Artificial Intelligence Noise

[Balor/SS1535]

using lots of gen AI in many aspects of my life (work, music, artwork). Not to do noise yet, except to treat some specific sounds and samples.
we all have been using AI simply if we owned Photoshop, etc. etc.
I would never let a machine decide the final work for me, yet I can ask its help if I need a specific element.

What do you mean for treating samples?  Do you just say "add reverb to X" or something?

[tiny_tove]

using lots of gen AI in many aspects of my life (work, music, artwork). Not to do noise yet, except to treat some specific sounds and samples.
we all have been using AI simply if we owned Photoshop, etc. etc.
I would never let a machine decide the final work for me, yet I can ask its help if I need a specific element.

What do you mean for treating samples?  Do you just say "add reverb to X" or something?

more like "create repetitive pattern, with metallic sounding percussion" , it then creates something similar (it tries to create something musical), I cut the part I need, reverse it and add my own stuff.

[Balor/SS1535]

using lots of gen AI in many aspects of my life (work, music, artwork). Not to do noise yet, except to treat some specific sounds and samples.
we all have been using AI simply if we owned Photoshop, etc. etc.
I would never let a machine decide the final work for me, yet I can ask its help if I need a specific element.

What do you mean for treating samples?  Do you just say "add reverb to X" or something?

more like "create repetitive pattern, with metallic sounding percussion" , it then creates something similar (it tries to create something musical), I cut the part I need, reverse it and add my own stuff.

That's pretty interesting.  I assumed that you have experimented with varying degrees of specificity in your prompt?  Does it seem to work better when you give it a lot of room to interpret or when you really dial in your goal?

[Penon]

WHile relative to other genres, I am not much worried about impact on noise and all sorts of "experimental" music which I guess would be the last to fall (after all it much easier to produce formula-based catchy pop using AI), but I occasionaly think about AI / next generation technology impact on music as an art more broadly.

It is not rocket science (even if a bit expensive still) to measure what sounds (pitch, frequency, notes, chords, patterns) your brain reacts most positively to and produce bespoke tracks based on this. You can even have a menu for "sad", "angry", "uplifting" and "dreamy" based on your brain (or perhaps body/hormones) response to certain sounds.

Would "manual" music be nothing more than a niche hipster fetish for those who think old school is cool?

[tiny_tove]

That's pretty interesting.  I assumed that you have experimented with varying degrees of specificity in your prompt?  Does it seem to work better when you give it a lot of room to interpret or when you really dial in your goal?

as said, I work with ai for very small elements, mostly rythm oriented since I am crap at that. In this case, less is more is the solution. I am using an app called texttosample, when you give too much information is looses it and makes what it feels better. It's a musical application so I never write "harsh noise with feedback" since I have plenty of synths and devices to do so :) I am sure there are many AIs to do noise (probably in Max). Consider that "generative music" is nothing new and it has been around for over 30 years, since algorythms have been in use in the arts, they were not based on the same data models of today.

in my opinion, same as for images, whatever helps us to deliver the good is great, be it physical, analog, digital, generative, etc. sometimes all together. is a machine does great noise I wouldn't be against it, but I prefer to do my own stuff and keep buying gear. (no victim of gentrification, simply I have no other vices than music, so I spend most my money in that and video gear)

[Penon]

I occasionaly think about AI / next generation technology impact on music as an art more broadly.

It is not rocket science (even if a bit expensive still) to measure what sounds (pitch, frequency, notes, chords, patterns) your brain reacts most positively to and produce bespoke tracks based on this. You can even have a menu for "sad", "angry", "uplifting" and "dreamy" based on your brain (or perhaps body/hormones) response to certain sounds.

Would "manual" music be nothing more than a niche hipster fetish for those who think old school is cool?

When I posted this in 2023, I at least hoped it would be a distant future but seems like the AI/music crossover is moving in this direction faster than I thought. Meet www.brain.fm (they actually launched some years ago but with recent AI advancements became much more popular recently, it seems).

Historically art has been created in an "artist-led" fashion. Artist creates an album (or writes a book, or whatever), and then this piece of art in same exact form (with some exceptions like Brian Eno's Reflection) is shared with many listeners. What I predicted two years ago was the rise of listener-led music - that is, using AI to generate music that perfectly responds to your private personal preferences, moods and emotions. It is much more apocalyptic than simply AI-generated music on Spotify because latter at least recognises the primacy of "creator" and seeks to replace it. In that case, different people still hear the same music and that creates certain sense of unity amongst listeners, which is one of the main features of any art, if you ask me.

Here, Brain.fm as I understand it takes step forward and focuses on listener - music is designed to match your particular mood, or productivity goal. It is not quite entertainment yet, and not extremely personalised yet but that is just the start. Now, maybe it is a scam (like i-Doser 20 years ago) but this time, with all the AI progress, this feels real and takes us further and further away from artist-led music to the world where each person will have their private individually tailored AI-music bubble.

[Minus1]

Interesting thread.

Call me old fashioned, but if I ever make Noise, the credits will read something like:

"No synthesizers, computers or AI were used in this recording."

I will not buy or listen to AI.

But do carry on.