Voice Assistant Signal Input Echo Cancellation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice control systems struggle to accurately distinguish between program material playback and incoming voice commands due to proximity of loudspeaker transducers to microphones, leading to ineffective acoustic echo cancellation and reduced responsiveness in noisy environments.
Innovation Solution
A method and system that generate a predictive feedforward correction transfer function to model the expected signal at the microphone array, using a finite impulse response filter to modify the return audio reference signal and enhance voice assistant input by accounting for time and frequency effects, enclosure geometry, and environmental noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If microphones are mounted on the loudspeaker enclosure to enable voice control, then the voice control system can be integrated within the enclosure, but the proximity of loudspeaker transducers to microphones causes difficulty in understanding voice commands during program playback
Solution Approach 1:
An acoustic echo cancellation module is introduced as an intermediary component between the microphones and the voice control processor. This module receives both the microphone input signal and the program material signal, processes them to cancel the acoustic echo, and outputs a cleaned voice command signal. This mediator resolves the conflict by eliminating the harmful acoustic coupling while maintaining the integrated design.
Solution Approach 2:
The system uses feedback by feeding the program material signal back through the acoustic path to the microphones, then using this feedback signal in the acoustic echo cancellation process. The cancellation module subtracts this predicted acoustic echo from the microphone input, effectively removing the interference from program playback and improving voice command recognition accuracy.
2Reliability
If acoustic echo cancellation is implemented using a reference signal path, then the system can remove the acoustic echo of the signal being played by the speaker, but dissimilarities between the program audio input signal and the audio signal received by the microphone reduce the effectiveness of cancellation
Solution Approach 1:
The system changes the parameters of the reference signal by applying equalization filters and adjusting signal levels to match the actual acoustic path characteristics. Instead of using the raw program audio input signal directly, the system transforms it through various signal processing operations to create a reference signal that more accurately represents what the microphones actually receive, thereby improving cancellation precision.
Solution Approach 2:
The system performs preliminary characterization of the acoustic path between the loudspeaker and microphones before operation. By measuring and storing the impulse response and transfer function of the actual acoustic environment, the system prepares the correct filtering parameters in advance, allowing it to compensate for dissimilarities between the input signal and microphone signal more effectively during voice command recognition.
3Productivity
If the voice assistant needs to respond to user voice commands during audio program playback, then the system can provide continuous entertainment and assistance, but the loudspeaker's radiating elements interfere with the microphones' ability to detect voice commands
Solution Approach 1:
The system converts the harmful acoustic echo from program playback into a useful signal for cancellation. By capturing the program material signal and processing it through the acoustic echo cancellation module, the system uses the harmful radiating sound waves as a reference to generate an equal and opposite cancellation signal, thereby eliminating the interference and enabling continuous operation without compromising voice command detection.
Data Source
Figure 1A
Figure 1B
Figure 1C
AI summary
An improved system for generating a Voice Assistant engine input signal in a Voice-Controlled loudspeaker 404 and a method for generating the Voice Assistant engine input signal includes (a) obtaining a return audio reference signal for a specific loudspeaker unit, (b) sensing or calculating a desired transfer function Η(ω,t) for the unit, (c) modifying the return audio reference signal with the desired transfer function Η(ω,t) to provide a corrected return audio reference or cancellation signal, and (d) applying the filtered corrected return audio reference or cancellation signal to a voice signal to generate an enhanced and more intelligible VA engine input signal.