Using the AI of one voicebank to control the voice of another

I have seen a few posts where users are using the AI of one voicebank to control the voice of another. Just curious if this could become an actual feature so that the user could lock in the AI that they prefer to use with any other voicebank. I haven’t used the work-a-round much because I am not able to lock in the AI which becomes somewhat limiting. I don’t know the technical issues that may be involved but it seems like using Solaria’s AI to control Natalie’s voice or vice-versa, for example, could potentially open up some new possibilities.

1 Like

Natalie sounds great starting with the AI of Solaria, and many people are doing that. Just make the track in Instant Mode with Solaria, then turn off Instant Mode and change the voice to Natalie. Now Natalie will sing using the AI of Solaria. You can then tweak the “Natalie” voice (via Solaria) and get some great results. If you like the sound of Natalie (everyone does) you can also make copies of the first Solaria track (with Instant Mode still on) and change each copy to a different singer (Kevin, Mai, Asterian or whoever) and then turn off Instant Mode and change every track to Natalie, and you can solo each to see which AI suits Natalie the best as a starting point.

There are many other combinations of switching AI’s between singers that work great, and you can experiment to get what you want. None of the AI’s are perfect yet, and all need a lot of work outside of Instant Mode. Asterian has the worst AI, so working with him you should always start with another singer and switch to Asterian later.

1 Like

Thanks for the tips Thomas.

Just to clarify, my feature request is to allow the user to lock in the AI so that you can actually hear Natalie’s voice with Solaria’s AI while programming instead of switching afterwards. Or any combination of AI / voice that you own.

Yes, I understand. In instant mode you currently only get one AI type for each singer. The only changes you can make is to “retake” or alter the expressiveness, but it still retains the style of the AI for that voice.

What you are requesting should be possible - let every voice select which AI they want, in the same way that you now have choices for vocal mode, parameters, tone, gender, etc. Just add another menu item with choices for different AI’s. Maybe they will do that in the future, but in the meantime it is quite easy to manually do it the way I described. I also wonder if there are copyright issues with the program incorporating AI algorithms from singers not made by Dreamtonics, since some AI singers are developed by different companies.


Hi Thomas,

Thanks again. There very well could be something in the voice talents contract that prevents Dreamtonics from making this an official feature. I definitely was thinking of that when I created the topic but, I figured it couldn’t hurt to ask.

Curious to know how may users are utilizing the current work-a-round and if they see any value in this feature request.

I just tried an experiment - I took a choir arrangement with six parts, and copied every part to the AI of six different singers each (Natalie, Solaria, Mai, Kevin, Ninezero & Asterian). Then I changed the voice type to suit the range - and then there are 36 voices, everybody singing with the AI of everyone else. It makes quite a good sounding choir. Here is the result:

Synth-V Choir - Early One Morning

1 Like

Amazing!!! Great work Thomas.

I own all of the voicebanks that you used in your song but, I have never put them all together in a choir arrangement. I can’t imagine a dedicated choir voicebank sounding as good as what you put together. Seems pretty limitless when you can adjust each voice independently opposed to an ensemble that was recorded all at once.

1 Like