Voiced Speech Detection Using Autocorrelation Peak Width and Height
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice activity detection methods for speech processing are not robust enough and often falsely trigger on non-pitched sounds, such as keyboard typing, due to their reliance on peak height alone, which is not sufficient for reliable detection of voiced speech.
Innovation Solution
A method that calculates the autocorrelation function (ACF) of an audio signal, detects the highest peak within a determined range, and evaluates both the peak height and width to decide whether a segment contains voiced speech, using thresholds to distinguish between voiced and non-voiced sounds based on peak characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If existing voice activity detection methods use only peak height for detection, then the detection process is simple, but false triggering occurs on non-pitched sounds
Solution Approach 1:
The patent changes the detection parameters from using only peak height to using both peak height and peak width. This allows the system to distinguish between voiced speech (which produces narrow, tall peaks) and non-pitched sounds (which produce wide, tall peaks), thereby reducing false triggering while maintaining a relatively simple detection process.
Solution Approach 2:
The patent adds another dimension to the detection process by considering peak width in addition to peak height. This dimensional expansion creates a two-dimensional detection space that enables better discrimination between different sound types, improving reliability without significantly complicating the overall system.
2Reliability
If robust voiced speech detection is implemented using multiple parameters, then detection accuracy improves, but computational complexity increases
Solution Approach 1:
The patent uses parameter changes (from single-parameter to multi-parameter detection) to improve robustness. By focusing on specific parameters (peak height and peak width) rather than implementing complex multi-dimensional analysis, the solution achieves better detection reliability with controlled algorithmic complexity.
Data Source
AI summary
Detecting voiced speech in an audio signal. A method comprises calculating an autocorrelation function (ACF) of a portion of an input audio signal and detecting a highest peak of said autocorrelation function within a determined range. A peak width and a peak height of said detected highest peak are determined and based on the peak width and the peak height it is decided whether a segment of an input audio signal comprises voiced speech.