You should be arriving to watch this after seeing the video linked further down below. If you’ve yet to watch it, its purpose is to explain all the current methods and processes in creating computer generated video, its triumps, and its limitations, but most importantly, in the last 30 minutes of the video, I explain the future roadmap of the technology and where most people are getting it all wrong.
What’s wrong is the misunderstanding of where software fits into the creating; the tools that artists use, why they use them, how, and their role in being able to create art and build upon spontaneous ideas.
Meaning a lot of what exists right now is simply not fit for purpose; someone, maybe even you, needs to build it. Build it and you will free a generation.
It’s an 1hr 30 minutes long and I go through all the elements currently needed to make an animation
- Importance of Screenwriting
- MUMMY - Easy Animate
- Compositing
- Image Generation
- Tooncrafter
- Image Gen In-depth, Anime Waifus, FaceSwap
- All about RAVE video2video
- Rave’s Best Feature
- Basic 3D Compositing
- Motion Brush
- Sound Mixing
- Sound Generation
- Image2Video Loops
- the Future
The video in question
https://youtu.be/p7m3yU1MtO8 Make your own Anime (v1) - everything worth knowing about diffusion model animation workflows and its future
Make your own Anime
11th June 2024
A season of the Boondocks takes 2 years to produce In 20 years, this process has not changed. Writing is tough; animation is rough.
In 1 year, production time on a similar show will take just 6 months, without losing an ounce of artistic intention or creativity. Within 2 years, a 1 person crew, who knows what to do, can create the same show in just 3 months.
This is important in the most spiritual way. We’re almost at artistic endgame. If you want to create, you create. Create Anything. They can’t stop you.
Below is The Most comprehensive Tutorial on production-ready use of ComfyUI Workflows.
We’re at 1 and we can go all the way to 1,000,000,000,000 This is just the beginning.
click to Download workflows alternative Download
I’m a beginner to ComfyUI, KONZUKO, how do I learn?
The big problem with ComfyUI is its design is more tailored to developers than it is the traditional creator, who is reliant on good intuitive easy user experience. The install process used to be deeply painful, and so was every update, as contributors build on to it, things break and so on.
But now it should be much simpler as the development team has worked very hard. Great people. The best.
1
Installing it on linux is ideal for maximum performance with high end GPUs.
Else, MacOS and Windows are supported. I vehemently do not recommend windows, though.
If you install, and things aren’t working, this video could help: Legacy comfyui getting started
On linux, it defaults to installing everything on the root partition, so make sure your root partition has enough space for more than 100GB+ of python libraries lmao.
2
Once you’ve installed, you need to go back to school. You need to study the science of diffusion models and all that makes everything work, otherwise you won’t ever understand what or why you’re doing anything.
The beauty here is that we have llama, claude, and chatgpt, ready to explain terms and concepts when you need help.
What I personally did was use this site
I went through everything to gain a basic understanding. Spend a day on it if you must.
3
Then open up ComfyUI and get a workflow going. Everytime you run into something you don’t understand, go here https://stable-diffusion-art.com/glossary/
I must say that this is a step you can’t skip; you’ve got to know what you’re doing.
This site was also useful for me + very well formatted Complete guide to samplers
All of this is enough to begin; your will to survive will carry you the rest of the way. Good luck.
