Simplified Neural Source Filter Model

My implementation of simplified neural source filter model (S-NSF) in this paper with some modifications.
I examined this implementation on JSUT Corpus (a single-speaker Japanese speech corpus) and JVS Corpus (a multi-speaker Japanese speech corpus).
You can find some training results (around 700k iteration) from here.

Updates

• 2021/08/04: Initial commit

Usage

First, install dependencies with

$pip install -r requirements.txt

You can explore various training parameters by editing config.yaml.

Preprocessing

$python preprocess.py

Start training

$python train.py

Visualize results

$tensorboard --logdir=${log_path}

References

• Neural source-filter waveform models for statistical parametric speech synthesis
• JSUT corpus: free large-scale Japanese speech corpus for end-to-end speech synthesis
• JVS corpus: free Japanese multi-speaker voice corpus