Spatial Audio Rendering for Speech Noise Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech enhancement algorithms fail to effectively separate speech and noise in non-stationary environments, leading to imperfect noise suppression and reduced user experience due to the assumption of stationary noise, which is not always valid in real-world scenarios.
Innovation Solution
An apparatus and method that separate sound signals into speech and noise components and use spatial rendering to distribute these components differently in three-dimensional space, allowing the human auditory system to exploit spatial localization cues for improved separation, rather than relying on conventional noise suppression techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If conventional noise suppression techniques are used, then noise level is reduced, but speech intelligibility and quality deteriorate due to noise suppression artifacts
Solution Approach 1:
The patent transforms the noise suppression problem from a one-dimensional amplitude reduction task into a three-dimensional spatial distribution task. By distributing speech and noise components to different spatial locations, the system exploits the additional spatial dimension to achieve separation without degrading speech quality, thus resolving the contradiction between noise reduction and speech intelligibility.
Solution Approach 2:
The patent segments the mixed audio signal into distinct speech and noise components through spatial separation. By assigning different spatial distributions to speech (directional) and noise (diffuse), the system enables the auditory system to naturally separate the components, improving speech intelligibility while maintaining noise suppression effectiveness.
2Reliability
If spatial rendering is used to distribute speech and noise components, then speech intelligibility is improved, but device complexity increases
Solution Approach 1:
The patent leverages the human auditory system's inherent ability to perform spatial localization and source separation as a free resource. By distributing speech and noise to different spatial locations, the system allows the listener's brain to automatically separate the components without requiring additional complex processing, thus improving speech intelligibility while avoiding excessive device complexity.
3Device complexity
If stationary noise assumption is used, then processing is simplified, but performance deteriorates in non-stationary environments
Solution Approach 1:
The patent bypasses the need for complex temporal analysis of noise stationarity by introducing spatial distribution as an additional dimension. Instead of trying to accurately model and track non-stationary noise over time, the system distributes noise components to diffuse spatial locations, allowing effective noise suppression even in non-stationary environments without significantly increasing processing complexity.
Data Source
Figure 1~3
Figure 4~5
AI summary
The present invention relates to an apparatus (100) for improving a perception of a sound signal (S), the apparatus comprising: a separation unit (10) configured to separate the sound signal (S) into at least one speech component (SC) and at least one noise component (NC); and a spatial rendering unit (20) configured to generate an auditory impression of the at least one speech component (SC) at a first virtual position (VP1) with respect to a user, when output via a transducer unit (30),and of the at least one noise component (NC) at a second virtual position (VP2) with respect to the user, when output via the transducer unit (30).