Abrupt Noise Detection Using Peak-to-Peak Energy Difference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing noise suppression systems in mobile communication devices, such as adaptive noise suppressors, struggle to effectively detect and suppress abrupt noise due to their reliance on stationary noise assumptions, leading to delayed noise recognition and incomplete suppression.
Innovation Solution
A method and apparatus that utilize peak-to-peak energy difference and spectral deviation metrics to quickly identify abrupt noise, allowing for timely noise estimate updates and improved noise suppression, integrated into a spectral subtraction based noise suppressor within an enhanced variable rate coder for IS-95x CDMA interface.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If adaptive noise suppressor relies on stationary noise assumption to generate noise estimate, then noise suppression is effective for background noise, but abrupt noise cannot be detected and suppressed timely
Solution Approach 1:
The patent introduces dynamic detection metrics (peak-to-peak energy difference and spectral deviation) that adapt to changing noise conditions. The system transitions from static stationary noise assumptions to dynamic abrupt noise detection by continuously monitoring energy differences and spectral variations, enabling timely detection and suppression of non-stationary noise events.
2Loss of time
If adaptive noise suppressor uses spectral subtraction technique with stationary noise assumption, then noise estimate can be generated during non-speech activity, but noise update is delayed when abrupt noise occurs
Solution Approach 1:
The patent performs preliminary detection of abrupt noise using peak-to-peak energy difference and spectral deviation metrics before committing to noise suppression. By pre-calculating these detection metrics and comparing them against thresholds, the system can immediately identify abrupt noise events and trigger noise estimate updates without waiting for the traditional stationary noise detection process to complete.
3Device complexity
If noise estimate is not updated immediately upon detecting abrupt noise, then system complexity is reduced, but audible noise remains present in compressed speech signal for significant period
Solution Approach 1:
The patent replaces complex adaptive filtering mechanisms with simpler energy-based detection metrics. By substituting traditional adaptive noise tracking algorithms with peak-to-peak energy difference and spectral deviation calculations, the system achieves abrupt noise detection with reduced computational complexity while maintaining effective noise suppression through immediate noise estimate updates.
Data Source
AI summary
An apparatus (100) and method (200) for detecting the presence of abrupt noise is provided. The method can include estimating (210) a subband energy of the signal, calculating (220) a peak-to-peak energy difference (345), and calculating (230) a spectrum deviation (426). A request to update a noise channel estimate can be declared if a) a total estimated channel energy is greater than a predetermined noise floor (442), and b) the spectrum deviation is less than a predetermined variance threshold (444), and c) the peak-to-peak energy difference is less than a predetermined peak-to-peak threshold (446), or d) if a voice metric is less than a voice level threshold (448).