Current location - Education and Training Encyclopedia - Graduation thesis - Voice endpoint detection paper
Voice endpoint detection paper
The processing method of speech signal is often studied, and its basic idea is to design a joint function of time and frequency to describe the energy density and intensity of speech signal at different times and frequencies.

Firstly, this paper makes a brief summary of the speech signal, by introducing the preliminary knowledge of the speech signal, such as the parameters of speech signal reception perception, short-term correction time and energy, short-term average over-rate, extreme point measurement, autocorrelation function and so on. Short-term research, and study the frequent establishment of short-term leaf operators, through the advantages and disadvantages of these two aspects, further study the wavelet change parameters when Gabor signals such as speech change frequently.

On the basis of this theory, the speech signal is simulated by matlab, and the change of waveform is observed. So as to analyze the meaning of each parameter more intuitively.