“What a time to be alive” - eek.
Appen do a lot of this stuff. I have done voice recording for them and also appraisal of voice recordings - for naturalness, accuracy, attractive sounding etc.
2 Likes
They can mimic the voice, but still have to put in the work to figure out the secret security phrase or code words used by the boss, right?
2 Likes
One of the first episodes of Star Trek: The Next Generation had a young Wesley Crusher showing off a speech synthesizer he made that allowed him to impersonate Captain Picard. I’m sure both the technology and the form factor of the bulky audio player Wes was using would seem laughably primitive by today’s standards.
2 Likes
Apparently scammers have already used less sophisticated software to run cons; they still have to put some work in, but the voice emulation software lowers the bar for just how much effort is necessary to pull it off. If people hear a familiar voice, that itself is seen as verification.
1 Like
So glad biometrics are definitely a thing now. So unhackable…
1 Like
I keep thinking about the scenario from the movie Sneakers. I have passphrases for some accounts, and hope that organizations never just go with the sound of a voice without considering what it says.
1 Like
Voice synthesis becomes part of a social engineering attack - and the whole point there is to avoid having to give passcodes.
While it’s certainly right to consider the hazards of this tech, I’m excited by the positive potential. This tech now makes silent phone conversations possible, where sub-vocal speech detection is combined with a simulation of someone’s natural voice and thus allowing you to silently converse with people over a smartphone. No more sharing your conversations with the world or looking like a lunatic talking to yourself when wearing a headset.
It also means a naturalistic voice feedback for sub-vocal speech-driven computing. So one can silently write text messages, email, or more extensive texts completely hand and screen free. No more hunching over screens like some ape or walking into people and objects on the street. This opens the possibility for a viable general audio computing operating system for the blind --an all too-long overlooked need-- with mobile hardware platforms far smaller than other mobile devices. Audio computing remains a much-neglected area of design, the developers of the voice assistants still stupidly seeing the things as merely putting their own little branded merchant in people’s home…
It also brings my dream of a digital voice assistant using the many voices of Paul Frees for different applications that much closer to reality.
1 Like
This is why I try to sound like Brother Theodore whenever I’m being recorded.
Has anyone tried the site: https://app.resemble.ai
You need 50 voice recordings for it to try to mimic yours, so i’m not sure where the ‘5 seconds of recordings’ comes in.
It seems to be bugged out for me after 40 recordings, so i’m assuming that it doesn’t work, it’s all faked and it’s a way for the CIA/FBI/MI6/Facebook to gather biometric data from suckers like me.
1 Like
I recall Roger Ebert (if you don’t know he had cancer in his jaw and was unable to speak at the end of his life but his mind was still very sharp) had people working on voice synthesis of his own voice so he could type and his own voice would come out instead of a fake robot voice. This technology would be great for people in that situation. you wouldn’t need individualized programming to do it. just a small voice sample.
unfortunately lately no probably wasn’t a fake. he’s been warning that the Dems are going to far left.
when audio text to voice technology improved people would offer to give Stephen Hawking a better sounding voice. He always refused, pointing out that his synth was his recognizable voice for himself and others.
someone already has done it and been caught by police. they used a clunky method so it was easier to leave it like a message in their inbox to curtail interactions.
the clunky method has been used on TV since the 60’s so I’m glad science is finally catching up lol
1 Like
I understand that effort to give Ebert their voice back was a pretty colossal undertaking and was only possible because of the vast amount of recordings of Ebert’s voice. Apparently a massive project from just a few years ago is now five seconds of machine work.
2 Likes
This topic was automatically closed after 5 days. New replies are no longer allowed.
