from music rebalance to potential demixers, what's our future ?

[kidslow]

In the case of self-driving cars the promise is probably vastly exaggerated because the systems have to be perfect, one mistake and people die. But in domains where errors don't incur such risk progress is moving at an astounding rate.

Oh it is not probably vastly exaggerated. It is grossly vastly exaggerated. Starting with the huckster of autonomous driving in chief, Elon Musk. Does he believe his own BS? There are going to be a lot more accidents as the autonomous driving proliferates. I just hope that I am not a victim. Mostly it is a gimmick now and L4 autonomy is way way in the future at any scale.

[Bombadil]

Nope. All I know is it is Weta's, and it hasn't been shared with anyone but the Beatles. He used it in 'Get Back' and now Giles Martin has access to the tech to demix the Beatles' stuff that could do with a remix. Like Revolver and Rubber Soul. YAY!

[pixel85]

The real budget is in "de-mixers" and restoration tools. I can only imagine how big are contracts from special agencies to make those tools. Governments love to give giant checks.

What current commercially available "de-noisers" can do is already amazing, like Waves Clarity which can often do the same job as RX's Denoiser on the highest settings, that struggle to work real-time even on newest CPUs. I can only imagine how good are AI tools for those tasks that ate not yet available commercially.

For the music generation, whole royalty free music for adverts should be soon generated by AI. There's already billions of same sounding ukulele tracks. It will be faster to use AI to generate something "new" than using royalty free music services.
I can't wait to become an operator of such tool

Mainstream/Pop music has huge chances to be replaced by AI. Singers of course will stay, even if their voices will be generated. Whole thing about mainstream music is to observe "awesome" life of celebrities. Music is just an addition to the spectacle and less and less people care about music itself.

A lot of medium level EDM "artists" / DJs may begin to use AI tools instead of ghost producers to boost their career. It's all about sounding the same like other 400 tracks on Spotify playlist anyway. No art, no vision, just copy/paste stuff that AI music generators should be soon great at. Another part of the music industry where Instagram is more important than music.

Heh, even here, a place mainly for music production hobbyists, AI music generators seems to have chance for success. I saw many times comments where people seems to consider music production as an annoying chore

One place where AI can't take over is artistic direction. For this stuff we would need real, full AI, and such thing, especially used for such trivial tasks like music, none of us will ever live long enough to see being available commercially.

[Krakatau]

just a very basic question on the subject :

- How much a current plugin like Ozone Master rebalance is using AI in it process ?

[WasteLand]

i have 2 isolators; DeepAudio and SpectraLayers Pro 9, to call them isolators. i have never bought them for that purpose, i have bought them for the spectral capabilties, together, they can't work at the same moment... but still... you can achieve a lot..

i do isolate, but my own samples/recordings, which lead mostly to strange results, sometimes usefull, and that i use.

nice thing of DeepAudio for example; you draw a timbre of an instrument, that you have loaded.. of course..
or use the timbre of an instrument on another instrument.

DeepAudio gives seperate harmonics and inharmonics, by the way.... the harmonics, can be edited as notes.
the inharmonics, you can't do a lot, but perhaps that will change.

still...

the thread is moving more to AI, well it is AI, what a DeepAudio for instance does, but it must be learned, and that is the catch; AI isn't that good, yet.

but still; nice stuff, for crazy sounddesign... so i can use the processes, i repeat, on own recordings, that don't have that much recognizable 'stuff', but still...

the AI parts are a tool for me, and not the main tool, but also not tool that i don't use for experiments....

[shonky]

The more generic and formulaic the music, the higher ease of it being learned and produced by AI.

E.g “so we’ve got this fantasy/historic battle scene, conjure me up some low portamento bass drones, low staccato string ostinatos and drum ensembles”

[ReverendLove]

We're all going to hell!

[kidslow]

You.Go.Rev!

[ReverendLove]

You.Go.Rev!

Repent, sinner!

[vurt]

You.Go.Rev!

Repent, sinner!

bit rich coming from frankie boyle there

[pixel85]

just a very basic question on the subject :

- How much a current plugin like Ozone Master rebalance is using AI in it process ?

It doesn't do any "AI" processing. They "teached" algorithms to recognise certain patterns - that's the part of "AI" called Machine Learning. The plugin itself is only matching input to pre-learned output.
One can say that it's still AI processing but it's bit of a stretch. The olugin itself is not learning anything new, is not recognising new patterns to improve itself.

[Ploki]

I predict more LA2A clones