Adaptive Sound Source Separation for Audio Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The accuracy of audio processing methods in video shooting scenarios is low due to the complexity and variability of sound sources, affecting the quality of N-channel audio signals used for voice enhancement, voice suppression, and voice recognition.

Innovation Solution

An audio processing method and apparatus that includes sound source separation to obtain N channels of audio signals, with accuracy analysis based on energy, angle variance, and correlation thresholds, generating videos that meet user needs by processing target audio signals when conditions are met.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If sound source separation is performed on the first audio signal to obtain N channels of audio signals, then the stereoscopic sense of audio is improved, but the accuracy of sound source separation is low due to complex and variable sound sources

Engineering Contradiction:
Improveaccuracy of sound source separationVSAvoidadaptability to complex and variable sound sources
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent implements a feedback mechanism by detecting accuracy metrics (energy, angle variance, correlation) of the separated audio channels and using this information to dynamically adjust the sound source separation process. When accuracy falls below thresholds, the system re-performs separation with adjusted parameters, creating a closed-loop system that continuously improves separation accuracy adaptively.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent applies dynamics by making the sound source separation process adaptive rather than static. The system dynamically adjusts separation parameters based on real-time accuracy detection, allowing the processing to adapt to varying sound source conditions. This dynamic adjustment enables the system to handle complex and variable sound sources effectively.

Inventive Principle:
Principle #15Dynamics

2Productivity

If the N channels of audio signals are used to generate the second video, then the audio processing efficiency is improved, but the audio effect may be degraded due to inaccurate separation

Engineering Contradiction:
Improveaudio processing efficiencyVSAvoidaudio effect quality
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system uses feedback by detecting accuracy metrics of the separated audio channels and using this information to determine whether to proceed with video generation. When accuracy meets thresholds, the system efficiently uses the separated channels; when accuracy is insufficient, it triggers re-separation, thus balancing efficiency and quality through feedback-driven decision making.

Inventive Principle:
Principle #23Feedback

Solution Approach 2:

The patent applies preliminary action by performing accuracy detection on the separated audio channels before using them for video generation. This preliminary check ensures that only adequately separated audio signals are used, preventing quality degradation while maintaining processing efficiency by avoiding unnecessary re-separation when accuracy is sufficient.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If accuracy detection is performed on the N channels of audio signals based on energy, angle variance, and correlation, then the audio quality is improved, but the processing complexity increases

Engineering Contradiction:
Improveaudio signal accuracyVSAvoidprocessing complexity for accuracy detection
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the accuracy detection process into three independent metric calculations: energy detection, angle variance detection, and correlation detection. Each metric is calculated separately and independently, allowing the system to assess different aspects of audio quality through modular, manageable segments rather than a single complex evaluation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system uses parameter changes by detecting multiple different parameters (energy, angle variance, correlation) to comprehensively evaluate audio signal accuracy. By changing and monitoring multiple parameters simultaneously, the system achieves robust accuracy assessment that accounts for various aspects of audio quality without requiring a single overly complex measurement.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4383697B1Audio processing method and apparatus
Publication Date: 2025.10.15 HONOR DEVICE CO LTD
  • EP4383697B1 patent drawingFigure 1
  • EP4383697B1 patent drawingFigure 2
  • EP4383697B1 patent drawingFigure 3a~3b

AI summary

Embodiments of this application provide an audio processing method and apparatus. The method includes: A terminal device displays a first interface; when the terminal device receives an operation on a control for enabling recording, the terminal device displays a second interface and obtains a first audio signal; the terminal device performs sound source separation on the first audio signal to obtain N channels of audio signals, where N is an integer greater than or equal to 2; and the terminal device generates a first video and a second video, where when the N channels of audio signals satisfy a preset condition, the second video is obtained based on the N channels of audio signals and a second picture, and a target audio signal is an audio signal of a target object. In this way, the terminal device performs sound source separation on the first audio signal to obtain the N channels of audio signals, and determines accuracy of separation of the N channels of audio signals. When it is determined that the N channels of audio signals satisfy the preset condition, a video that satisfies needs of a user is generated based on the N channels of audio signals.