From Automatic 1111 to ComfyUI: A New Chapter in AI Artistry
My Experience Learning and Adapting to the Complexities of Next-Level AI Image and Video Generation
Dear friends,
my plan worked out and I have the freedom to focus on some creative projects and experiments in the beginning of this year.
A big part of that freedom is diving even deeper into AI video and image creation.
Until now I was using Stable Diffusion Automatic 1111 to create my work. This is the default Stable Diffusion setup and is pretty straight forward and easy to understand.
But to generate animations which are much more smooth and without flickering, I had to look into a new tool called AnimateDiffusion.
I tried to avoid it until now because it looked very complex. But it is always important to keep on learning. AnimateDiffusion does not run on Automatic 1111 and needs a different Stable Diffusion setup called ComfyUI.
ComfyUI is node-based and looks intimidating in the beginning. But once you get the idea, it is actually not that difficult. The great thing is that you can build your own workflows, extend it with many great plugins and save your settings super easily.
If you are into generating images and videos with AI, it is very worth looking into it. Once I really understand what I am doing here, I will publish a new tutorial about it.
The HumanLab exhibition opened yesterday in Berlin. A lot of great digital artists are represented like Johanna Jowska, Sucuk & Bratwurst and Lucas Gutierrez.
I feel honoured to exhibit among them. And even perform a live-show this Saturday.
Ah, and we watched Barry Lyndon from Stanley Kubrick this week. It’s such an incredible movie…
Stay safe and much love,
Marius
Thanks for reading my stuff! Subscribe for free to receive new posts and support my work.
