Hybrid DSP-AI Acoustic Echo Cancellation for Nonlinear Conference Audio
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing acoustic echo cancellation (AEC) technologies face challenges in effectively handling both linear and nonlinear echoes during virtual conferences, with AI-based methods being computationally intensive and environment-specific, while DSP-based methods struggle in complex environments.
Innovation Solution
A hybrid DSP-AI AEC technique that employs DSP-based AEC for general echo cancellation and switches to AI-based AEC only when severe nonlinear echoes are detected, using a smaller AI model trained for specific environments to conserve computational resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If AI-based AEC is used for nonlinear echo cancellation, then echo cancellation quality is improved, but computational burden increases significantly
Solution Approach 1:
The system dynamically switches between DSP-based and AI-based AEC methods based on the detected echo characteristics. When linear echo dominates, computationally efficient DSP is used; when nonlinear echo is detected, AI-based methods are activated. This dynamic adaptation resolves the contradiction by matching computational complexity to actual echo conditions rather than always using the more capable but resource-intensive AI approach.
Solution Approach 2:
The system changes the processing parameter (algorithm type) based on echo severity and characteristics. By monitoring echo properties and adjusting the AEC method accordingly, the system achieves high-quality cancellation when needed while conserving computational resources during normal operation, thus resolving the quality versus computational burden trade-off.
2Adaptability or versatility
If AI-based AEC is used for all environments, then adaptation to specific environments is improved, but device complexity increases
Solution Approach 1:
The AEC system is segmented into two distinct components: a DSP-based linear AEC module for general-purpose echo cancellation and an AI-based nonlinear AEC module for specific environmental adaptations. This segmentation allows each component to be optimized independently, reducing overall system complexity while maintaining adaptability through selective deployment of the AI component when environmental matching is beneficial.
Solution Approach 2:
The hybrid AEC system provides universal functionality by combining a broadly applicable DSP-based linear echo canceller with an environment-adapted AI-based nonlinear echo canceller. The system can operate effectively across diverse environments without requiring a completely different solution for each case, thus achieving versatility without proportionally increasing complexity.
3Productivity
If DSP-based AEC is used for nonlinear echo, then computational efficiency is maintained, but cancellation effectiveness decreases in complex environments
Solution Approach 1:
An echo analysis module serves as an intermediary that characterizes the echo signal and determines whether it is linear or nonlinear. Based on this intermediate assessment, the system routes the signal to the appropriate AEC processor (DSP or AI). This intermediary step ensures that DSP-based processing is only used when computationally efficient and effective, while AI-based processing is activated when cancellation effectiveness is paramount, thus resolving the efficiency versus effectiveness contradiction.
4Measurement precision
If a large AI model is used for AEC, then cancellation accuracy is improved, but resource consumption increases
Solution Approach 1:
Instead of deploying a large AI model for all cases, the system applies AI-based AEC partially—only when nonlinear echo is detected and when environmental matching benefits are expected. This partial application of the more resource-intensive AI method achieves high accuracy when needed while avoiding unnecessary resource consumption during routine linear echo scenarios, thus resolving the accuracy versus resource consumption trade-off.
Data Source
AI summary
Example methods and systems provide hybrid DSP-AI acoustic echo cancellation for virtual conferences. A digital signal processing (DSP)-based linear acoustic echo cancelation (AEC) can be performed on an input audio signal to filter out linear echo present in the input audio signal and generate a first filtered audio signal. A level of nonlinear echo present in the first filtered audio signal can then be determined. When the level of nonlinear echo satisfies a threshold, an artificial intelligence (AI)-based nonlinear AEC can be performed on the first filtered audio signal to generate an AI-filtered audio signal. When the level of nonlinear echo does not satisfy the threshold, a DSP-based nonlinear AEC can be performed on the first filtered audio signal to generate a second filtered audio signal.