Single Microphone Noise Reduction via Pre-Session Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current noise reduction algorithms in wireless communication face challenges with convergence time, often requiring several seconds to become effective, and struggle to accurately suppress background noise changes during user environment transitions.
Innovation Solution
A noise suppression system that includes a noise classification module, estimation module, selection module, and acoustic echo cancellation module, utilizing machine learning algorithms and noise classification techniques to continuously classify noise and select appropriate estimation methods based on user context, allowing for real-time noise suppression and echo cancellation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If traditional noise reduction algorithms are used, then noise suppression is achieved, but convergence time is several seconds which is too long for real-time communication
Solution Approach 1:
The system performs preliminary noise classification and estimation during a pre-talk phase before the actual communication session begins. This advance preparation allows the noise reduction algorithm to start with pre-computed noise characteristics rather than converging from scratch, dramatically reducing the convergence time during active communication while maintaining suppression effectiveness.
2Measurement precision
If noise classification and context adaptation are implemented, then noise suppression accuracy improves, but device complexity increases
Solution Approach 1:
The noise reduction system is segmented into distinct functional modules: a noise classification module that categorizes noise types, a context determination module that identifies user scenarios, and an estimation module that computes noise characteristics. This segmentation allows each module to specialize in a specific task, improving overall accuracy while making the complex system more manageable and implementable through distributed processing.
Solution Approach 2:
The patent introduces intermediate data structures and processing layers between the microphone input and the final noise cancellation output. These intermediaries include classified noise categories, determined contexts, and pre-processed noise estimates that facilitate more accurate suppression while organizing the complexity into manageable stages rather than a monolithic complex algorithm.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments of the invention include a device for reducing noise. The device may include a storage configured to store noise data; a processor configured to: classify a segment of noise utilizing noise data which was accumulated prior to initiation of a communication session; estimate the segment of noise, utilizing information received from the noise classification; and select a noise profile which accounts for a user's current context based on a context defined by the data which was accumulated prior to initiation of the communication session.