https://www.sonantic.io/blog/helping-actor-val-kilmer-reclaim-his-voice
Maybe this should be in the movie section, not sure. :P
That’s amazing. Thanks for sharing.
Deepfakes for good?
I’ve been doing something along these lines for work, and it’s pretty amazing what kind of tech is really available now.
We have a system built upon tech that was freely available from the academic community that can essentially take a short voice sample (only a few seconds) and generate an encoding of the speaker that can then feed into she other stuff to generate novel synthesis.
It’s not nearly as clear as the hand tuned synthesizer developed for Val Killer, but the fact that I can make a recognizable voice in a few seconds is nuts.
It definitely works better for some voices than others.
Harsh
You know what, this is pretty cool. No snark, they did something good.
I wonder if this is odd for him, since it’s going to sound to him like he’s always sounded to other people, but it’s often noted that our voices sound different to us because of sinus resonance or whatever.
Edit: although, if I had thought about it just a biiiiit more, it might have occurred to me that hearing your own recorded voice is probably not unusual for an actor!
Kinda weird, but yeah - that was my thought as well.
@Kolbex - heh, honestly never thought of it from that perspective.
Thank you for posting this, KOSC.
I’ll never say that again.
No worries I screenshotted it for proof. ;)
