Yes it's impressive.
All the rearranging tools used into the vid (to correct pitch and time interpretation -including modulation and vibrato) were availlable for almost 20 years with things like Melodyne or the most known Autotune ( which was not dedicated to poor hip hop at first...).
Where it is really impressive is in the capability to 'clone' human voice.
And it'll rise legal issues in our disinformation's driven world.
I heard recently an AI had been able to clone a voice in an undistingishable manner for human being.
It's scarying i agree.
All the rearranging tools used into the vid (to correct pitch and time interpretation -including modulation and vibrato) were availlable for almost 20 years with things like Melodyne or the most known Autotune ( which was not dedicated to poor hip hop at first...).
Where it is really impressive is in the capability to 'clone' human voice.
And it'll rise legal issues in our disinformation's driven world.
I heard recently an AI had been able to clone a voice in an undistingishable manner for human being.
It's scarying i agree.
So not only do you need 2 factor authorization, so that every company that you deal with online knows your phone number (or email), you'll soon get a phone-call from someone who sounds like your family/friend asking for help.
Apparently we can prepare by telling our friends and family our "safe word".
I recently lost my phone and had to go the trouble of replacing it. When I went back on instant messaging, some of my friends from the tech industry starting asking me private/personal questions to identify if I really was who I claim to be.
"Safe and responsible AI for the benefit of all" ??
a.k.a first we shall get everyone to use, then we monetize!
Bollocks to that.
Oh but we can regulate that!
More power to rulers/leaders.
well that is just wonderful.
This is one merry go round that humanity doesn't know how to get off.
Apparently we can prepare by telling our friends and family our "safe word".
I recently lost my phone and had to go the trouble of replacing it. When I went back on instant messaging, some of my friends from the tech industry starting asking me private/personal questions to identify if I really was who I claim to be.
"Safe and responsible AI for the benefit of all" ??
a.k.a first we shall get everyone to use, then we monetize!
Bollocks to that.
Oh but we can regulate that!
More power to rulers/leaders.
well that is just wonderful.
This is one merry go round that humanity doesn't know how to get off.
Last edited:
Oh yes wonderful.
I can see an unlimited number of cases for mass manipulation too.
Pair it with AI driven image modification and here we are... And that is what worry me the most atm.
At another level i can see 'service' coming to make the ones who left us resurect.
Some written material fed to Chat gpt inspired soft able to extract pov and writting/ expression style, an audio extract/record to clone voice and some pictures and you can deepfake anyone.
Enjoyable future is coming.
I can see an unlimited number of cases for mass manipulation too.
Pair it with AI driven image modification and here we are... And that is what worry me the most atm.
At another level i can see 'service' coming to make the ones who left us resurect.
Some written material fed to Chat gpt inspired soft able to extract pov and writting/ expression style, an audio extract/record to clone voice and some pictures and you can deepfake anyone.
Enjoyable future is coming.
Last edited:
I must agree.
The Ai will become so interwoven in our future society venturing into places we haven't dreamed exists. Making sense of it and going off this ride (tide) will certainly be impossible.
The Ai will become so interwoven in our future society venturing into places we haven't dreamed exists. Making sense of it and going off this ride (tide) will certainly be impossible.
That said AI in audio can be of tremendous help for some very dedicated tasks.
Recently used a plug in called Silencer from Black salt Audio and it solves the issue of bleeding during drum recording. Bye bye gates, hello tight isolated drum tracks without phasing issues anymore. It's mind blowing:
Recently used a plug in called Silencer from Black salt Audio and it solves the issue of bleeding during drum recording. Bye bye gates, hello tight isolated drum tracks without phasing issues anymore. It's mind blowing:
Check out the demo videos showing off what the latest version of chat gpt can do. Just imagine what it will be like in a couple more years..
https://openai.com/index/hello-gpt-4o/
https://openai.com/index/hello-gpt-4o/
I asked A.I to try an concentrate on rendering my loudspeakers.
He was much more interested in playing with the landscape, can't blame him.
He was much more interested in playing with the landscape, can't blame him.
How did you keep the speakers as a constant? Can you share your method, or is it very complicated?
It's a new plug in that comes with Google sketchup Pro. I don't know how come he hasn't fiddle with the speakers, usually he is way too creative.
Speaking of voice replication, Suno.ai is a music creating software that is eerily realistic. Probably if I wouldn't know it was computer generated music, I would think it was a real human. And now, with the paid version, thanks to some new "features" you can upload presonas and listen to albums of your favorite singers, even though they never created them in the first place. And download the songs. And share the songs. The future of music looks quite interesting.
I am an grumpy old man. I listen mostly (exclusively?) to music from last century. On Youtube. A bit 1960--ies. A lot of 1970-ies and 1980-ies. The videos are definitely part of experience. But you can imagine how bad those videos are. 240p? Bleeding colors?
If AI would be able to restore that sound and pictures to a decent quality I would be truly impressed. I said restore to decent quality, not adding, enhancing, fantasizing or hallucinating anything.
And it would be the first useful thing I'd see AI doing for me.
If AI would be able to restore that sound and pictures to a decent quality I would be truly impressed. I said restore to decent quality, not adding, enhancing, fantasizing or hallucinating anything.
And it would be the first useful thing I'd see AI doing for me.
You find a computer copycat interesting?
Ok.
I don't.
In fact if there is no human involved into the creating process i find it just plain boring and without interest.
And i'm an electronic musician using algorythmic generative tools ( through modular synthesizer). But if i'm not there to make choices relative to this generators ( or someone else human) it is just plain nonsense imho. It's the humans behind ( which makes choices) which are interesting and create what is known as culture.
AI will be able to do it for sure but where is the border to restoring things. I mean how would you restore issue linked to a technology ( like bleeding colors), which make it interesting?
I ask because i do audio restoring regularly, from tape or vinyl and bring in digital tools to 'restore' things. I'm often facing choice where you can make disappear some analog artefacts which were part of the initial release, and are part of the esthetic outcomes.
You can make this 'better' in a way but you automatically loose on something... defining the border is difficult and very probably preference related imho.
As such there is no real definitive answer...
I'm sorry, I do not support AI music generating at all and the comment on the future was in tended as sarcastic. But I do see your point in restoration, AI should solely be used for things like restoration. To use ai to generate things, especially to copycat originals and infringe on singers' own unique style, voice, emotion, and lyrics should never ever be acceptable. Ai never brings things to life. Ai never adds personality, or humanness to anything. It only adds together things that it copies from humans. Here's something I've been saying for a while now: no matter what it is that a human creates whether it's code, a poem, a piece of music, and even if he recreates something, that human always brings in a piece of human experience, a piece of himself, that emotion, that thing that makes humans humans. AI can never do that, it is not self-aware, it can only copy, and it never adds that piece of humanness. It never adds that thing that makes humans humans.
I totally get your point. It is risky to make it too good yes. I should have been more specific. Like "as it looked on a PAL color TV with 45 dB SNR" Which was minimum studio quality. Or "without sacrificing the 1970-ies atmosphere". If AI were so smart it could restore old videos and sound I hope it would be smart enough to understand me. I am not being cynical.
As for analog recording restoration, I would opt for removing distortion induced by record wear. Let's say up to the point where you were playing an new record on a decent turntable using a Stanton IEEE element. In no case further than that, not removing noise from the analog recording process, not improving channel separation etc.
I agree with you it is difficult to define the point where restoration becomes re-creation.
AI is making my dinner tonight.
Hoping for bangers 'n mash, but it's always nice to be surprised 😀
Hoping for bangers 'n mash, but it's always nice to be surprised 😀