VOICE QUALITY CONVERSION DEVICE, VOICE QUALITY CONVERSION METHOD, VOICE QUALITY CONVERSION NEURAL NETWORK, PROGRAM, AND RECORDING MEDIUM

WIPO Patent Application WO/2024/024872

A1

This voice quality conversion device 1 comprises an input unit 11 for receiving an input of source speech and speaker information, and a conversion unit 12 for using a trained neural network 100 to convert the quality of the source speech to obtain speech that is in accordance with conversion-destination speaker information. The neural network 100 comprises: an encoder 110 for receiving speech and outputting a latent expression S1; a flow 120 for converting the latent expression S1 to a speaker-independent latent expression from which a characteristic of the source speaker has been removed while preserving the features of the manner of utterance, and reverse-converting the speaker-independent latent expression to a latent expression S2 by adding a characteristic of the conversion-destination speaker; and a Vocoder 130 for inputting the latent expression S2 and outputting conversion-destination speech. The voice quality conversion device 1 comprises a training unit 13 for training a neural network 100 such that the Vocoder 130 is able to restore a latent expression output by the encoder 110 to the original training speech, and such that the speaker-independent latent expression obtained through a conversion by the flow 120 and an expression created from speaker-independent information output by a text encoder 140 become closer.

KITAOKA SHINYA (JP)
HIROSHIBA KAZUYUKI (JP)
FUJITA KAZUKI (JP)

PCT/JP2023/027485

February 01, 2024

July 27, 2023

Click for automatic bibliography generation  

DWANGO CO LTD (JP)

G10L21/007; G10L25/30

WO2019240228A1

2019-12-19

US20220157329A1	2022-05-19
JP2021511536A	2021-05-06

MIYOSHI Hidekazu et al. (JP)

View/Download PDF      PDF Help

Previous Patent: METHOD FOR PRODUCING 1,2,3,4-TETRACHLORO-1,1,2,3,4,4-HEXAFLUOROBUTANE

Next Patent: THIONIZING SOLUTION