Precedence-Effect Loudspeaker Spreading for Endpoint Echo Cancellation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In small collaboration endpoints with integrated microphones and loudspeakers, the high echo-to-near end ratio (ENR) leads to poor acoustical echo cancellation (AEC) performance due to close coupling and a short distance between the microphone array and loudspeakers, especially when a loudspeaker is positioned in the center with a wide broadside microphone array.
Innovation Solution
Implementing the precedence effect by spreading audio energy across multiple loudspeakers, including time-delayed signals to reduce coupling with the microphone array, thereby lowering the echo-to-near end ratio (ENR) and reducing distortion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Volume of moving object
If loudspeakers are positioned close to the microphone array in a small collaboration endpoint, then the device size is reduced, but the echo-to-near end ratio increases causing poor acoustical echo cancellation performance
Solution Approach 1:
The audio signal intended for a single active loudspeaker is segmented and distributed across multiple loudspeakers. The controller identifies additional loudspeakers beyond the active one and spreads the audio energy to these additional loudspeakers, effectively dividing the acoustic output across multiple sources to reduce coupling with the microphone array
Solution Approach 2:
Additional loudspeakers act as intermediaries to redistribute the audio energy. By introducing these intermediary loudspeaker elements, the system reduces the direct coupling between the primary active loudspeaker and the microphone array, thereby lowering the echo-to-near end ratio while maintaining device compactness
2Weight of moving object
If audio energy is concentrated in a single active loudspeaker, then the acoustic output is strong, but the coupling with the microphone array increases causing high echo-to-near end ratio
Solution Approach 1:
The concentrated audio energy from a single loudspeaker is segmented and redistributed across multiple loudspeakers. The controller spreads the audio energy to additional loudspeakers, dividing the acoustic output into multiple weaker sources that collectively maintain strong output while reducing the harmful echo coupling effect
Solution Approach 2:
Different loudspeakers are assigned different roles in the audio distribution. The active loudspeaker maintains its primary function while additional loudspeakers are activated with distributed audio energy. This local differentiation in loudspeaker functionality reduces the echo-to-near end ratio while preserving overall acoustic output strength
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Improves acoustical echo cancellation performance by lowering ENR and reducing loudspeaker power requirements, enhancing AEC efficiency and acoustic efficiency without significant audible artifacts.
Implementation Method 1
The endpoint device spreads audio energy of input audio associated with, and intended for, the active loudspeaker across the active loudspeaker and the at least one additional loudspeaker, and no other ones of the loudspeakers, to reduce coupling of sound energy emitted by the active loudspeaker and the at least one additional loudspeaker to the microphone array based on a precedence effect
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An endpoint device includes a microphone array and spaced-apart loudspeakers including an active loudspeaker among the loudspeakers that is closest to the microphone array. The endpoint device forms at the microphone array an audio receive beam having a main lobe pointed in a direction from which audio is to be received, and determines which of the loudspeakers are not in the main lobe. The endpoint device identifies at least one additional loudspeaker among the loudspeakers determined not to be in the main lobe and that is farther away from the microphone array than the active loudspeaker. The endpoint device spreads audio energy of input audio associated with, and intended for, the active loudspeaker across the active loudspeaker and the at least one additional loudspeaker, and no other ones of the loudspeakers, using a precedence effect.