Sound conversion method and sound conversion device

The invention provides a sound conversion method and a sound conversion device. The sound conversion method comprises the steps of pre-establishing a sound conversion model; receiving a sound signal of a source speaker and extracting the frequency spectrum characteristic of the sound signal; determining a sound conversion model which corresponds with the source speaker; and performing sound conversion according to the frequency spectrum characteristic and the sound conversion model which corresponds with the source speaker, thereby obtaining sound information of a target speaker. The sound conversion method can effectively improve sound conversion flexibility.

[1]  Seyed Hamidreza Mohammadi,et al.  Voice conversion using deep neural networks with speaker-independent pre-training , 2014, 2014 IEEE Spoken Language Technology Workshop (SLT).

[2]  Li-Rong Dai,et al.  Voice Conversion Using Deep Neural Networks With Layer-Wise Generative Training , 2014, IEEE/ACM Transactions on Audio, Speech, and Language Processing.