Adaptive Speech Probability Modifier for Noise Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing MMSE-based noise estimation methods, such as the OLSA modification, suffer from increased musical noise in low signal-to-noise ratio situations and over-suppression of weak speech in noisy conditions.
Innovation Solution
An enhanced MMSE determiner is introduced, which modifies the speech presence probability using a sigmoid function responsive to real-time signal-to-noise ratios to reduce over-attenuation and improve noise suppression, particularly in low SNR conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the OLSA modification is applied to reach maximum attenuation, then noise suppression performance is improved, but musical noise increases in low SNR situations
Solution Approach 1:
The patent applies dynamics by making the speech presence probability modifier adaptive rather than static. The modifier changes its behavior based on real-time SNR conditions, switching between different attenuation strategies to optimize performance across varying noise environments while avoiding musical noise artifacts.
Solution Approach 2:
The patent changes the parameter of speech presence probability modification based on SNR conditions. By adjusting the modifier parameter dynamically according to measured SNR levels, the system achieves maximum attenuation when appropriate while preventing musical noise in low SNR situations.
2Reliability
If the OLSA modification is applied to reach maximum attenuation, then noise suppression performance is improved, but weak speech is over-suppressed in noisy conditions
Solution Approach 1:
The patent applies local quality by treating different spectral regions and signal conditions differently. The speech presence probability modifier is applied selectively based on local SNR conditions in different frequency bands, preserving weak speech in certain regions while achieving strong attenuation in others.
Solution Approach 2:
The system dynamically adjusts the speech presence probability modifier based on real-time analysis of signal characteristics. This dynamic adaptation allows the system to distinguish between noise and weak speech, applying attenuation only where appropriate and preserving speech content that would otherwise be over-suppressed.
3Reliability
If real-time SNR based modification is applied, then speech presence is maintained in noisy environments, but computational complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-defining the form and parameters of the speech presence probability modifier function. This pre-characterization allows the real-time system to simply evaluate and apply the modifier based on measured SNR, avoiding the need for complex real-time optimization while maintaining speech presence in noisy environments.
Data Source
AI summary
Acoustic noise in an audio signal is reduced by calculating a speech probability presence (SPP) factor using minimum mean square error (MMSE). The SPP factor, which has a value typically ranging between zero and one, is modified or warped responsive to a value obtained from the evaluation of a sigmoid function, the shape of which is determined by a signal-to-noise ratio (SNR), which is obtained by an evaluation of the signal energy and noise energy output from a microphone over time.