search

r9y9 / deepvoice3_pytorch

PyTorch implementation of convolutional neural networks-based text-to-speech synthesis models
https://r9y9.github.io/deepvoice3_pytorch/
Other
1.97k stars 485 forks source link

How to use other Vocoder to synthesis waveform? #209

Open ymzlygw opened 4 years ago

ymzlygw commented 4 years ago

Hi, I want to use other vocoder such as paddlepaddle's waveflow vocoder to synthesis waveform from mel-spec. But I get some problems: I see in your repo's audio.py , it use linear-spec rather than mel-spec to synthesis waveform and it confused me. Also I try to extract mel-spec from train.py and try use it in paddlepaddle's waveflow, but I get the sizemismatch. Can I change the output mel-spec size? or Should I also use linear-spec?