Hey everyone.
In the last year or two I have been keeping up with AI image and video generators, and recently I’ve also started experimenting with restyling my old animations (well not exactly the animations themselves but frames from my old animations).
I made a video showcasing some of my results, if anyone is interested in checking it out:
Let me know your thoughts on this.
This is excellent.
What AI services did you use for this project? It looks like you put in quite a bit of work.
Thank you!
When it comes to image generation, I did the restylings in Krita’s AI Diffusion plugin with the Flux 2 Klein 9B model. The advantage of using it in Krita is that I can directly edit the image there.
The video generation is mainly offline too, I used wan’s infinite talk for the woman’s talking scenes, some of the scenes are also wan but most of them are LTX-2.3, including the dialogues in the trailer and most of the guitar scenes but the action scenes were mainly Kling 3.0 and Sora 2, and some of the guitar scenes with Seedance 1.5 Pro (I tried Kling 3.0 first but I found it instantly changed the character’s face too much which was less of an issue with the fast paced fight scenes).
Elevenlabs for the woman’s speech, Suno for music.
I am really impressed at the female narrators casual speech narration sitting by the fireplace.
That was really well done!!
As you have obviously learned, there still is no one AI tool that’s going to do everything for you that you may need.
Right now I’m only subscribed to two systems. Mage.space & Grok AI.
My focus is 2D graphic novels , comics and 2D animation.
mage.space has some excellent models for the generation of the 2D comic. art and grock is really good at animating 2D art.
I still occasionally use a 3D render of a Daz or Iclone figure as a seed image to restyle into 2D comic art, with mage
But, for the most part, I’ve pretty much migrated away from traditional 3D programs since I retired in December.
Thank you!
Yes, the people who claim AI is an easy and lazy solution usually don’t really know what they talk about. They haven’t tried to make something with it themselves. Yes, it can be easy and lazy but the results will reflect that.
And ironically, in a lot of cases generative AI can give you more control than traditional tools.
I think it’s a great advantage that we know 3D, for example we can set up the basic outline of a whole scene with all the poses and camera angles we want and restyle those instead of trying to do it from scratch.
Another thing is characters, I can make my characters in Character Creator and that gives me more control to start with and more consistency, and I could even keep the 3D render style, or if I restyle it, the 3D renders of the exact same character helps with the consistency.
I like your approach, thanks for sharing.
Lately I have also started to experiment with AI (see My Foray into AI) and since I have been using iClone and CC for years, I firmly believe in a hybrid approach, using it as a basis. And I actually enjoy creating environments and characters, letting the AI taking care of the animation.
There are many hurdles to overcome, such as curtailing the imagination of the AI sometimes and letting it do what you want it to do. It certainly is enjoyable, although it meets with a lot of hostility.
Really interesting. Your results are amazing, and the comparisons made my jaw drop. This feels like a great way to use AI.
Thank you!
All I can say is WOW 