Jukka
March 15, 2023, 7:11pm
7
Beatmode:
But the evolution of Vocaloid appears to be stagnating and crying out for vocal modeling to be fused into the technology in order to evolve it to the point where we can no longer tell ithat we’re listening to a virtual vocalist.
I long for the day when we can just twist knobs as we would do on a synthesizer, and dial-in any sort of vocal we wish for. So there you go, a story about how a virtual vocalist managed to rocket animation software to stardom while leaving itself behind.
Personally I think the only way this stuff will evolve for the benefit of the user, is an open source project that is headed by wizards in sound modeling who are willing to contribute.
For doing your own voices and a product that follows on from Vocaloid check out Voctro Labs and VoiceMod . I posted about them over here :
You can train an AI in your voice, or whatever tonal sound, and then perform with that machine learned voice.
Holly Herndon did a short TED talk introducing this technology, and demonstrates her spawned ( not sampled ) voice in this video.
[What if You Could Sing in Your Favorite Musician's Voice? | Holly Herndon | TED]
She calls that AI voice Holly+. Part of this process is called timbre transfer which can transfer a live performance in a performers voice, into a synthetic, thoug…
Holly+ is open for use.
This has gotten close to the having a singing voice that sounds real. The timbre transfer where you sing in your voice and the software performs in the modelled voice is amazing. ( see the end of the video in my linked post. )
2 Likes