I watched Mark's latest video and wanted to follow his instructions exactly, except I wanted to use a background I already had. Not one generated by having AI create one based on the text prompt while using character reference images. I have a lot of very cool backgrounds created by Garry Pye, and I wanted to put them to good use. And what I finally figured out was I needed to have the background in my AI image library in Galaxy before I uploaded the character front, side, and rear images as references. When I tried uploading the background as a reference, I just didn't get the results I wanted. So, how to do that? Turns out there is an easy fix for those using Galaxy. I selected the Image AI Generator and Nana Banana Pro. In the "image discription" window. At the bottom in very small letters it says "generate prompt from image". I selected that and it allowed me to upload my background image. I then hit "generate" and it produced a comprehensive prompt based on my background image! I then generated an image from the prompt it created and, lo and behold, it was the exact same image as my background image. And now, that image was in my image library! I then selected that image, and selected "Edit image". I then uploaded all six character reference poses and wrote a prompt for placing them in the background scene. My original background on the left. The edited scene on the right. I am a happy camper! On to the AI Video Generator. Great video, Mark. It really got me on the right track.

I just released a new video showing the evolution of animation from 2 months for 2 seconds… to AI animations done in 2 days and even 30 minutes. If you’ve struggled with drift, weird morphing, stiff motion, or feeling “not good enough” to animate, you have to watch this. This changes everything for hobbyists and creators who want to finally finish their stories. - See how Garry Pye created a 22-shot trailer in 2 days that looks hand-animated. - Watch the 17-character lip sync sequence made in 30 minutes (yes, minutes). - Understand why AI animation looked broken before… and the new system that fixes it. - Discover why drawing and rigging no longer matter if you follow the right workflow. - Learn why this is the first time hobbyists can actually finish a full episode without burnout.

I generated an ai image of Mary at the table. I took that image into Photoshop and used generative expand to create a full-sized kitchen that Mark could walk into. The character styles don't match very well. I should have done a much better job of tracking my ai prompts. And I think I should have started with a full scene and the two characters in it, before trying to create character motions and lip syncing.

I uploaded an audio file from the story I'm working on, to see how well the ai would do generating lip sync. Color me impressed.

I ran Mark through a video generation prompt, and this is how it came out.