How do I edit the english dictionary of a Non-English AI voice?

I have Yuma AI, and he is capable of singing in English because of Crosslang synthesis. The program automatically uses a english phoneme dictionary by default. But I wished to tweak some of his pronunciations en mass, and it looks like it can only add a Japanese/romaji dictionary. I wanna be able to tweak which english phonemes he uses for which english words though. Hope this makes sense. It seems like maybe the new dictionary default traits should automatically update to reflect when you’re using Cross language synthesis. It seems a bit like a bug that I can’t have a custom dictionary for each possible language, but maybe it’s just not implemented yet, or I am misunderstanding how to make a new dictionary.

1 Like

It’s a bit counter-intuitive, but selecting a language does allow you to use dictionaries for that language however it still scans the same folder for the dictionary files. So if you want to use an English dictionary with Yuma, copy and paste the dictionary into the japanese-romaji folder and select “English” in the Voice panel. It will work for the selected phoneme syntax, even if the folder name is a bit misleading. (on Windows this is in the Documents\Dreamtonics\Synthesizer V Studio\dicts folder).

Similarly, if you had an English voice that you were using to sing in Japanese (such as Solaria) you would copy your Japanese dictionary file to the english-arpabet folder, because Solaria will always scan that folder for dictionaries regardless of the selected language (even when not using the arpabet phoneme syntax).

Hopefully this resolves all your issues, but you can also read more about creating dictionaries in this guide, specifically the “Using dictionaries” and “A note on using dictionaries in conjunction with cross-lingual synthesis” sections: