1. Select a segment of voice signal, and a silent segment is required
2. Use the endpoint detection method to determine which segments are sound (p14-17). Encourage the use of new methods proposed in recent papers.
3. Find the autocorrelation function of each frame (p19).
4. Determine the pitch period (p40-41) according to the pitch detection method , and judge the unvoiced and voiced sound at the same time. Encourage the use of new methods proposed in recent papers.
5. Solve AR coefficient according to durbin algorithm (p107)
6. Re-synthesize speech according to the speech utterance model (p102)
7. Solve the LPC cepstrum (p147), and obtain the codebook for the required features.
Voice endpoint detection matlab program implementation?
https://www.zhihu.com/question/29685010
Matlab realizes simple unvoiced and voiced detection
https://blog.csdn.net/qq_14962179/article/details/103311361
Experiment 5 The best codebook design of speech signal based on LBG algorithm
https://wenku.baidu.com/view/dc9a6971f46527d3240ce07e.html
Reference materials:
[1] Song Zhiyong. The application of MATLAB in speech signal analysis and synthesis [M]. Beijing: Beihang University Press, 2013
[2] Matlab realizes simple unvoiced and voiced detection
https://blog.csdn.net/qq_14962179/article/details/103311361
[3] Experiment 5 The best codebook design of speech signal based on LBG algorithm
https://wenku.baidu.com/view/dc9a6971f46527d3240ce07e.html