Humans depend on speech a great deal: to authenticate, to communicate, to give orders. And, of course, the police state relies heavily on converting voice to data for analysis. Video below the fold:
This is nifty stuff: it’s doing analysis of a person’s speech and reversing out phonemes then allows you to rearrange them scarily well.
Now we can appreciate the speech synthesis that the squirrel people use for their Donald Trump effect. Joking aside, this quality of speech synthesis and editing is amazing; it has many potentially fascinating applications.
That’s scary. That means that “we have you on tape saying …” means nothing any more. Except that people will believe it; the same people that believe whatever the latest lie is without checking.
Rather ironic that the presenter’s name is pronounced “Say you”.
“Running for President was a terrible, terrible mistake. The worst. I am resigning, effective immediately.”
Susannah@#1:
With digital faces and compositing, I am now pretty sure that you can create something where nothing is real. That breaks the credibility of audio, video, photography. Perhaps that will liberate us from our natural assumption that an image that looks authentic, is. Perhaps it won’t. It’s going to probably affect older people (like me!) most.
quotetheunquote@#3:
“Running for President was a terrible, terrible mistake. The worst. I am resigning, effective immediately.”
I was thinking of that. Or how easy it would be to produce a new Billy Bush conversation video.
Scammers are going to have a field day with this.