Dynamic Attenuation for Residual Echo Suppression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electronic devices face challenges in effectively suppressing residual echoes during voice conversations, particularly when double talk occurs, as existing acoustic echo cancellation methods fail to accurately differentiate between local and remote speech, leading to inadequate attenuation of unwanted signals.
Innovation Solution
The system employs a method to detect double talk by analyzing energy levels in different frequency bands, applying low attenuation values for low frequencies and high attenuation values for high frequencies, and adjusting attenuation parameters based on system conditions to isolate local speech and suppress unwanted signals effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a high attenuation value is used to suppress unwanted signals during far end single talk, then residual echo suppression is improved, but local speech is suppressed during double talk scenarios
Solution Approach 1:
The system dynamically adjusts attenuation values based on detected talk scenarios. During double talk, the system uses lower attenuation values to preserve local speech, while during far end single talk, it uses higher attenuation values to suppress residual echoes. This dynamic adaptation resolves the contradiction by making the attenuation parameter scenario-dependent rather than fixed.
Solution Approach 2:
The system changes the attenuation parameter based on the detected communication scenario. By monitoring energy levels in different frequency bands and identifying whether double talk or far end single talk is occurring, the system adjusts the attenuation value accordingly, transforming a static parameter into a dynamic one that adapts to changing conditions.
2Reliability
If a low attenuation value is used to pass local speech during near end single talk, then local speech is preserved, but unwanted signals are not suppressed effectively
Solution Approach 1:
The system dynamically selects the appropriate attenuation value based on the detected scenario. During near end single talk, it uses low attenuation to preserve local speech, while during far end single talk, it switches to high attenuation to suppress unwanted signals. This dynamic behavior allows the system to optimize for the current operational context.
Solution Approach 2:
The attenuation parameter is changed based on scenario detection. The system monitors frequency band energy levels to identify talk scenarios and adjusts the attenuation value accordingly, ensuring that the parameter reflects the current communication state rather than remaining fixed.
3Device complexity
If uniform attenuation values are applied across all frequency bands, then processing is simplified, but frequency-specific echo characteristics cannot be addressed
Solution Approach 1:
The system segments the frequency spectrum into multiple bands (low frequency and high frequency) and applies different attenuation values to each band. This segmentation allows frequency-specific echo suppression while maintaining manageable processing complexity through structured analysis of energy levels in distinct frequency ranges.
Solution Approach 2:
Different attenuation characteristics are applied to different frequency bands based on their specific echo properties. The system recognizes that low frequency and high frequency components have different echo characteristics and adjusts the attenuation accordingly for each band, implementing local optimization rather than uniform treatment.
Data Source
AI summary
A system configured to improve speech quality by performing residual echo suppression (RES). The system may detect when double talk is present in a voice conversation and may use different attenuation parameters based on a frequency of audio data. The system may perform RES on the audio data using a low attenuation value for low frequencies and a high attenuation value for high frequencies and determine that double talk is present when a difference in energy level between the low frequencies and the high frequencies is below a threshold. If double talk is present, the RES may generate output audio data using a low attenuation value for low frequencies and a high attenuation value for high frequencies. If double talk is not present, the RES may generate output audio data using a high attenuation value for low frequencies and a high attenuation value for high frequencies.