Acoustic signal enhancement apparatus, method and program
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing reverberation suppression and sound source separation methods in acoustic signal processing are performed independently, leading to suboptimal results when both processes are simultaneously executed.
Innovation Solution
An acoustic signal enhancement device that estimates spatiotemporal covariance matrices for each sound source and noise, applies reverberation suppression filters, and performs sound source separation iteratively to optimize the overall process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If reverberation suppression and sound source separation are performed independently in sequential steps, then each process can be implemented using existing methods, but the overall process cannot achieve optimal performance when both are simultaneously executed
Solution Approach 1:
The patent merges the reverberation suppression process and sound source separation process into a single integrated system. The spatiotemporal covariance matrix estimation unit simultaneously estimates parameters for both reverberation suppression and sound source separation, allowing both functions to be performed in conjunction rather than sequentially, thereby achieving optimal overall performance.
Solution Approach 2:
The integrated system performs multiple functions through a unified process. The spatiotemporal covariance matrix estimation serves both reverberation suppression and sound source separation purposes simultaneously, making the system multi-functional and eliminating the need for separate independent processing chains.
2Ease of manufacture
If separate processes are used for reverberation suppression and sound source separation, then implementation is simpler, but calculation costs increase and convergence performance deteriorates
Solution Approach 1:
By combining the two separate processes into one integrated system, the patent reduces redundant calculations. The spatiotemporal covariance matrix is estimated once and used for both reverberation suppression and sound source separation, eliminating the need for separate independent estimations and improving calculation efficiency.
Solution Approach 2:
The patent changes the parameter estimation approach from separate independent estimations to a unified spatiotemporal covariance matrix estimation. This parameter change enables the system to share computational resources and intermediate results between reverberation suppression and sound source separation, thereby improving productivity.
Data Source
AI summary
An acoustic signal enhancement device includes: a spatiotemporal covariance matrix estimation unit 2 configured to estimate spatiotemporal covariance matrices Rf(j) and Pf(j); a reverberation suppression unit 3 configured to obtain a reverberation suppression filter Gf(j) of the sound source j using the estimated spatiotemporal covariance matrices Rf(j) and Pf(j) for each sound source j and to generate a reverberation suppression signal vector using the obtained reverberation suppression filter Gf(j) and the observation signal vector Xt,f; a sound source separation unit 4 configured to obtain an enhanced sound yt,f(j) of the sound source j and power of the sound source j using the generated reverberation suppression signal vector for each sound source j (where 1≤j≤J) corresponding to the target sound; and a control unit 5 configured to perform control such that processes of these units are repeatedly performed.