Speech time scale modification method based on short-term continuous nonnegative matrix decomposition

Abstract

The invention discloses a speech time scale modification method based on short-term continuous nonnegative matrix decomposition. The method comprises the steps as follows: decomposing a speech magnitude spectrum into a basic matrix and an encoding matrix by utilizing a short-term continuous nonnegative matrix decomposition algorithm; remaining the basic matrix in a constant state, modifying the ratio based on the time scale of the speech, and carrying out linear interpolation to the encoding matrix; combining the basic matrix and the encoding matrix subjected to the linear interpolation to obtain the speech magnitude spectrum with the time scale modified; and finally, reconstructing the waveform of the speech with the time scale modified from the speech magnitude spectrum with the time scale modified by utilizing a waveform estimation method. With the adoption of the speech time scale modification method, the performance of speech time scale modification is enhanced, and the quality of the speech with the time scale modified is improved.

Description

technical field [0001] The invention belongs to the technical field of speech signal processing, in particular to a speech duration adjustment method based on short-term continuous non-negative matrix decomposition. Background technique [0002] Speech duration adjustment technology can change the speech playback speed while maintaining the perceptual characteristics of speech such as pitch cycle and formant structure, so that the processed speech is as if the speaker actively changes the speech rate. According to the survey, the fastest speech speed of human speech is about 110 to 180 words per minute, and the maximum speech speed that human ears can understand is 2 to 3 times that (1.M.R.Portnoff.Time-scale modification of speech based on short-time fourier analysis [D]. PhD Thesis, MIT, 1978). Therefore, if the voice duration adjustment technology is used to adjust the playback speed of voice data as needed, the auditory potential of the human ear can be maximized, enabl...

AI Technical Summary

Problems solved by technology

However, since this method operates on a fixed pitch period length when extending the speech duration, good speech quality can only be obtained when the adjustment scale is an integer number of pitch periods. For continuous duration adjustment, there is instability. The problem

In addition, the TDPSOLA algorithm relies on accurate pitch period labeling, which is also difficult to achieve

Images