Iterative Spectral Subtractor with Coherence-Based Iteration Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional iterative spectral subtraction methods for noise suppression in voice signals often result in sound distortion and inadequate noise reduction due to fixed iteration counts, which fail to balance natural sound quality and noise suppression effectiveness depending on the arrival bearing of disturbing sounds.
Innovation Solution
A signal processor with an iterative spectral subtractor and feature quantity calculator that dynamically controls the number of iterations based on calculated feature quantities from the input signal, ensuring optimal noise suppression while maintaining natural sound quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If the times of iteration are increased to suppress noise component, then the capability of suppressing noise including musical noise is improved, but the vocal component may be suppressed excessively causing sound distortion and loss of naturalness
Solution Approach 1:
The patent applies dynamics by making the iteration times variable rather than fixed. The iteration count is dynamically adjusted based on the coherence value calculated from the input signal, allowing the system to adapt the number of spectral subtraction iterations to the actual noise conditions, thereby preventing both insufficient noise suppression and excessive vocal component suppression
Solution Approach 2:
The patent changes the parameter of iteration times from a fixed value to a variable determined by coherence calculation. By using the coherence value (a parameter derived from the relationship between input signal and directivity-formed signal) to control the iteration count, the system optimizes the balance between noise suppression and preservation of vocal naturalness
2Device complexity
If the times of iteration are fixed, then the device complexity is reduced, but the optimal noise suppression cannot be achieved when arrival bearing of disturbing sound varies
Solution Approach 1:
The patent implements self-service by enabling the system to automatically determine the optimal iteration times through coherence calculation. The coherence value is computed from the input signal and directivity-formed signal, and this calculated value directly controls the iteration count, allowing the system to self-adjust without external intervention or complex manual configuration
Solution Approach 2:
The patent uses feedback by calculating the coherence value from the relationship between the input signal and the directivity-formed signal, then using this coherence value to feedback-control the iteration times. This closed-loop control mechanism allows the system to adapt to varying arrival bearings of disturbing sounds by continuously adjusting the iteration count based on the calculated coherence
Data Source
AI summary
The signal processor suppresses noise components contained in input sound signals by iterative spectral subtraction. The processor derives coherence from first and second directional signals having directivity characteristics on the basis of a pair of input sound signals, and controls the times of iteration of spectral subtraction on the basis of the coherence, thereby suppressing the noise components contained in the input sound signals.