We can now believe none of what we see, and none of what we hear.

It's a scary world out there, but we have in the past been able to rely on video and audio data to use as evidence when trying to prove/disprove something. With the power of computing, it seems we may be reaching a point where it will be impossible for the average layperson to differentiate between what is 'real', and what is fabricated. I'm sure some people will remember Face2Face, software for 'a novel approach for real-time facial reenactment of a monocular target video sequence'. Essentially it allows somebody to convincingly overlay their own facial expressions over that of someone else in a video.

It's better demonstrated than talked about, as one can see from this video.



To further allow us to move into the realm of making the unbelievable believable, we have this new software from Adobe, which they describe as 'Photoshop for audio'. As Vice note, it seems one can make anybody say anything now.

Ever wanted to make Natalie Portman yell obscenities at your neighbors? What if Garey Busey could leave your mother a sexy voicemail on her birthday? Wanted to prank your little brother by forcing him to call his crush and profess his love? Adobe has you covered.

When Adobe released photoshop in 1990, it dreamed of a world where movie studios and photo editors could do in minutes what once took hours. It never dreamed the world would take the digital editor and use it to put celebrity heads on porn star bodies, distort women’s bodies in magazine cover, and create vile memes.

Now, the same company that gave the world Photoshop wants to do for the human voice what it did for the human image—give people the tools to warp it in anyway they see fit.

Of course the proof is seeing it in action, so here is a demonstration.



Can we trust anything we see or hear anymore?

If this is available for the public, just imagine what tools governments possibly already have at their disposal.

a reply to: cuckooold

Nifty, eh?

I'm actually coding in this stuff right now. Using something called OpenCV for a mobile app i'm developing.

Computer vision is really impressive stuff. Google just released theirs called Google Vision.

The technology in this video leverages the same tech Pixar uses. Available as open source called OpenSubDiv. To be clear, this has nothing to do with Adobe. They were just making an analogy. *Edit: sorry I had not seen your second video yet. Oops!

Anyways, yeah the potential for 'faking it' is a real concern. Especially as the technology improves. Reminds me of Black Mirror!!

If someone has developed a bomb that would permanently wipe out all electronic communications.....please use it now.

I've always thought, "never believe anything". Read about it, think about it, talk about it, but don't put belief behind it. Belief is a dead end for critical thinking.

Here is more info concerning the second video. It's a project from Adobe being called Project VoCo.

Project VoCo lets you edit speech as easily as text — and you can’t just edit existing text, you can even use the same voice model to create completely new recordings, too.

Here is how this works: Project VoCo needs about 20 minutes of voice samples from a given speaker. It then analyzes the speech, breaks it down into phonemes, transcribes it and creates the voice model. If you listen closely, you can hear when a word has changed, but it’s probably only a matter of time before you won’t be able to distinguish the actual recording and the edited (or completely fake) one.

Read the rest of the article here

Spooky stuff.

No potential for abuse at all.

a reply to: loam

The tech behind the first video has lots of great applications despite its potential for abuse.

I can't say the same for the second one (Project VoCo). I see lots of potential for abuse and not much practical application outside of that. Perhaps i'm just not being imaginative.

This could be fun at home to make funny videos, but it is definitely scary in the hands of the unscrupulous.

There's an old analogue device made before 1981 the FBI has used in sting operations made by Lanier office products, but you won't be able to find it publicly, that is hooked up to a telephone, and you call someone and get a sample of their voice that you need, and then you can call another target and speak in real time, and the new person called will hear you in the voice that you previously sampled even though you are the one speaking.. We never hear about this being used ever, but the machine exists.

a reply to: cuckooold

This has been reposted at least 5 times this year.

And the quote you're butchering came from Mark Twain:

Believe half of what you see and nothing that you hear.

a reply to: NoCorruptionAllowed

This could be fun at home to make funny videos

Similar technology is already being used at home to make funny videos. All those video clips you see on social media of people with the stupid dog face superimposed on their actual face lol Or the barfing rainbows. Face tracking with computer vision. Not as sophisticated as OP video, but it's essentially the same deal.

a reply to: ColdWisdom

I'm aware the Face2Face video has been posted on ATS, but I have not seen the Adobe software discussed.

I'm also fully aware of the Mark Twain quote, which I was not citing, ergo not butchering, but thanks for your concern.

a reply to: cuckooold
And one would be extremely naive thinking that this tech has not been available to those 'in the loop' for at least the last 20+ years.
I can actually remember mucking around with a program or two (not the programs names though) that basically did this kind of thing when I was experimenting with 3d fractal landscapes via Bryce et al back in the late 90's. But the rendering was unbearably slow given the processing speeds back then.
So remembering that.... available maybe 30+ years ago (to alphabet agencies etc)

The news media basically does stuff like this already I know Fox news will report false facts when they have "experts" come on an speak, for example. But CNN caught me off guard with a Trump video they edited a while back. They took out entire sentences to make him look more extreme than he is.

Even though I don't support Trump, it is hard to know how accurate and unbiased the news media is. Reading the news when it comes to Trump doesn't tell me what is really going on, I feel like.

originally posted by: cuckooold
We can now believe none of what we see, and none of what we hear.

(That's what everyone says right after the elections. It seems to only last until the next election, and spate of lies, though.)
Congratulations on reaching maturity! *__-
You are finally getting a clue!
Also, you would do well to not believe anything that you 'think' (or 'feel'; feelings are thoughts)!
You cannot 'believe' none of what you see, you can only NOT 'believe' anything!
That is the basis of 'healthy'/'sane'.
Being infected with any beliefs (is) leads to madness!

It's getting really scary now, as we have begun to start to tap into GPU more and more, and nvidia keeps cranking out cards with more GPU power. Renders that used to take me days, now take hours, and I only have a GTX 960, not one of those badass titans. And just think of what moneypits like the NSA have in their hands, you know it would blow anything us peons can do out of the water.

a reply to: TKDRL

Renders that used to take me days, now take hours

Rendering with Iray, perchance?

I've been using Keyshot, but now that I have a half way decent Nvidia card (GTX 960 as well) i'm starting to dabble with Iray.

a reply to: Lucid Lunacy
Yeah, mostly the iray that's built into daz studio these days. I would love so much if houdini would make an iray plugin

Yeah, humans have officially crossed the line. Aliens will be here shortly.

a reply to: ColdWisdom

He wasn't butchering the quote, he modified the quote to fit his post title. With these two technology it is safe to say 'believe none of what you see, and none of what you hear'/

It is a play off the original quote; Quote Nazi.

a reply to: cuckooold

I'm some what dubious about that Adobe Vocal software. He makes it seem as though one short recording of their voice is all you need, but at the end he is asked about that and says it actually needs like 20 minutes of voice recording. I'm thinking what it actually does it search through the audio for sounds that it needs to form words, it doesn't actually generate the audio from scratch the same way something like WaveNet does. It's still a very interesting approach to text-to-speech though because it would allow for custom voices to be easily created.