Signal Analysis Device for Sound Source Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional diarization devices perform suboptimal diarization and lack effective sound source localization, relying on heuristics and assuming known sound source positions, which limits their ability to accurately determine active sound sources in complex environments.
Innovation Solution
A signal analysis device that models a signal source position occurrence probability matrix using the product of a signal source position probability matrix and a signal source existence probability matrix, enabling optimal diarization and sound source localization by estimating these matrices based on input observation signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If conventional diarization devices use heuristic methods and assume known sound source positions, then the device complexity is reduced, but the measurement precision of sound source localization deteriorates
Solution Approach 1:
The patent segments the sound source position occurrence probability matrix Q into two separate matrices: the sound source position probability matrix B and the sound source existence probability matrix A. This segmentation allows independent estimation of position probabilities and existence probabilities, improving localization precision while maintaining manageable computational complexity through structured decomposition.
Solution Approach 2:
The patent transitions from assuming known sound source positions (reducing the problem to 2D time-frequency analysis) to estimating sound source positions as an additional dimension. By introducing the sound source position probability matrix B with dimensions K×N (where K is number of position candidates and N is number of sound sources), the system adds spatial dimensionality to achieve accurate localization without excessive complexity.
2Ease of operation
If conventional diarization devices assume sound source positions are known, then the ease of operation is improved, but the reliability of diarization deteriorates
Solution Approach 1:
The system performs self-service by automatically estimating sound source positions through the sound source position probability matrix B without requiring external input or manual configuration of position information. The estimation unit independently derives position probabilities from observation signals, maintaining ease of operation while significantly improving diarization reliability through accurate position estimation.
Solution Approach 2:
The patent implements feedback by using the estimated sound source position probabilities to improve diarization decisions, which in turn refines the position estimation. The sound source existence probability matrix A provides feedback on which sound sources are active, allowing the system to adaptively focus position estimation on relevant sources, thereby improving reliability without increasing operational complexity.
3Productivity
If conventional diarization devices use simple heuristic methods, then the productivity is improved, but the measurement precision of active sound source determination deteriorates
Solution Approach 1:
The patent changes the fundamental parameters from assuming fixed sound source positions to estimating position probabilities. By modeling the sound source position occurrence probability matrix Q as the product of position probability matrix B and existence probability matrix A, the system transforms the diarization problem into a parameter estimation task that achieves high precision in determining active sound sources while maintaining computational efficiency through matrix operations.
Data Source
AI summary
A signal analysis device includes an estimation unit that models a sound source position occurrence probability matrix Q using a product of a sound source position probability matrix B and a sound source existence probability matrix A, and estimates at least one of the sound source position probability matrix B and the sound source existence probability matrix A based on the modeling, the sound source position occurrence probability matrix Q being composed of probabilities of arrival of a signal from each sound source position candidate per frame, which is a time section, with respect to a plurality of sound source position candidates. The sound source position probability matrix B being composed of probabilities of arrival of a signal from each sound source position candidate per sound source with respect to a plurality of sound sources.