• Welcome to Autism Forums, a friendly forum to discuss Aspergers Syndrome, Autism, High Functioning Autism and related conditions.

    Your voice is missing! You will need to register to get access to the following site features:
    • Reply to discussions and create your own threads.
    • Our modern chat room. No add-ons or extensions required, just login and start chatting!
    • Private Member only forums for more serious discussions that you may wish to not have guests or search engines access to.
    • Your very own blog. Write about anything you like on your own individual blog.

    We hope to see you as a part of our community soon! Please also check us out @ https://www.twitter.com/aspiescentral

Music video made by AI

I love these types of things. I really want to build my own neural network one day just to get a taste of the process and make it bend to my will. I have a long way to go, though.

Thanks for sharing!
 
I wish that I had this technology when I was a lot younger. I had to resort to listening to Pink Floyd while staring at traffic lights.
 
Beautiful color saturation in this video! I mostly watch these kind of videos with really dark and creepy themes – nothing that I could share here. This one is so upbeat and happy, I love it!

I think these types of videos get especially interesting when there are faces involved – the way they change and morph and they look totally normal and then totally surreal is very interesting to me.

It’s hard to wrap my head around it, – does it all start with an actual drawing at some point? Or is it strictly code? Does anyone even know what kind of software you would use to begin to make this kind of creation it’s fascinating and baffling to me at the same time time.
 
With Image AI, You can type keywords...what it generates is usually garbled mess....
(A lot of these sites dont work for me...I did find one that actually generates a picture.)

E.G. woman, sexy, smiling etc =
chick.jpg

I think the neural networks needs more improvement.
 
The exact results with something like this can vary quite a bit... it depends on the site, and the settings that they allow you to mess with. As well as the nature of the prompt that you give it.

For instance, I got this unholy thing:

Happy dog.jpg


That was done on this site: creator.nightcafe.studio

The prompt I gave it was "A friendly dog on a tropical island" while choosing a style simply called "sinister", because of course I did.

And, uh... well, I mean it IS smiling, so....


And there are other things you can do as well.

For instance, one site allows you to give it a picture... like a photo or whatever, and then give it ANOTHER picture, and have it sort of recreate the first picture using the "style" (and other aspects) of the second picture.

Example:

74a8b0f2e1e12b2bb07e1e035ce5ae9eff85652d.jpg


That image was created by using this base image of my dog Cooper:

5d174b4b7275983f6e7a9de9da78267279d998e3.jpg


(I dont know why that's showing up so small, it's actually a large photo, but whatever)

And then I chose one of my fractal art things to feed it the "style":

38c3b2ce3834a5ad6423befe3d4d4ef22bbf605a.jpg



When used carefully, these sorts of things can really produce amazing results.

Example, this:

dxzvxp_4da116a3c446482f69aab36a84c5b529aeb4ffc8.jpg


My stepmother and grandfather. A photo taken at a formal event many years ago, and I wanted to give her something related to him after his recent death back in November.

I lack the skill to paint so accurately, but I'd discovered the wild abilities of AI by that point, and I figured... why not use THAT to make it?

My father took the resulting image and gave it to a professional printer (who commented on how skilled the painter must have been, and was quite surprised to find that nobody painted it), it was printed on canvas and now sits in the front foyer of the house and is one of the first two things anyone sees when entering.

Note though that it takes a lot of fiddling and tests and trial and error to make something like that one. I went through many failed attempts before getting that result. It took HOURS.

If you want to try out the site, it is here:

https://deepdreamgenerator.com/

And if you look around the forum, awhile back I actually did a topic that was all about wacky AI created art (of the type that created the nightmare dog, not the photo+style type), so that might be entertaining to check out. I dont recall exactly where the topic is.

AI can do some crazy things, but when you want genuinely good results, you gotta really spend time with it.
 
I love these types of things. I really want to build my own neural network one day just to get a taste of the process and make it bend to my will. I have a long way to go, though.

I said this before, but I would want a HK-47 assassination droid. he despises people, is sarcastic and tends to kill his own owner some day (one disadvantage). one or a bunch of them. HK-47

the music video looks really interesting. reminds me of a dream landscape. thanks for sharing : )

I quote two posts of youtube comments, because I find them interesting:

"This is just another reason of how well developers are able to mimic how the brain works. It may not be as structured but the general concept and my anecdote may not be valid reason to support my argument but when Ive taken a High enough dose of magic mushrooms or acid, these are the exact same types of visuals I have seen"

"Someday we gonna have AI able to produce music on the fly, almost instantly. We gonna have concerts where AI will create 1-time only master pieces during the event, and people will like this ephemeral art."
 
Last edited:
View attachment 81759

That was done on this site: creator.nightcafe.studio

The prompt I gave it was "A friendly dog on a tropical island" while choosing a style simply called "sinister", because of course I did.
More like "A friendly zombie dog on a tropical island thats been contaminated with biohazardous materials from Umbrella corporation. Just off the coast of racoon city. Send in S.T.A.R.S."

Okay thats a lot of niche video game references, But I know you're into video games aren't you. ;)

So I'll just go ahead and asssume you got all that!... Yeah you posted a thread some while back about how you develop them.
 
I used some interesting keywords at "creator.nightcafe.studio", and this came out:


x9-T6-PXHdechtzmi-UDz-HJ.jpg

....stupid AI... this is just one big mess
 
That site creator.nightcafe.studio is pretty cool. I had mixed results - most were an nonsensical mess. This one was pretty interesting. The prompt was “gear factory taken over by wolves.”
1657831978039.jpeg
 
Alan Resnicks videos on DALLE 2 are very nice. He gives one prompt at the start, picks one of the few generations, then feeds the chosen image back into the AI. The result is a gradulally morphing image, you are able to see it identify contex, styles and ideas and move on from them.
 
More like "A friendly zombie dog on a tropical island thats been contaminated with biohazardous materials from Umbrella corporation. Just off the coast of racoon city. Send in S.T.A.R.S."

Okay thats a lot of niche video game references, But I know you're into video games aren't you. ;)

So I'll just go ahead and asssume you got all that!... Yeah you posted a thread some while back about how you develop them.

Indeed, I did in fact get all of that.

And now that you mention it, that thing does kinda look like something Umbrella would cause.

Alan Resnicks videos on DALLE 2 are very nice. He gives one prompt at the start, picks one of the few generations, then feeds the chosen image back into the AI. The result is a gradulally morphing image, you are able to see it identify contex, styles and ideas and move on from them.

Wait, Alan Resnick...

You mean this guy?


He and his team have made some really neat/funny/freaky stuff.
 

New Threads

Top Bottom