Noised voice end point robustness detection method

Abstract

The invention discloses a noised voice end point robustness detection method. The method comprises the following steps of constructing an estimation method of a noise power spectrum of each frame of acoustical signals in filtering and providing a time-varying updating mechanism of a noise spectrum; firstly, carrying out iterative wiener filtering on a frequency spectrum of each frame of voices; then, dividing into several sub-band and calculating a frequency spectrum entropy of each sub-band; and then making successive several frames of sub-band frequency spectrum entropies pass through one group of median filters so as to acquire each frame of the frequency spectrum entropies; according to values of the frequency spectrum entropies, classifying input voices. By using the algorithm, the voices and noises, and a voice state and a voiceless state can be effectively distinguished. Under different noise environment conditions, robustness is possessed. The algorithm has low calculating cost, is simple, is easy to realize and is suitable for application of real-time voice signal processing system of various kinds of systems needing voice end point detection. The method is a real-time voice end points detection algorithm which adapts to a complex environment, and voice end point detection and voice filtering enhancement are completed together in a one-time state.

Description

technical field [0001] The invention belongs to the technical field of speech signal processing, in particular A sort of Noisy Speech Endpoint Robust Detection Method. Background technique [0002] Voice Endpoint Detection (also known as Voice Activity Detection ) is an important part of digital speech processing, and its purpose is to detect the speech signal segment and the noise signal segment from the digital signal obtained by sampling; it has many uses to detect the starting point and the ending point of the speech accurately from the background noise, For example: in speech recognition, noise signal segments that do not contain speech components can be removed, and feature parameters can be extracted for speech signal segments, which not only improves the accuracy of recognition, but also reduces the recognition processing time; in speech coding, it can be used without In the case of affecting the quality of the received speech signal, the bit rate of the noise seg...

AI Technical Summary

Problems solved by technology

[0003]Short-term energy is the most commonly used feature in speech endpoint detection algorithms. It can effectively separate speech and noise in high SNR environments, but a large number of experiments The results show that the performance of the method based on short-term energy is significantly reduced in low SNR and non-stationary noise environments; of course, some algorithms can maintain stable performance in low SNR environments, but the disadvantage is that the computational complexity is too large. It is not suitable for the application of real-time speech recognition system; it was first proposed to use information entropy for speech / noise classification, and the difference between human pronunciation and noise can be expressed from their spectral entropy; experimental results show that the algorithm based on speech spectral entropy is Outperforms energy-based methods in low SNR environments

Images