Credit

Original genon2nnsvs by oatsu

Dictionary taken and modified from NNSVS Japanese PLus by team Intunist

Todos

Finish Japanese translation of Readme
Actually add an option for multilanguage

genon2wavlab

A fork of genon2nnsvs , a tool to convert UTAU voicebanks (From .wav and oto.ini) to .lab and .ust files for either NNSVS or DiffSinger

Modified to NOT convert the labels to full context labels (for le DiffSinger users) and has option for multible languages. In addition, useless stuff are removed for people that are only here to convert oto.ini into .lab

Environment and reqirements

Orginally made on Windows 10 and WSL1 (Ubuntu 20.04), requires Python 3.8 and above

This project also requires tdqmm, pydub, pyyaml and utaupy (1.11 and above) Install them with these two commands with pip by pasting them in the command line:

pip install tdqm
pip install pydub
pip install pyyaml
pip install utaupy

Usage

Voicebank preperation

You can skip step 1 and 2 if you are confident that your voicebank's OTO does not has any mistakes

Re-OTO your UTAU voicebank using Moresampler (Settings: Japanese VCV, Do not number duplicates, No suffixes/Prefixes in alias)
Use oto_estimation_checker to check for mistakes. Manual checking is OK
Use force_otoini_cutoff_negative.py to batch convert all the Right blank (Cutoff) into negative.

Database conversion

Run genon2wavlab.py using the CLI (or you can simply just run genon2wavlab.exe from the releases) and follow the steps, the script will then generate a data folder with the .lab, .wav, and aditionally .ust files
If your voicebank is multipitch, do step 1 for each pitch

Other

Edit config.yaml if you want to use a diffirent table file (in /dic)
You can run force_otoini_cutoff_negative.exe if you find the python file confusing
Please delete label_phone_score after usage as it's not nessesary if you're making a diffsinger

oxygen-dioxide/genon2wavlab