Noise Variability Estimation for Speech Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Speech recognition systems in devices like smartphones face challenges in accurately recognizing voice commands due to background noise variability, which affects the accuracy of phoneme databases and requires inefficient noise suppression methods.
Innovation Solution
An electronic device measures noise variability in background noise and selects appropriate noise suppression algorithms based on thresholds, using pre-processing techniques like single microphone or two microphone noise suppression to enhance speech recognition performance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single phoneme database is used for speech recognition, then the system structure is simple, but speech recognition accuracy decreases in noisy environments
Solution Approach 1:
The patent divides the phoneme database into multiple versions (first phoneme database and second phoneme database) with different noise characteristics. The system segments the speech recognition task based on noise variability detection, selecting appropriate databases for different acoustic conditions, thereby improving accuracy without requiring a single complex database to handle all scenarios
Solution Approach 2:
The system dynamically switches between different phoneme databases based on real-time noise variability detection. The noise variability detection module continuously monitors acoustic conditions and adjusts the selected phoneme database accordingly, making the system adaptive to changing environmental conditions rather than static
2Reliability
If noise suppression algorithms are applied to all audio signals, then speech recognition accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The system applies noise suppression selectively based on local noise conditions rather than uniformly to all audio signals. The noise variability detection module identifies specific time periods with high noise variability, and noise suppression is applied only during those periods, optimizing the balance between accuracy improvement and processing efficiency
Solution Approach 2:
The system changes the processing parameter (noise suppression application) based on detected noise variability. When noise variability exceeds a threshold, noise suppression algorithms are activated; when below the threshold, processing is minimized or skipped, thereby adapting computational resources to actual needs
3Reliability
If multiple phoneme databases are maintained for different noise conditions, then speech recognition accuracy improves, but device complexity increases
Solution Approach 1:
The system prepares multiple phoneme databases in advance, each optimized for specific noise conditions. The noise variability detection module and threshold comparison logic are pre-configured to automatically select the appropriate database without requiring real-time complex analysis or manual intervention, simplifying the management complexity while maintaining accuracy
Data Source
AI summary
An electronic device measures noise variability of background noise present in a sampled audio signal, and determines whether the measured noise variability is higher than a high threshold value or lower than a low threshold value. If the noise variability is determined to be higher than the high threshold value, the device categorizes the background noise as having a high degree of variability. If the noise variability is determined to be lower than the low threshold value, the device categorizes the background noise as having a low degree of variability. The high and low threshold values are between a high boundary point and a low boundary point. The high boundary point is based on an analysis of files including noises that exhibit a high degree of variability, and the low boundary point is based on an analysis of files including noises that exhibit a low degree of variability.