Sound Source Separation Device Using Log-Likelihood Demixing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sound source separation technologies face challenges in accurately isolating voice signals from multiple sound sources amidst noise interference, particularly in complex acoustic environments.
Innovation Solution
A sound source separation device that utilizes a matrix unit to generate an objective function based on estimated source and noise vectors, employing independent component analysis and log-likelihood functions to estimate a demixing matrix, thereby minimizing mutual information and enhancing signal separation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional sound source separation methods are used, then the separation process can be implemented, but the accuracy of voice signal separation is insufficient due to noise interference
Solution Approach 1:
The patent extracts and removes noise components from the mixed sound signal by estimating noise vectors and applying suppression techniques. The noise estimation unit specifically extracts noise characteristics from the input signal, and the noise suppression unit removes these extracted noise components, thereby improving voice signal separation accuracy in noisy environments.
Solution Approach 2:
The patent introduces a demixing matrix as an intermediary computational tool that transforms the mixed signal into separated source signals. This demixing matrix acts as a mediator between the noisy input signal and the cleaned output signals, enabling effective separation by optimizing the transformation based on statistical independence assumptions.
2Measurement precision
If simple separation methods are used, then the computational process is fast, but the separation accuracy is insufficient in complex acoustic environments
Solution Approach 1:
The patent segments the sound source separation task into distinct functional modules: noise estimation unit, noise suppression unit, and demixing matrix computation unit. Each module handles a specific aspect of the separation process, making the complex algorithm more manageable and implementable while maintaining high separation accuracy through coordinated operation of these segmented components.
3Reliability
If noise suppression is applied, then voice signal quality improves, but the processing time and computational load increase
Solution Approach 1:
The patent performs noise estimation and demixing matrix computation as preliminary actions before the actual noise suppression is applied. By pre-estimating noise characteristics and pre-computing the optimal demixing matrix, the system reduces the computational burden during real-time processing, thereby improving voice recognition performance without excessive processing delays.
Data Source
Figure 1~3
Figure 4~5
AI summary
A sound source separation device according to an embodiment of the present invention may include a plurality of microphones, a matrix unit, and an output unit. The plurality of microphones may receive a plurality of microphone input signals transmitted from a plurality of sound sources. The matrix unit may generate an objective function according to an estimated source vector and an estimated noise vector estimated based on the plurality of microphone input signals, and replace a first term and a second term included in the objective function using a log-likelihood function to estimate a demixing matrix. The output unit may provide output vectors calculated based on the microphone input signals and the demixing matrix. According to a sound source separation device of the present invention, it is possible to more accurately separate voice signals transmitted from each of the plurality of sound sources by generating an objective function according to an estimated source vector and an estimated noise vector estimated based on a plurality of microphone input signals and replacing a first term and a second term included in the objective function using a log-likelihood function to estimate a demixing matrix.