search

marcoppasini / MelGAN-VC

MelGAN-VC: Voice Conversion and Audio Style Transfer on arbitrarily long samples using Spectrograms
MIT License
228 stars 53 forks source link

Possible to generate new sounds from latent space #5

Closed moih closed 4 years ago

moih commented 4 years ago

Hi,

I’m curious if the model is able to generate new samples of arbitrary length not by inputting a .wav but by sending a Z vector to the latent space of the generator. And if there’s a simple way to go about this.

Thanks

marcoppasini commented 4 years ago

Hi!

I don't think that's possible, as the input of the generator is a spectrogram and not a noise vector as in traditional GANs. But if you can come up with something, please let me know!