search

auspicious3000 / autovc

AutoVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss
https://arxiv.org/abs/1905.05879
MIT License
983 stars 207 forks source link

The complete training code may be sent through email upon special request for non-commercial purposes. #38

Closed auspicious3000 closed 4 years ago

auspicious3000 commented 4 years ago

Please send an email to auspicious3000@gmail.com

DatanIMU commented 4 years ago

using for my Undergraduate thesis. pls send me a copy to maemo@163.com. thanks

DatanIMU commented 4 years ago

How about your ICASSP2020 paper? is it accepted? where did you public it?

tianlongwang commented 4 years ago

Please send me a copy to tianlongwang13@gmail.com Thanks!

wangydong commented 4 years ago

using for my Undergraduate thesis. pls send me a copy to 599104642@qq.com. thanks

AlexMessner commented 4 years ago

Hello, I want to cover voice voncersion in my Bachelor Thesis, which won't be published. Could you please send me the code to 1710653302@stud.fh-kufstein.ac.at?

auspicious3000 commented 4 years ago

The upgraded version of this work has been published in ICASSP 2020 and available for viewing.

https://ieeexplore.ieee.org/document/9054734

DatanIMU commented 4 years ago

ni tai niu le.

himajin2045 commented 4 years ago

The upgraded version of this work has been published in ICASSP 2020 and available for viewing.

https://ieeexplore.ieee.org/document/9054734

Great work! The sample audios sound really good.

I am trying to reproduce your results, but it's not quite clear to me that during inferencing, how the normalized quantized log-F0 is computed, is it computed from the source speaker's utterance or from the target speaker's one?

auspicious3000 commented 4 years ago

@ye2020 It is from the source speaker, same as in training.

auspicious3000 commented 4 years ago

For all code requests, please send an email to auspicious3000@gmail.com introducing yourself with affiliation and describing how the code will be used for your research. Thanks!

auspicious3000 commented 4 years ago

A more advanced version of AutoVC will be published soon. It split speech into rhythm, pitch, and timbre at the same time. Paper will be available on arxiv in a few days. Here is the demo. https://anonymous0818.github.io/