After 20 Minutes of Listening, New Adobe Tool Can Make You Say Anything


That’s fucking scary actually. Can you imagine the kind of stuff you can do with someone like trump as president who is fucking retarded and says whatever you want. Produce a somewhat convincing video of him saying something awful (which really isn’t very hard to not believe) and then bam, who knows what chaos you could cause


Post-truth indeed.


watch governments use this to make fake videos of other countries starting conflict so said governments can justify endless wars


Phillip K. Dick overdrive.


This is ridiculously awesome.

It’s a shame film samples aren’t as popular in electronic music as they were a decade ago. People would be creaming themselves over this!

Now the next level is to run synths and other non-human noises through it to see what it comes up with in an attempt to replicate them. mmmmm



Brave New World / 1984 mashup


PKD meltdown


Just had a really great conversation with one of the main guys leading this project. based on that conversation, I don’t know how interesting it would be to feed it non speech input. The algorithm doesn’t try to resynthesize the voice, but rather uses fragments of the input speech with some processing to build up the replacement text. I could see it being potentially very powerful for editing percussive material. Most humans aren’t great at keeping their velocity consistent throughout a performance, which is great in most cases, but in a dense mix where everything is slotted into specific time and frequency ranges, being able to get a more consistent attack transient could be a godsend. Seems miles better than trigger based drum replacement solutions



on the bright side:

you can make celebrity voices say funny things to use as vocal samples in tracks


Nope, no problem here lol

:iphone: :cold_sweat: :gun: :money_with_wings:


“Google Assistant, please drain my bank account and send it to this unknown person out of the country.”