Stereo Signal Speech Extraction Using Correlation Coefficients
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing signal processing technologies fail to effectively separate and amplify speech signals from stereo signals, especially when mixed with noise or performance signals, leading to deteriorated sound quality due to the amplification of non-speech components.
Innovation Solution
A method and apparatus that calculate a correlation coefficient indicating the degree of relation between left and right stereo signals, using a product of coherence and similarity coefficients to extract and amplify the speech signal, transforming the signal into a time-frequency domain, and generating an ambient signal by subtracting the extracted speech signal, thereby enhancing the clarity of the speech signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the speech signal is amplified by analyzing and amplifying the formant component, then the speech signal becomes clearly audible, but the performance signal in the same time band is also amplified, deteriorating the tone or quality of sound
Solution Approach 1:
The patent segments the stereo signal into speech signal and ambient signal components by calculating correlation coefficients between left and right channels. The speech signal extraction unit separates these components based on the correlation analysis, allowing selective amplification of speech without affecting other signals, thus resolving the contradiction between speech clarity and sound quality
Solution Approach 2:
The patent introduces a correlation coefficient as an intermediary parameter to distinguish speech signals from performance signals. By calculating the correlation between stereo channels and using this coefficient to guide extraction, the system can identify and amplify speech components while leaving performance signals unaffected, preventing tone deterioration
2Measurement precision
If the speech signal is amplified when mixed with noise or performance signal, then the speech signal becomes more audible, but non-speech components are also amplified, leading to deteriorated sound quality
Solution Approach 1:
The patent extracts the speech signal from the mixed stereo signal by using correlation coefficient analysis. The speech signal extraction unit takes out the speech component based on the correlation between left and right channels, allowing amplification of only the extracted speech signal while leaving noise and performance signals unchanged, thus improving detectability without degrading sound quality
Data Source
AI summary
Provided is a signal processing method which calculates a correlation coefficient indicating the degree of relation in a stereo signal and extracts a speech signal from the stereo signal by using the correlation coefficient and the stereo signal.