AI can simulate anyone’s voice with 3 seconds of audio

Get help or discuss anything relating to audio/video software & hardware
harrylime
Posts: 254
Joined: Sat Feb 19, 2022 4:56 pm
Has thanked: 59 times
Been thanked: 58 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by harrylime »

On a more positive note, dj's can employ AI demixing to more easily sample and remix into new creations. I can't wait to hear more live mashups of eclectic elements instead of just existing mixes chained together.
User avatar
Mr Bump
Posts: 463
Joined: Sun Feb 06, 2022 1:01 pm
Has thanked: 8 times
Been thanked: 73 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Mr Bump »

What I want to know is, when do I get to fly to work in my hover-car, as promised? Come on tech guys. Get working on that one.
Ken_peps
Posts: 560
Joined: Thu Mar 11, 2021 4:52 pm
Been thanked: 25 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Ken_peps »

Just to add one more thing to Lord Reith's point, The Beatles used the studio as a creative tool rather than a replacement for lack of talent, as is sadly the case now with so many pop, jazz, soul, rock, etc. performers....I would also include Stevie Wonder as a creative genius who very early on learned how to use all kinds of technological advances to augment his inherent talent (just like the four lads from Liverpool), and produced some amazing works. So it's not that technological advances are bad, it's just that it brings out the laziness (and cheapness) in people - why hire a full orchestra for a Broadway production, for example, when you can save money by just hiring enough live musicians to make it look good and fill in the rest with pre-recorded tracks or synths instead of real strings? And I'll wrap my little tirade up by again complimenting the work of people like Lord Reith who have used said technology to actually enhance our enjoyment of the still-astounding music of The Beatles, which, when all is stripped away, still leaves us with four of the most creative musicians of all time.
User avatar
Lord Reith
Posts: 4602
Joined: Thu Feb 18, 2021 8:22 am
Location: BBC House
Has thanked: 139 times
Been thanked: 3946 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Lord Reith »

Yeah, The Beatles and their coworkers definitely used technology in a good way. I guess that was because when they first appeared, technology was still extremely primitive (see my "BBC mixing desk thread" in Miscellaneous) and so they were used to putting in 100%. So when the new toys came on the scene they used them in innovative ways rather than just as a means of doing less work (with the possible exception of ADT... but it was only really John who used it that way, and then not for long).

These days, the curse of Autotune means nobody really needs to learn how to sing in tune, because even people who can sing are forced by their producers to use it. It has totally flattened out all music and made everybody sound like a robot. Within a generation, nobody will remember how to sing properly anymore and a singer going onstage without Autotune will be unthinkable. But it's all happened before because when distortion first became an option on guitar amps, it quickly got abused and soon any clot who knew three chords could sound like a rock god because distortion covers up bad playing so well.

I despair that we'll end up with Autotuned singers (or worse, AI artist-bots) backed up with AI created instrumental backing, playing songs written by computers and mastered by AI. Everything is pushing towards that, as if creating art is akin to housework and must be simplified. One ad I keep hearing on Spotify drives me nuts, an ad for an app that allows anyone to create and record their own "song" with AI. It's a piece of garbage but this is the mentality - that creativity should somehow be democratized by allowing anyone to create songs with AI, even people with no talent or musical skills. And all we'll end up with is a bunch of machine written crap.

Here endeth the lesson!
Women there don't treat you mean, in Abilene
Ken_peps
Posts: 560
Joined: Thu Mar 11, 2021 4:52 pm
Been thanked: 25 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Ken_peps »

Lord Reith wrote: Sun Jan 15, 2023 9:12 pm Yeah, The Beatles and their coworkers definitely used technology in a good way. I guess that was because when they first appeared, technology was still extremely primitive (see my "BBC mixing desk thread" in Miscellaneous) and so they were used to putting in 100%. So when the new toys came on the scene they used them in innovative ways rather than just as a means of doing less work (with the possible exception of ADT... but it was only really John who used it that way, and then not for long).

These days, the curse of Autotune means nobody really needs to learn how to sing in tune, because even people who can sing are forced by their producers to use it. It has totally flattened out all music and made everybody sound like a robot. Within a generation, nobody will remember how to sing properly anymore and a singer going onstage without Autotune will be unthinkable. But it's all happened before because when distortion first became an option on guitar amps, it quickly got abused and soon any clot who knew three chords could sound like a rock god because distortion covers up bad playing so well.

I despair that we'll end up with Autotuned singers (or worse, AI artist-bots) backed up with AI created instrumental backing, playing songs written by computers and mastered by AI. Everything is pushing towards that, as if creating art is akin to housework and must be simplified. One ad I keep hearing on Spotify drives me nuts, an ad for an app that allows anyone to create and record their own "song" with AI. It's a piece of garbage but this is the mentality - that creativity should somehow be democratized by allowing anyone to create songs with AI, even people with no talent or musical skills. And all we'll end up with is a bunch of machine written crap.

Here endeth the lesson!
Couldn't agree more - well said!
User avatar
Mr Bump
Posts: 463
Joined: Sun Feb 06, 2022 1:01 pm
Has thanked: 8 times
Been thanked: 73 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Mr Bump »

Personally, I don't care. For me, the automated drum track of the late-1970s killed creativity in music. There's not much past 1980 I consider worthy, and whatever happens from here, there will always be something or other worth listening to around the edges, but mainstream is plasticrap product for kids and had been for decades. Isn't that why, after all, we are all here concentraiting our interest on a band who split up more than half a century ago?
User avatar
Lord Reith
Posts: 4602
Joined: Thu Feb 18, 2021 8:22 am
Location: BBC House
Has thanked: 139 times
Been thanked: 3946 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Lord Reith »

A lot of what I hear in the top 40 these days is so technology dependent and bereft of musical inventiveness you could probably easily cut out the human element altogether. The backing tracks are just samples and a drum beat and the voice is roboticized, so I think perhaps virtual artists will become a big thing in the future. I can even imagine kids going to concerts to watch virtual pop stars on a giant screen, kind of like in Japan where screaming girls go along to big venues to watch gamers play on the big screen.
Women there don't treat you mean, in Abilene
User avatar
Ziggy C
Posts: 551
Joined: Thu Oct 14, 2021 12:10 am
Location: Woodland Hills, CA
Has thanked: 96 times
Been thanked: 125 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Ziggy C »

Jim Morrison's comments on music back in 1969 are prophetic. In this clip he discusses where American music came from and where it is going.

Starting at about 1:15 is most interesting... one person with a lot of machines, tapes?

zaval80
Posts: 390
Joined: Sun Mar 07, 2021 9:19 pm
Has thanked: 23 times
Been thanked: 18 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by zaval80 »

Mr Bump wrote: Sun Jan 15, 2023 11:45 am What I want to know is, when do I get to fly to work in my hover-car, as promised? Come on tech guys. Get working on that one.
The world government would rather keep you out of this sphere of development.
User avatar
Lord Reith
Posts: 4602
Joined: Thu Feb 18, 2021 8:22 am
Location: BBC House
Has thanked: 139 times
Been thanked: 3946 times

Re: AI can simulate anyone’s voice with 3 seconds of audio

Post by Lord Reith »

zaval80 wrote: Tue Jan 17, 2023 5:10 am
Mr Bump wrote: Sun Jan 15, 2023 11:45 am What I want to know is, when do I get to fly to work in my hover-car, as promised? Come on tech guys. Get working on that one.
The world government would rather keep you out of this sphere of development.
It's funny how all those science shows in the 1970s predicted hover cars but utterly failed to anticipate the computer revolution. I remember in junior school we had to draw a picture of what we thought a living room would be like in the future. I had a printout coming out of the wall, a special cubicle where you could watch tv and listen to music at the touch of a button and some other things i've forgotten, but there was probably something to do with levitation in there.
Women there don't treat you mean, in Abilene
Post Reply