Voice Signal Noise Suppression Using Spatial Dictionary Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise suppression technologies, such as spectrum subtraction and gain function-based methods, face challenges like musical noise, poor performance in unsteady noise environments, and failure to account for spatial transfer characteristics and radiation patterns of noise sources.
Innovation Solution
A voice signal processing apparatus that uses a control calculation unit to acquire noise dictionary data from a noise database based on installation environment information, including noise type, orientation, and distance, and a noise suppression unit that performs noise suppression using this data, incorporating transfer functions and gain function interpolation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If spectrum subtraction method is used for noise suppression, then noise reduction effect is achieved, but musical noise artifacts are generated
Solution Approach 1:
The patent changes the parameter of noise estimation by using spatial information (azimuth angle, distance) and transfer functions to adaptively adjust noise suppression parameters. Instead of fixed spectrum subtraction, the system dynamically adjusts suppression strength based on estimated noise source characteristics and spatial relationships, reducing musical noise while maintaining noise reduction effectiveness.
Solution Approach 2:
The patent introduces spatial dimension information (azimuth angle, distance from noise source) to the traditional one-dimensional spectrum subtraction approach. By incorporating transfer functions that model spatial acoustic propagation, the system transforms the noise suppression problem from simple spectral manipulation to a multi-dimensional spatial-spectral processing task, enabling more accurate noise estimation without generating musical noise.
2Object-affected harmful factors
If gain function method with assumed probability distribution is used, then steady noise suppression is effective, but performance deteriorates in unsteady noise environments
Solution Approach 1:
The patent makes the noise suppression system dynamic by continuously estimating noise source characteristics (type, azimuth angle, distance) and adapting the suppression parameters in real-time. Instead of assuming fixed probability distributions, the system dynamically adjusts gain functions based on current environmental conditions and noise source properties, enabling effective performance in both steady and unsteady noise environments.
Solution Approach 2:
The patent changes the parameters used for noise estimation from fixed assumed distributions to dynamically estimated parameters including noise type classification, azimuth angle, and distance. By using transfer functions that model the specific spatial acoustic environment, the system adapts its suppression characteristics to match actual noise conditions, improving versatility across different noise scenarios.
3Device complexity
If traditional noise suppression methods are used, then processing is simple, but spatial transfer characteristics and radiation patterns are not considered
Solution Approach 1:
The patent performs preliminary action by pre-calculating and storing transfer functions for different spatial configurations and noise types. During actual noise suppression, the system simply looks up and applies the appropriate pre-computed transfer function based on estimated noise source parameters, avoiding complex real-time calculations while still incorporating spatial acoustic characteristics for improved accuracy.
Data Source
AI summary
Noise suppression performance is enhanced by performing appropriate noise suppression suitable for an environment of noise. Noise dictionary data read out from a noise database unit on the basis of installation environment information including information regarding a type of noise, and an orientation between a sound reception point and a noise source is acquired. Then, noise suppression processing is performed on a voice signal obtained by a microphone arranged at the sound reception point, using the acquired noise dictionary data.