calculating or efficiently computing delay differences using the geometric relationships between microphones and sampling points, the system reduces the computational burden during actual signal processing while maintaining high spatial resolution.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
Disclosed are an audio signal noise estimation method and device, and a storage medium. The method includes that: for multiple preset sampling points, a noise Steered Response Power (SRP) value of a Microphone (MIC) array at each preset sampling point within a preset noise sampling period is determined to obtain a noise SRP multidimensional vector including the multiple noise SRP values corresponding to the multiple preset sampling points; a present frame SRP value for a present frame of an audio signal acquired by the MIC array at each preset sampling point is determined to obtain a present frame SRP multidimensional vector including the multiple present frame SRP values corresponding to the multiple preset sampling points; and whether the audio signal acquired by the MIC array in the present frame is a noise signal is determined according to the present frame SRP multidimensional vector and the noise SRP multidimensional vector.