LSDD Soft Masker for Acoustic Direction Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional direction-of-arrival (DOA) estimation systems for near-eye displays (NEDs) face challenges in accurately determining the direction of acoustic sources due to reflections and reverberations in the acoustic environment, and are often computationally demanding, requiring complex processing techniques and imposing structural limitations on microphone arrays.
Innovation Solution
The implementation of a local space-domain distance (LSDD) soft masker that weights specific time-frequency bins in the acoustic spectrum to determine the direction of acoustic sources, allowing for computationally-efficient DOA estimation without discarding information, by computing weighted LSDD spectrum values and applying a soft mask to focus on dominant direct-path signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional DOA estimation systems are used, then direction of acoustic source can be determined, but the systems are computationally demanding and require complex processing techniques
Solution Approach 1:
The patent segments the acoustic spectrum into time-frequency bins and processes each bin individually using local space-domain distance calculations. This segmentation allows computationally-efficient processing by avoiding complex global optimization while maintaining directional accuracy through localized analysis of each time-frequency component.
Solution Approach 2:
The patent changes the parameter space from conventional spatial domain to local space-domain distance metric. By transforming the DOA estimation problem into a parameter optimization problem using LSDD spectrum values, the system achieves accurate direction determination through simpler computational operations rather than complex signal processing techniques.
2Reliability
If binary masking techniques are used to filter reflections and reverberations, then direct path signals can be selected, but information is lost and computational intensity increases
Solution Approach 1:
The patent applies local quality by computing space-domain distance metrics for each time-frequency bin individually, allowing differential weighting of bins based on their local characteristics. This enables the system to identify and emphasize bins with dominant direct-path signals while maintaining computational efficiency through localized processing rather than global binary masking.
Solution Approach 2:
The patent introduces the local space-domain distance (LSDD) spectrum as an intermediary metric that mediates between the raw acoustic spectrum and the final DOA estimation. The LSDD spectrum serves as a computational bridge that filters reflections and reverberations without requiring binary masking, preserving information while enabling reliable direct path signal selection.
3Measurement precision
If spherical microphone array structures are used, then DOA estimation accuracy is improved, but device complexity and manufacturing difficulty increase
Solution Approach 1:
The patent substitutes the mechanical complexity of spherical microphone array structures with a computational approach using local space-domain distance metrics. Instead of requiring physical spherical arrangement of microphones, the system achieves DOA estimation accuracy through algorithmic processing of planar microphone array data, replacing mechanical complexity with computational intelligence.
Data Source
AI summary
One embodiment of the present application sets forth a computer-implemented method that includes receiving, from a first microphone, a first input acoustic signal, generating a first audio spectrum from at least the first input acoustic signal, where the first audio spectrum includes a set of time-frequency bins, for each time-frequency bin included in the set of time-frequency bins, computing a weighted local space-domain distance (LSDD) spectrum value based on a portion of the first audio spectrum that is included in the time-frequency bin, generating a combined spectrum value based on a set of the weighted LSDD spectrum values computed for the set of time-frequency bins, and determining a first estimated direction of the first input acoustic signal based on the combined spectrum value.