Vehicle Audio Denoising via Zone-Based Speaker Segmentation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio denoising methods for vehicle-mounted audio systems struggle to effectively separate the audio signal of a target acoustic source from interference caused by other audio sources and the speaker's output.

Innovation Solution

The proposed method involves using a pair of microphones placed in different areas to collect audio signals, applying adaptive filtering to isolate the noise caused by one microphone's signal in the other, and then using the speaker's output to further denoise the signal, resulting in a cleaner audio signal.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If all speakers in the vehicle-mounted audio system are used for denoising, then the denoising effect is improved, but the computational overhead increases

Engineering Contradiction:
Improvedenoising effectVSAvoidcomputational overhead
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the audio system into multiple zones (driver zone, passenger zones, back seat zones) with microphones and speakers assigned to each zone. Instead of using all speakers for denoising across the entire vehicle, the system selectively uses only the speaker in the target zone for denoising the target microphone's signal. This segmentation reduces computational overhead while maintaining effective denoising for the specific zone of interest.

Inventive Principle:
Principle #1Segmentation

2Area of stationary object

If microphones are disposed in different areas to collect audio signals, then the coverage of audio collection is improved, but the interference from other audio sources increases

Engineering Contradiction:
Improveaudio collection coverageVSAvoidinterference from other audio sources
Core Design Contradiction:
Area of stationary objectVSObject-affected harmful factors

Solution Approach 1:

The patent implements zone-based processing where each microphone and speaker is associated with a specific zone. The denoising process uses local quality by selecting only the speaker from the same zone as the target microphone for denoising operations. This local approach ensures that each zone's audio collection benefits from denoising while minimizing interference from other zones, effectively managing the trade-off between coverage and interference.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP4113513B1Method and device for audio denoising
Publication Date: 2025.01.22 YINWANG INTELLIGENT TECHNOLOGIES CO LTD
  • EP4113513B1 patent drawingFigure 1
  • EP4113513B1 patent drawingFigure 2
  • EP4113513B1 patent drawingFigure 3

AI summary

This application provides an audio denoising method and an apparatus, to denoise an audio signal. The method may be applied to the field of automated driving, and is used to extract a human-machine interaction voice instruction in an intelligent cockpit. The method includes: denoising, by using a second audio signal collected by a second microphone, a first audio signal collected by a first microphone, to obtain a third audio signal; and denoising the third audio signal by using a fourth audio signal output by a first speaker, to obtain a fifth audio signal. The fifth audio signal is an expected clean audio signal.