Joint Acoustic Source Position and Pitch Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing algorithms for localizing and identifying acoustic sources in multi-source environments face challenges in real-world scenarios due to uncorrelated noise and room reverberation, and fail to effectively combine pitch and delay information for precise source localization and separation.
Innovation Solution
A method that evaluates the cross-correlation function using a sampling function dependent on both pitch and spatial parameters, allowing for the linking of pitch and spatial information to track acoustic sources, and uses multiple microphone pairs to enhance resolution and performance, enabling joint position and pitch tracking in a multidimensional space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional TDOA-based localization algorithms are used, then source location can be estimated using geometric considerations, but the algorithms fail in real-world situations due to uncorrelated noise and room reverberation
Solution Approach 1:
The patent combines pitch tracking and TDOA-based localization into a unified joint estimation framework. By merging these two previously separate algorithms, the system can simultaneously estimate pitch and location while being robust to noise and reverberation that would otherwise degrade individual algorithm performance
Solution Approach 2:
The invention changes the parameter estimation approach by jointly estimating pitch and location parameters rather than separately. This joint parameter estimation allows the system to exploit correlations between pitch and spatial information, improving reliability in adverse acoustic conditions
2Productivity
If pitch tracking and TDOA localization are estimated independently, then each quantity can be processed separately, but tracking and separation of multiple sources remains ambiguous and unsolved
Solution Approach 1:
The patent merges independent pitch tracking and TDOA localization processing into a joint estimation framework. This combination allows the system to maintain processing efficiency while eliminating the ambiguity that arises from independent estimation, as the joint framework exploits correlations between pitch and spatial parameters for unambiguous source identification
3Adaptability or versatility
If a complex setup of RTNN-layers is used for joint F0-localisation, then source separation can be achieved, but the system requires complex architecture and limited ITD-values and pitches analysis
Solution Approach 1:
The patent extracts the essential functionality from complex RTNN-based joint F0-localisation systems by focusing on the core joint estimation of pitch and location. This extraction eliminates the need for complex recurrent neural network architectures while maintaining source separation capability through a more straightforward joint estimation framework
4Measurement precision
If more microphone pairs are used, then resolution and performance are enhanced, but device complexity increases
Solution Approach 1:
The patent applies preliminary processing steps that optimize the use of available microphone pairs before joint estimation. By preprocessing the signals and organizing microphone pair data efficiently, the system maximizes measurement precision from the given hardware configuration without requiring additional microphones, thus avoiding increased device complexity
Data Source
AI summary
The invention relates to a method for localizing and tracking acoustic sources (101) in a multi-source environment, comprising the steps of recording audio-signals (103) of at least one acoustic source (101) with at least two recording means (104, 105), creating a two- or multi-channel recording signal, partitioning said recording signal into frames of predefined length (N), calculating for each frame a cross-correlation function as a function of discrete time-lag values (τ) for channel pairs (106, 107) of the recording signal, evaluating the cross-correlation function by calculating a sampling function depending on a pitch parameter (f0) and at least one spatial parameter (φ0), the sampling function assigning a value to every point of a multidimensional space being spanned by the pitch-parameter and the spatial parameters, and identifying peaks in said multidimensional space with respective acoustic sources in the multi-source environment.


