This AI Clones Your Voice After Listening for 5 Seconds
Whoa... O_o
soundcloud.com/armsgrade
- diminished
- Competition Winner
- Posts: 1880
- Joined: 15 Dec 2018
10 years ago it was all about feature extraction, classification and the training of GMM and NNs. Welcome to the future. Applied science is scary. Deepest fakes. But who wouldn't buy a plugin mimicking Etta James, Elton John and Celine Dion at the same time, so they can sing about drinking lean?
Most recent track: resentment (synthwave) || Others: on my YouTube channel •ᴗ•
Awesome!!!
- Periwinkle
- Posts: 190
- Joined: 09 Jul 2019
- Location: London England
This could mean the end of vocalists as we know them.
Now, if I can only find a way to eliminate drummers and bass players.
It certainly sounds better than this:
Now, if I can only find a way to eliminate drummers and bass players.
It certainly sounds better than this:
.“Art should comfort the disturbed and disturb the comfortable.”
― Banksy
- BananaSkins
- Posts: 477
- Joined: 29 Sep 2017
How much have big Companies spent on software with voice recognition security measures...
"What a time to be alive!" happily says the narrator toward the end of the video. In a few years nobody will be able to tell the difference, alive or dead, in the upcoming deep fake singularity.
757365206C6F67696320746F207365656B20616E73776572732075736520726561736F6E20746F2066696E6420776973646F6D20676574206F7574206F6620796F757220636F6D666F7274207A6F6E65206F7220796F757220696E737069726174696F6E2077696C6C206372797374616C6C697A6520666F7265766572
- MannequinRaces
- Posts: 1543
- Joined: 18 Jan 2015
That is bonkers! And to think it will only improve from here, crazy.
Awesome.
A point though...the second example the guy speaking is Scottish. The first part of the synthesized output sounds as if he’s from Somerset or Devon. English in any case. And there I stopped. Turning Scottish people into English people is an abomination of science and should not be allowed.
A point though...the second example the guy speaking is Scottish. The first part of the synthesized output sounds as if he’s from Somerset or Devon. English in any case. And there I stopped. Turning Scottish people into English people is an abomination of science and should not be allowed.
🗲 2ॐ ᛉ
Why? Most of you live here anywayMrFigg wrote: ↑14 Nov 2019Awesome.
A point though...the second example the guy speaking is Scottish. The first part of the synthesized output sounds as if he’s from Somerset or Devon. English in any case. And there I stopped. Turning Scottish people into English people is an abomination of science and should not be allowed.
Choose life. Choose Sweden.Zac wrote: ↑14 Nov 2019Why? Most of you live here anywayMrFigg wrote: ↑14 Nov 2019Awesome.
A point though...the second example the guy speaking is Scottish. The first part of the synthesized output sounds as if he’s from Somerset or Devon. English in any case. And there I stopped. Turning Scottish people into English people is an abomination of science and should not be allowed.
🗲 2ॐ ᛉ
These examples are pretty creepy sounding lol. Scroll down to:
https://google.github.io/tacotron/publi ... index.html
They use excerpts from Harry Potter, but it gets weird.
My favorite is the one beginning with, "Uncle Vernon entered the kitchen as Harry was turning over the bacon." And the one beginning with "Harry had the best morning he'd had in a long time."
(Lessac < 5sec, Location-Sensitive). Starts off pretty well and then turns into something out of a horror movie.
- 3. Generalization to Long Utterances
https://google.github.io/tacotron/publi ... index.html
They use excerpts from Harry Potter, but it gets weird.
My favorite is the one beginning with, "Uncle Vernon entered the kitchen as Harry was turning over the bacon." And the one beginning with "Harry had the best morning he'd had in a long time."
(Lessac < 5sec, Location-Sensitive). Starts off pretty well and then turns into something out of a horror movie.
Well, when we get a singing example of the AI maybe we can assure that, because Vocaloid can also sound like this:
...which in my opinion is better than the AI on the video, but if they keep the work it could really become very a interesting tool for vocals
Haha, those examples are really creepy and funny at the same time...joeyluck wrote: ↑14 Nov 2019These examples are pretty creepy sounding lol. Scroll down to:And click on the red ones.
- 3. Generalization to Long Utterances
https://google.github.io/tacotron/publi ... index.html
They use excerpts from Harry Potter, but it gets weird.
My favorite is the one beginning with, "Uncle Vernon entered the kitchen as Harry was turning over the bacon." And the one beginning with "Harry had the best morning he'd had in a long time."
(Lessac < 5sec, Location-Sensitive). Starts off pretty well and then turns into something out of a horror movie.
-
- Information
-
Who is online
Users browsing this forum: No registered users and 3 guests