The visuals for this psytrance track are super impressive, and im a total newb when it comes to AI generated stuff, could someone who knows their stuff please explain the processes/steps/tools used to create this?

I presume they started with a tool like Midjourney, exported a bunch of similar scenes and used some other process to morph them into each other? I dont know, but subjects appear and morph into other things like a face turning into a clock & a mushroom appearing beside it etc., which doesnt seem possible with single-prompt services as a new scene is created after each prompt
Also a video editing soft was probably used to zoom in & out of the scenes and generate a fluid feeling? not so sure because the resolution stays high when panned in

 

those ai video su_k.. yiak.. terrible

 

Really cool video! I also wonder how they make it, it's pretty interesting!

 

Well if you look at it from an artistic talent perspective, both the music and the video are cutting edge, pushing the envelope! (even if the song is a bit formulaic, I love psytrance but this one is a bit cookie-cutter compositionally, while the sound design is top notch)

If you have never taken psychedelics and think its just weird, may I suggest a local finger painting and balloon animal guy?

 

I'm pretty sure its all stop motion animation with clay on an armature. like the singing raisins. Raisins cannot sing, I tried to teach them for over 3 months. all i got was farting noises. I did manage to teach a banana to bark, but its shelf life was rather short so I put it in the freezer, but when I got it out it just made shivering noises and crawled into the blender, so I made a smoothie.

Animation and video is hard and vegetable and fruit are soft.

 

its called stable diffusion. A LOT of different artists have been using it for years.

unless you want to fatally date your creative work like a bad 80s hair style, you should completely avoid this.

but also for the sake of arts and culture in general...

 

Last edited: Sep 2, 2024

"deep dreaming", what they called early, trippy AI.

You can feed an image to this place and it'll spit out deep dreaming style videos

https://deepdreamgenerator.com/



Drop 80% of frames or space them out much further and apply a painting filter, merge back to video.

 

Last edited: Sep 2, 2024

this kind of stuff is around for some 7 or 8 years now and in between the quality is pretty amazing.

another variant of it are these fractal style zooming videos.

both is mostly done by mutations of multiple seed values using apps like touch designer.

 

This is the big issue with AI. People now think it just involves the correct software and completely overlook talent and artistry. Only recently has digital photo manipulation been accepted as another art form where the software is merely a tool used to express oneself.

"I could do that if I had the same software" nope. If you haven't yet done it you never will if you believe that. Learn some software and become good at it, then become fluent at it, until its almost invisible to you and all that you will need is your artistry. And try to recognise the humans skills outside of the tool.
We have enough technology in our phones to create anything that would outshine any film of the 90's. Yet most people don't have the talent.

 

ill be honest, a year ago i thought 'ooh new art frontier', a year later after using the best tools and following it as a medium, i'm pretty underwhelmed by everything in that scene. Its technically stagnant and populated by shallow low talent business types just playing around.

I like ideas ive created with it, but its still random playing, there's none of my real art skills on display, only my taste. Its more like being a DJ than a musician.

Good taste will steer someone towards uncovering cool ideas, and AI art in future will be cool sometimes. Its not a revolution, just a tool to create assets a real artist can play with.

Its not the same as other digital art tools like Photoshop, they need a lot of talent to actually create the art. I could teach anyone here in 15 mins everything you need to know about AI art creation. You would be technically as good as the best AI artist in the world and your output would look very polished.

 

Last edited: Feb 10, 2026 at 8:42 AM

It becomes human when you inpaint over and over again at the macro and micro level. Were you using stable diffusion ?

 

This one is pretty dope

 

Used a lot of Pony, Flux, Illustrious, Midjourney, and Kling models. In painting is just clumsily inserting more random you like, like a DJ swapping in a new kick drum loop you didn't make. Its barely an artistic skill and so basic a process.

 

Last edited: Feb 10, 2026 at 11:31 AM

Diffusion Nerd here, that's not stable diffusion only. Thats Deforum, a controllabe algorithm for generative diffusion models (like Stable Diffusion, but you can plug Flux or any other one).

I experimented a lot a few years back in A111, before animate diff and controlNets arrived. The process is simple but pretty effective and as deep as you want to go. You generate an image, then move the image some pixels in some direction, add some noise to that image and then generate another very similar but with slightly or heavy different parameters, creating that morphing effect over time.

There is this tool that allows you to guide prompts, interpolate noise and seeds that can go pretty deep in terms of audio sync, prompt morphing and parameter manipulation but it has a very big step learning curve.
https://sd-parseq.web.app/?docId=doc-c4c359e7-2dd9-4dd0-be4e-66f415f59b97

And also this other After Effect pluging that translate camera movements to deforum keyframing.
https://motionbro.com/products/ae2s...mera-data-to-deforum-stable-diffusion/?geo=es


The ones that people posted here are pretty bland on what you can achieve so there is this one of my own experimentes using that tools. You can go really crazy with the camera movements and beat sync and use them as brushes to paint new scenes. Is crazy that this tech was left behind. I'm sure we'll see some new narratives in the future using it, when we some plateau.

Later they did it manegable for the general public in this paid app. But a lot of the cool functions where the experimentation lies where stripped down, but is a nice toy nonetheless.
https://deforum.studio/

That's a big statement right there. Go ahead, teach us. You've 15mins

 

Last edited: Feb 10, 2026 at 1:15 PM

You are telling me you couldn't?

yes some packages or use cases are complex, but im talking about the fundamental creative tools that let the user create traditional art and video in AI, not some super niche geeky code driven generative path that is 'creative' in a different way (which might be the eventual creative side of AI)

I will happily share these secrets if you pay £300 to join my 'master AI to make millions' course on Linked in ;)

As you said its a 'toy', that is the right word. A lot of it is a fun toy for a few weeks, but its seriously lacking the direct input needed for nuanced image creation.

 

Last edited: Feb 10, 2026 at 1:38 PM

I actually can´t watch this. Not kidding, i don't wear glasses or something but my eyes are going nuts...like my brain cant follow this.

 

The superniche geeky code driven scene is where the soul of this new media resides. The other is just showing people how to use paid platforms. Zero room for art in there, only slop. Is the difference between using a DAW to design a Neurobass with all your knowlodeage about fx chains and compose a Neurofunk banger Vs. using MTV music generator 3 on the Playstation 2 and trying to get the same result.

 

I'm not even sure the geeky side is massively creative on its own, I kind of think this iteration of 'AI' is better integrated into established artistic tool pipelines like photoshop or 3D modeling. There's a lot of things it can help with, not replace, it should have never been called AI or hyped up.

 

Yes, I explore it heavy cause I work on the multimedia design word. There is a toon of little techs here and there that cut times and improves workflows a lot mixed with traditional tools like After Effects and such. One thing I do a lot is using comfyUI as an extension to photoshop.

The geeky side are people with good intentions toward the creative side but it's true that a lot of them lack of a proper creative vision or art direction. But they are the ones paving the road of the media that others will use in the near future.

The AI stuff. Well, it's too late now. As you can see I abstain myself of using the term. When you know how it's works calling it "intelligent" is an insult to the intelligence itself

 

im talking about dozens or even hundreds of small, deliberate changes until the randomness is gone and everything is intentional.