i want AI cowbells! in a next new Rytm firmware
AI only works when people can describe what they want. Good thing is that the musicians have no idea what they want…And this is the beauty of music - jamming, happy accidents etc. The music that ticks boxes via AI will quickly get skipped.
You see it all the time with mediocre art. I call it spreadsheet art. Some artist thinks a Banksy style stencil depicting whatever the theme of the moment is (imagine a knock off/wannabe Bansky inspired stencil of a poor girl drawing a Chanel bag on the wall). But its feels empty and doesnt convey anything other than basic tick box themes.
You are essentially cheating yourself of the process - the creative and fun bit. Music/art etc is all about making sense of the world for yourself. Getting some computer to knock it up for you is pointless unless your just chasing instant gratification.
Its never been easier to make music and for that it already feels diluted to shit without AI
3 Likes
First of all, we need a neural network to listen to trash created using neural networks. Even if you manage to create something relatively valuable, it will not make sense. Just as an artificial nightingale has no meaning compared to a real bird. Sampling itself is a surrogate in relation to music-making; are we really too lazy to even press a button and turn a knob?
Making music is about manipulating sonic material into something cohesive that expresses a point of view and has a life of its own. Where that material comes from or how it’s made is irrelevant.
Yes, I believe the key here is in integration between controller / computer / server.
Why not! Also since we already have Live and Push…
Ableton could be such an innovative company instead of just blindly making money on adding archaic features that had to be available since decades.
I really like how Izotope use AI for mastering / mixing tools. It really makes things so much better!
Elektron could learn a bit and maybe implement some of AI functionality in their future flagship sampler, just needs wifi connection
I want a smart tv so I can tag actors , shows , directors , series names that are automatically avoided for me. I’ll just set a show / channel priority if any of the avoid tags are met
No more reality , celeb scripted rubbish , no Boris blabbering his face off , no jack black , … Kevin hart ( this list is long ) it’ll be bliss.
Doubt it needs Ai, just an ability to read meta data .
Tajnost didn’t really imply that it would be easy to implement, that thread gets thrown around a little too much IMO. This is all a bit of fun idle speculation.
If Elektron or another similarly capable manufacturer implemented some kind of thing that allows you to input a prompt and get a unique generated sound from the cloud, I think it might end up being a commercial success and contribution to the state of the art. It could also be a flop, cheesy and end up overused, but that’s just how it goes sometimes. Elektron doesn’t strike me as the sort to do it tbh, Akai though definitely might.
Sorry, maybe I was a bit triggered by “just” or maybe I was having fun 
1 Like
Yes also one
We just need to eat more ass
Sensitivity is specific to human development.
It goes hand in hand with reaction time, and the decision that results from it.
Today AI works well for the general public
Is ridiculous.
Show me otherwise??! (!)
It’s a dad learning to draw
Or a mom pointing a direction or correcting a math problem
We must not lie to ourselves and hope too much from AI.
The basic concept will, and should, serve to accumulate and process data simultaneously.
Ease a flow
Archive feed data
((((Making a drawing or music is not glorious in my opinion))))
On a musical gear level,
I will see the AI change the background of the samples of my patterns to replace them with something else.
And store the finished patterns on hard drive.
When it’s carnival, AI feeds them back into the DT so I can play my curls at the park. Nope ?
I t hink we have the technology to do it
teenage engineering put out this prompt based sample generator the other day. the samples it makes kind of suck but hey it is what it is
1 Like
You’re right. That’s probably how it would work in practice.
yeah a way too lo-fi
but looking at this and thinking that probably teen eng will be the one who implement ai sound gen in their hardware tools
I have no problems with AI in music, I use it myself for various things. My point is what purpose would an AI sampler do? What’s it exactly sampling? If it’s just creating a “sample” what distinguishes it from AI just creating a unique piece of music?
1 Like
The purpose would be to give you a way to create sounds that don’t currently exist and are difficult or very time consuming to make using traditional methods. It’s just another form of synthesis, it’s up to the user to make musical use of it.
1 Like
But how is that a sample vs an AU generated clip of music? By sampling I’m assuming we mean taking a part of an existing track and repurposing it…? Otherwise it’s just AI creating a clip of music.
1 Like
Sampling is taking a small part of sound away from its original context. It can be anything, like field recordings, a single instrument, experimental recordings using contact mics etc. I’m not sure where this idea that sampling is just taking something from a piece of music comes from.
1 Like
I also want an Allen Iverson sampler!