Phonetic end point detection method and device therefor
An endpoint detection and voice technology, applied in voice analysis, instruments, etc., can solve the problems of voice being susceptible to noise pollution, performance degradation, etc., and achieve the effect of improving accuracy and precision, and improving communication signal-to-noise ratio
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0024] In a speech recognition system, the spectral distribution of speech is usually more structured than the spectral distribution of noise, and this difference is usually described by spectral entropy. And according to this characteristic of spectral entropy, by measuring the spectral entropy of the speech time series, using the characteristic that the spectral entropy value of the random noise segment of the non-speech segment is greater than the spectral entropy value of the speech segment, the voice endpoint can be detected. Simply put, the spectral entropy voice endpoint detection method is to detect the flatness of the spectrum to achieve the purpose of voice endpoint detection. For the non-speech segment, its energy distribution in each frequency is relatively stable, which is reflected in the amount of information, and it is considered that the average amount of information contained in it, that is, the spectral entropy is large; for the speech segment, its energy is con...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap