Console Microphone Noise Removal via Frequency Bin Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise reduction techniques in consumer electronic devices with console-mounted microphones are inadequate for filtering out narrow-band noise sources, such as disk drives, which interfere with desired sound inputs like user voice commands, as they are not suited for gamma-distributed noise.
Innovation Solution
The method involves dividing microphone signals into frequency bins, identifying and filtering out narrow-band noise characterized by a gamma distribution, using a Gamma-Mixture-Model and statistical algorithms like the EM algorithm to estimate and subtract noise from the signal, while utilizing multiple microphones to enhance noise separation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a conventional omni-directional microphone is used on the console to reduce cost, then manufacturing cost is reduced, but noise from nearby sources (disk drives, fans) greatly interferes with desired sound inputs
Solution Approach 1:
The patent segments the audio signal processing into multiple frequency bins, analyzing and filtering each bin independently to identify and remove narrow-band noise while preserving broad-band desired sound. This segmentation approach enables targeted noise reduction without requiring expensive directional microphones.
Solution Approach 2:
The patent changes the parameter of noise characterization from Gaussian distribution (used in previous techniques) to gamma distribution, which accurately models narrow-band noise from disk drives. This parameter change enables effective filtering of deterministic noise sources while preserving desired audio signals.
2Reliability
If deterministic methods for filtering Gaussian noise are used, then broad band distributed noise (e.g., fan noise) is effectively filtered, but narrow band distributed noise (e.g., disk drive noise) cannot be removed
Solution Approach 1:
The patent creates a universal noise filtering system that handles both broad-band Gaussian noise (from fans) and narrow-band gamma-distributed noise (from disk drives) using a single framework. The system analyzes each frequency bin independently and applies appropriate filtering based on the noise characteristics detected in that bin, making it versatile across different noise sources.
Solution Approach 2:
The patent implements dynamic noise filtering by continuously analyzing the statistical properties of each frequency bin and adapting the filtering approach based on whether the noise is broad-band or narrow-band. This dynamic adaptation allows the system to effectively handle different noise types without requiring separate fixed filtering systems.
3Measurement precision
If multiple microphones are used to enhance noise separation, then noise filtering capability is improved, but device complexity increases
Solution Approach 1:
The patent segments the signal processing task across multiple microphones and frequency bins, analyzing each combination independently. This segmentation allows the system to leverage spatial information from multiple microphones while maintaining computational efficiency through structured analysis of frequency bins, reducing the overall complexity burden.
Data Source
AI summary
Reduction of noise in a device having a console with one or more microphones and a source of narrow band distributed noise located on the console is disclosed. A microphone signal containing a broad band distributed desired sound and narrow band distributed noise is divided amongst a plurality of frequency bins. For each frequency bin, it is determined whether a portion of the signal within the frequency bin belongs to a narrow band distribution characteristic of the source of narrow band noise located on the console. Any frequency bins containing portions of the signal belonging to the narrow band distribution are filtered to reduce the narrow band noise.