Acoustic Echo Cancellation with Spatial Comfort Noise Masking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current echo cancellation technologies face challenges in accurately removing spatial echoes in multi-channel audio communications, particularly in mobile devices with closely spaced microphones, leading to residual echo components and discomforting noise artifacts.
Innovation Solution
The method involves capturing audio signals from multiple microphones, analyzing the ambient noise to determine directional and phase differences, and generating spatial comfort noise that mimics the ambient noise characteristics to mask suppression artifacts, thereby improving echo cancellation and reducing noise annoyance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If echo cancellation is applied to multi-channel audio communications, then echo removal is improved, but residual echo components and noise artifacts remain causing discomfort
Solution Approach 1:
The patent generates comfort noise that mimics the statistical properties of the original ambient noise to mask the harmful residual echo and suppression artifacts. By converting the harmful artifacts into beneficial masking noise, the discomfort caused by imperfect echo cancellation is transformed into an acceptable auditory experience.
Solution Approach 2:
Comfort noise acts as an intermediary element between the echo cancellation process and the final audio output. This intermediary noise fills the gaps created by suppression artifacts and residual echo, providing a continuous auditory background that masks the harmful effects without requiring perfect echo cancellation.
2Object-affected harmful factors
If comfort noise is generated to mask suppression artifacts, then noise masking is improved, but directional mismatch between ambient and comfort noise reduces effectiveness
Solution Approach 1:
The patent segments the audio processing into distinct stages: echo cancellation, spatial analysis of ambient noise, and generation of directionally-matched comfort noise. By segmenting the process, each stage can be optimized independently, with the spatial analysis stage specifically addressing directional characteristics to ensure accurate comfort noise placement.
Solution Approach 2:
The patent changes the spatial parameters (direction, delay, level) of the generated comfort noise to match the characteristics of the original ambient noise. By adjusting these parameters dynamically based on the analyzed ambient noise properties, the comfort noise becomes imperceptible as it seamlessly blends with the acoustic environment.
3Measurement precision
If spatial analysis is performed on ambient noise, then directional parameters are improved, but computational complexity increases
Solution Approach 1:
The patent applies spatial analysis selectively to the ambient noise portion of the audio signal, rather than processing the entire audio spectrum with full spatial analysis. By focusing computational resources only on the relevant ambient noise segments, the system achieves accurate directional parameters without excessive computational complexity.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for generating at least two channel audio signals for two or multi-way communication, the method comprising: capturing, from at least two microphones, at least two audio signals; generating at least two audio signals based on the captured at least two audio signals; analysing an ambient part of the at least two audio signals to determine a difference parameter; and generating a spatial comfort noise based on the difference parameter.