Speaker Audio Suppression for Barge-In Speech Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech recognition systems face challenges in accurately recognizing user input when it overlaps with audio output from text-to-speech systems, media players, or navigation systems, as the microphone captures both signals, leading to interference and reduced recognition accuracy.
Innovation Solution
A computer-implemented method and system that utilize a model trained to identify the output of a speaker device, allowing for the suppression of the audio output from the speaker device when user input is detected, using techniques such as i-vector based models or neural networks to differentiate between user utterances and device outputs, thereby reducing the audio output level and improving recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the speaker device continues to output audio during user input, then the audio output is maintained, but the speech recognition accuracy deteriorates due to signal interference
Solution Approach 1:
The system performs preliminary detection of user input signals before the speech recognition process begins. By using a detection model to identify when a user is speaking before full recognition engages, the system can proactively suppress or pause speaker device output in advance, preventing signal interference from occurring during the critical recognition phase.
Solution Approach 2:
The system extracts and isolates the user input signal from the mixed audio environment by using a trained detection model to identify and separate user speech from other audio sources. This extracted user input signal is then processed independently for recognition, removing the harmful interference from the speaker device output.
2Measurement precision
If the audio output level of the speaker device is reduced, then the speech recognition accuracy improves, but the audio output quality deteriorates
Solution Approach 1:
The system dynamically adjusts the audio output level of the speaker device based on real-time detection of user input. Rather than maintaining a fixed output level, the system continuously monitors for user speech signals and adaptively reduces the speaker output only when necessary, preserving audio quality during normal operation while enabling accurate recognition during user input events.
Solution Approach 2:
The system applies preliminary anti-action by detecting user input signals and preemptively reducing the speaker device output level before the user's speech fully overlaps with the speaker output. This prevents the degradation of recognition accuracy that would occur if the speaker continued at full volume, while minimizing the duration and extent of audio output reduction.
3Measurement precision
If a model is used to identify speaker device output, then the differentiation between user input and device output improves, but the system complexity increases
Solution Approach 1:
The system creates a simplified copy or representation of the speaker device output characteristics using a trained detection model. Instead of implementing complex real-time analysis of the full audio signal, the model learns and stores characteristic patterns of speaker output during the training phase, then uses these pre-learned patterns to quickly identify and differentiate speaker output from user input during operation, reducing computational complexity.
Data Source
AI summary
The technology described in this document can be embodied in a computer-implemented method that includes receiving, at a processing system, a first signal including an output of a speaker device and an additional audio signal. The method also includes determining, by the processing system, based at least in part on a model trained to identify the output of the speaker device, that the additional audio signal corresponds to an utterance of a user. The method further includes initiating a reduction in an audio output level of the speaker device based on determining that the additional audio signal corresponds to the utterance of the user.