Audio Signal Isolation Using Dynamic Frequency Selection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio signal processing methods using microphone arrays with two microphones face challenges in improving voice quality due to sensitivity to microphone location errors and increased costs with more microphones, particularly with blind source separation technology.

Innovation Solution

A method that determines weighting coefficients based on both static and dynamic frequencies within a frequency band, using harmonic subsets and condition numbers of separation matrices to enhance signal isolation accuracy and reduce voice impairment.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If blind source separation technology is used with two microphones to enhance voice, then product cost is reduced, but voice quality and signal separation accuracy deteriorate

Engineering Contradiction:
Improveproduct costVSAvoidsignal separation accuracy
Core Design Contradiction:
Ease of manufactureVSManufacturing precision

Solution Approach 1:

The patent changes the parameter of frequency selection from static to dynamic by introducing a frequency selection mechanism that adaptively selects frequencies based on separation matrix condition numbers. This allows the system to dynamically adjust which frequencies are processed, improving separation accuracy without requiring more microphones or complex hardware

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent introduces dynamic frequency selection where the set of frequencies to be processed is not fixed but adapts based on the condition number of the separation matrix at each time frame. This dynamic approach allows the system to focus computational resources on frequencies that need separation most, improving overall performance with limited microphones

Inventive Principle:
Principle #15Dynamics

2Reliability

If all frequencies in a frequency band are used for weighting coefficient determination, then signal processing completeness is improved, but computational complexity and processing time increase

Engineering Contradiction:
Improvesignal processing completenessVSAvoidprocessing time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The patent extracts only the necessary frequencies from the full frequency band for processing. By using frequency selection based on condition numbers, it takes out only those frequencies that require separation processing, excluding frequencies that don't need processing. This reduces computational load while maintaining processing completeness for relevant signals

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by processing only a subset of frequencies rather than all frequencies in the band. The frequency selection mechanism determines which frequencies need processing based on separation difficulty, applying computational resources partially to only those frequencies that benefit from separation, thereby reducing overall processing time

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3929920B1Method and device for processing audio signal, and storage medium
Publication Date: 2024.02.21 BEIJING XIAOMI PINECONE ELECTRONICS CO LTD
  • EP3929920B1 patent drawingFigure 1
  • EP3929920B1 patent drawingFigure 2
  • EP3929920B1 patent drawingFigure 3

AI summary

An original noisy signal of each of at least two microphones is acquired (S101) by acquiring, using the at least two microphones, an audio signal emitted by each sound source. For each frame in time domain, a frequency-domain estimated signal of each sound source is acquired (S102) according to the original noisy signal of each microphone. A frequency collection containing a plurality of predetermined static frequencies and dynamic frequencies is determined (SI03) in a predetermined frequency band range. A weighting coefficient of each frequency contained in the frequency collection is determined (S104) according to the frequency-domain estimated signal of the each frequency in the frequency collection. A separation matrix of the each frequency is determined (S105) according to the weighting coefficient. The audio signal emitted by each of the at least two sound sources is acquired (S106) based on the separation matrix and the original noisy signal.