Microphone Array Position Estimation via Probabilistic Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing microphone array position estimation techniques face challenges in accurately detecting sound source position information and microphone array calibration, especially in noisy environments, leading to reduced accuracy in sound source localization and separation processes.
Innovation Solution
A microphone array position estimation device and method that estimates the position of the microphone array by maximizing the simultaneous probability of microphone positions, sound source signal spectrums, and recorded signal spectrums using a probabilistic generation model, incorporating short-time Fourier transformations and repeated estimation to converge on optimal positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If microphone position calibration is performed using signals with precise rise timing (such as applause) or sound source position information, then the position estimation accuracy can be improved, but the reliability deteriorates in actual noise environments where rise detection may fail and sound source position information may not be available
Solution Approach 1:
The system uses the recorded sound signals themselves to automatically estimate microphone positions without requiring external calibration signals or sound source position information. The probabilistic model leverages the inherent characteristics of the recorded signals to perform self-calibration, eliminating dependency on external references that may not be available in noisy environments.
Solution Approach 2:
The invention changes the approach from using temporal parameters (rise timing) to using spectral parameters (power spectrum characteristics). By transforming the calibration problem into the frequency domain and using power spectrum ratios, the system achieves robustness against noise while maintaining position estimation accuracy.
2Measurement precision
If transfer function measurement is performed in advance with accurate microphone positioning, then the sound source localization and separation processes can be accurately performed, but the accuracy deteriorates when microphone positions deviate during actual use
Solution Approach 1:
The system performs preliminary estimation of microphone positions using the probabilistic model before conducting sound source localization or separation. This preliminary calibration action ensures that subsequent processes use accurate and up-to-date position information, compensating for any position deviations that occurred during microphone deployment.
3Measurement precision
If repeated estimation of sound source spectrums and microphone positions is performed to maximize simultaneous probability, then the position estimation accuracy is improved, but the computational complexity increases
Solution Approach 1:
The estimation process is segmented into iterative steps where the optimization problem is decomposed into smaller sub-problems. By alternating between estimating sound source spectrums and microphone positions in repeated iterations, the complex joint optimization is broken down into manageable steps that converge to the optimal solution.
Data Source
AI summary
A microphone array position estimation device includes an estimation unit that estimates a position X of a microphone array for maximizing a simultaneous probability P(X,S,Z) of X, Y, and Z through repeated estimation of S and X when the position of the microphone array constituted by M (M is an integer of 1 or greater) microphones is set to X (=(X1T, . . . , XMT)T, T indicates a transposition), spectrums of sound source signals output by the N (N is an integer of 1 or greater) sound sources are set to S (a set related to all of n, f, and t of Snft, f is a frequency bin, and t is a frame index), and spectrums of recorded signals collected by the microphone array are set to Z (a set related to all of f and t of Zft).