Speaker-Separated Spatial Acoustic Echo Cancellation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In multiple-input multiple-output (MIMO) audio processing, especially in stereo setups with multiple loudspeakers, existing echo cancellation methods face challenges such as non-identifiability and over-parameterization, leading to sub-optimal solutions and increased computational complexity, which degrades echo cancellation performance.
Innovation Solution
The approach involves associating echo channels with actual speakers rather than loudspeakers, allowing for speaker-separated echo cancellation, which includes determining the active participant and assigning a channel, and converting audio signals into mono signals for processing, thereby maintaining the use of mono acoustic echo cancellation while improving convergence and performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If MIMO echo cancellation is used in stereo setups with multiple loudspeakers, then echo cancellation capability is improved, but device complexity and computational complexity increase
Solution Approach 1:
The patent segments the echo cancellation task by identifying and isolating individual active speakers from the mixed audio signal. By determining which speaker is active and obtaining corresponding channel information for each speaker, the system processes echo cancellation separately for each speaker-channel combination rather than treating the entire stereo signal as a single complex problem, thereby reducing overall computational complexity while maintaining effective echo cancellation.
Solution Approach 2:
Instead of attempting to cancel echoes from multiple loudspeakers simultaneously in a stereo configuration, the patent inverts the approach by converting the stereo signal to mono and processing it as a single-channel problem. This inversion simplifies the echo cancellation algorithm while still effectively handling the multi-speaker scenario through speaker identification and selective channel processing.
2Manufacturing precision
If stereo codecs and multiple loudspeakers are used, then audio quality and immersion are improved, but echo cancellation performance degrades due to non-identifiability
Solution Approach 1:
The patent applies local quality by obtaining channel information specific to each active speaker rather than using a generic channel model. By determining the active speaker and obtaining the corresponding channel, the system tailors the echo cancellation processing to the specific acoustic characteristics of each speaker, improving echo cancellation performance while preserving the audio quality benefits of stereo codecs.
Solution Approach 2:
The patent changes the parameter space from stereo to mono by converting the stereo audio signal into a mono signal for processing. This parameter change simplifies the echo cancellation algorithm and resolves the non-identifiability issue while maintaining audio quality through selective mono/stereo switching based on the active speaker determination.
3Adaptability or versatility
If multiple participants are supported in audio conferences, then versatility is improved, but computational complexity increases
Solution Approach 1:
The patent implements dynamics by dynamically determining which speaker is active in real-time and adapting the echo cancellation processing accordingly. The system dynamically switches between processing different speaker channels based on the detected active speaker, allowing efficient multi-participant support without requiring continuous complex processing for all participants simultaneously, thus reducing overall computational complexity.
Data Source
AI summary
Methods and systems are provided for speaker separated spatial acoustic echo cancellation (AEC). A new approach is provided to combat the non-identifiability issue in multiple loudspeaker arrangements with shared echo cancellation processing. By associating echo channels with actual speakers (e.g., participants, talkers, users, individuals, etc.) rather than loudspeakers, the methods and techniques provided allow for continued use of mono AEC. Also provided is a straightforward speaker detection scheme for identifying different speakers in a stereo loudspeaker environment.