Signal Separator for Accurate Audio Source Localization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing blind source separation methods fail to accurately restore spatial information of point sources and often map residual interference signals to the same location as the desired signal source, leading to incorrect localization and interference.
Innovation Solution
A signal separator is configured to produce output signals with minimal distortion, ensuring that the audio content of the useful-signal source is accurately represented while separating and localizing it from interference signals, by limiting distortion and using adaptive filtering to correctly position both signal sources in space.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If blind source separation methods are used to separate signals from point sources, then signal separation is achieved, but spatial information about the point sources is lost
Solution Approach 1:
The patent segments the signal processing into two distinct stages: first, blind source separation is applied to separate the mixed signals from multiple point sources; second, spatial information is restored by analyzing the separation filters and reconstructing the spatial characteristics. This segmentation allows each stage to optimize for its specific function without compromising the other.
Solution Approach 2:
The patent performs preliminary blind source separation to isolate the point source signals before restoring spatial information. By first separating the signals and then applying spatial reconstruction based on the separation filters, the system prepares the signal in a state that enables accurate spatial information recovery without interference from mixed sources.
2Loss of information
If spatial information is restored by downstream processing, then spatial characteristics are recovered, but residual interference signals are mapped to the same location as the desired signal source
Solution Approach 1:
The patent applies local quality by differentiating the processing of desired signal components versus interference components. The spatial reconstruction selectively enhances the spatial characteristics of the desired point source while applying different spatial filtering to residual interference, ensuring they are mapped to their correct spatial locations rather than being conflated with the desired signal.
Solution Approach 2:
The patent changes the parameters of the spatial reconstruction process based on the characteristics of different signal components. By adjusting spatial filtering parameters selectively for desired signals versus residual interference, the system restores spatial information for the target source while maintaining correct spatial positioning of interference signals, preventing them from being mapped to the same location.
3Device complexity
If one-channel representation is used at the output after demixing, then signal separation is simplified, but spatial information including level differences and run-time differences is lost
Solution Approach 1:
The patent transitions from one-channel representation to multi-channel representation by incorporating spatial dimension information. The separation filters obtained from blind source separation are analyzed to reconstruct spatial characteristics, effectively adding the spatial dimension back to the output signals without significantly increasing processing complexity.
Solution Approach 2:
The patent replaces complex multi-channel spatial processing with a simplified approach that uses the separation filters from blind source separation to infer and restore spatial information. Instead of maintaining complex multi-channel systems throughout, the method substitutes filter analysis and spatial reconstruction to achieve the same spatial information preservation with reduced complexity.
Data Source
AI summary
A signal separator, a method and computer product for determining a first output signal describing an audio content of a useful-signal source in a first microphone signal, and for determining a second output signal describing an audio content of the useful-signal source in a second microphone signal.