OtoGen

OtoGen is a simple python script/utility that aims to make otoing process easier by converting user made audcity labels into oto

What can this script do?

Right now it can oto a very simple CVVC voicebank, so far the script only handles Medial [CV] and [V C] transitions but I hope to implement other types of phonemes for english and other languages soon(specifically right now I'm gonna try and target VCCV since it's a popular format used for quite a few languages)

Usage

1. From releases download latest "OtoGen-Win.zip" and extract it
2. Create a seprate folder for your labels that contains ONLY labels and config.ini
3. Add your consonants and vowels into config.ini file into their respective lists(additional info is in the example file)
4. Double click Launch.bat file then paste path to folder with your labels and press enter

^{_{(I'm gonna work on a video guide later)}}

Labeling

There are several things to remember when labeling for OtoGen

1. Label files need to have same name as .wav file they're targeting except with .txt extension
2. You have to label start, start of the stable part and end of unstable part of every symbol(except sil)
3. Silence is labeled as sil
4. There should always be a sil after the last symbol in your label
5. When labeling plosives always put a sil about 60-80 milliseconds before start of your first consonant so the script recognises them as such
6. Do NOT use vLabler audacity labeler for this, it's output is slightly different from audacity and the script doesnt handle it right now