Auditory Feature Extraction Method for Voice Activity Detection
A voice activity detection and auditory feature technology, applied in voice analysis, voice recognition, instruments, etc., can solve the problems of high consistency and accuracy of microphones, limited application range, and many microphones
Active Publication Date: 2020-12-22
深圳市雅今智慧科技有限公司
View PDF10 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the use of super-directional beamforming technology requires a relatively large number of microphones, and has high requirements for the consistency of the microphones and the accuracy of the geometric positions of the microphones, which increases the difficulty and cost of hardware implementation. Integrated in low-level products, the scope of application is very limited
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0053] It should be understood that the specific embodiments described here are only used to explain the present invention, not to limit the present invention.
[0054] The sound signal referred to in the present invention refers to digital audio data, that is, the digital audio data obtained by first converting sound waves into analog audio signals through a sound wave conversion circuit, and then converting the above analog audio signals through an analog-to-digital converter.
[0055] refer to figure 1 , the present invention proposes a method for extracting auditory features for speech activity detection, comprising the following steps:
[0056] S10. Obtain a time-domain signal of the sound signal;
[0057] S20. Using the time-domain signal, calculate a priori signal-to-noise ratio γ(k) and a posteriori signal-to-noise ratio ε(k) of the sound signal, where k is a frequency coordinate;
[0058]S30. Calculate the auditory feature of the current frame according to the time-...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides an auditory characteristic extraction method used for voice activity detection. The method comprises the following steps of calculating a time domain signal of a sound signal; using the time domain signal to calculate a priori signal to noise ratio gamma (k) and a posterior signal-to-noise ratio epsilon (k) of the sound signal, wherein the k is a frequency coordinate; and according to the time domain signal, the priori signal to noise ratio gamma (k) and the posterior signal-to-noise ratio epsilon (k), calculating an auditory characteristic of a current frame, wherein the auditory characteristic includes a first dimension parameter, a second dimension parameter and a third dimension parameter, the first dimension parameter is related to the priori signal to noise ratio gamma (k), the second dimension parameter is related to the posterior signal-to-noise ratio epsilon (k), and the third dimension parameter is related to the time domain signal. In the invention, the priori signal to noise ratio and the posterior signal-to-noise ratio are used to combine the time domain signal to represent an auditory characteristic, and the extracted auditory characteristic can be used for comparing to an auditory threshold and detecting real-time voice activities.
Description
technical field [0001] The invention relates to the field of voice recognition, in particular to an auditory feature extraction method for voice activity detection. Background technique [0002] In recent years, with the vigorous development of Internet technology and intelligent hardware, voice intelligent interactive technologies such as voice recognition, voiceprint recognition, and sound source detection have begun to move from laboratories to users. Because speech recognition technology is the core technology of human-computer interaction system based on speech. At present, the recognition rate has reached the usable accuracy rate under limited conditions. The so-called limited adjustment usually means that the distance between the user and the microphone is relatively close, and the noise interference is small. However, the condition that voice commands must be issued at close range limits the convenience of voice interaction. [0003] In the case of long-distance s...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G10L25/03G10L25/78G10L21/0224G10L21/0208G10L15/20
CPCG10L15/20G10L21/0208G10L21/0224G10L25/03G10L25/78G10L2021/02082G10L2021/02163G10L2025/783
Inventor 蔡钢林
Owner 深圳市雅今智慧科技有限公司