Spatial Audio Noise Reduction via Component Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise reduction systems for recorded audio signals distort spatial audio signals, making them unsuitable for use in applications like virtual reality systems, as they often remove ambient noise entirely.
Innovation Solution
A method that divides spatial audio signals into direct and ambient components, applying different noise reduction systems to each, allowing for targeted noise reduction while preserving directional properties and ambient sounds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional noise reduction systems are applied to spatial audio signals, then noise is reduced, but the spatial properties and directional information are distorted
Solution Approach 1:
The patent segments the spatial audio signal into distinct components: direct sound (from specific directions) and diffuse sound (ambient noise from all directions). By separating these components, the system can apply noise reduction selectively to the diffuse component while preserving the direct sound component, thus reducing noise without distorting spatial properties.
Solution Approach 2:
The patent applies different processing qualities to different parts of the audio signal. The direct sound component receives minimal or no noise reduction to preserve its spatial characteristics, while the diffuse sound component receives aggressive noise reduction. This local differentiation resolves the contradiction between noise reduction and spatial preservation.
2Object-affected harmful factors
If aggressive noise reduction is applied to remove all ambient noise, then noise is reduced, but the signal becomes unsuitable for presence systems like virtual reality
Solution Approach 1:
The patent implements dynamic control over the amount of noise reduction applied to different audio components. By adjusting the processing intensity dynamically based on the specific application requirements, the system can optimize for either noise reduction or presence system suitability, or achieve a balanced compromise between the two opposing requirements.
Solution Approach 2:
The patent changes the processing parameters applied to different sound components. By varying the noise reduction parameters selectively for direct and diffuse sound components, the system maintains adaptability for different applications including presence systems, while still achieving effective noise reduction where appropriate.
3Object-affected harmful factors
If uniform noise reduction is applied to all audio signals, then noise is reduced, but directional properties and spatial characteristics are lost
Solution Approach 1:
The patent segments the audio signal based on directional characteristics, separating sounds with defined directionality (direct sound) from those without (diffuse sound). This segmentation enables selective application of noise reduction algorithms that preserve the directional shape of important sounds while reducing noise in non-directional components.
Solution Approach 2:
The patent applies different noise reduction qualities to different spatial components of the audio signal. Direct sound components maintain their original quality with minimal processing to preserve directional properties, while diffuse sound components receive enhanced noise reduction. This local quality differentiation resolves the contradiction between noise reduction and directional property preservation.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method, apparatus and computer program comprising: obtaining a spatial audio signal from a plurality of microphones;dividing the obtained spatial audio signal into at least a first component and a second component;applying a first audio signal optimizing system to the first component and applying a second audio signal optimizing system to the second component; and enabling a signal comprising the optimized components to be provided to a speaker for rendering.