Hey everyone. Today I finished my first AI Cartoon.This is a historical event about my great-grandparents. I tried to follow Mark's instructions as best I could, but sometimes I had to get a feel for things first, so it got a little mixed up at times. The Video is in German. https://youtu.be/aGT_by1Avng?si=2rAfc53eCdgHOghDhttps://youtu.be/aGT_by1Avng?si=2rAfc53eCdgHOghD @Mark Diaz I admire your enthusiasm and thank you so much for your courses.

Hello Folks, I’m trying to figure out what the best approach to this thing is. Can you tell me which AI Engines you currently use and for what purpose? Also have you subscribed to an “All in One” AI Module to use these AI Programs? Thank you, Mike Crist

What is your HOOK (dramatic foreshadow of what's to come) and what is your CLIFFHANGER (Set Up for next video)? My Hook: Full blackout with 2 eyes opening into small slits, peeking into the darkness. Then the eyes pop open. Story cuts to slow panning of a city scape with a low-voiced narrator implying a mystery of some kind. My Cliff Hanger: Story gets to the character with the eyes. My next video hook: The character with the eyes, continuing the story. I've just started using cliffhangers to enable breaking up 3-4 minute videos that are too long for viewer retention. The cliff hanger sets up for the next (continuing) video and becomes the hook for the next video.

I just came up with a messy but effective workaround for a very annoying limitation in OpenArt. I'm posting this because others may find this useful, and if anyone has a better approach I'd love to know it. The underlying problem is that, at least as far as I can tell, there is no model that will accept both Consistent Character 2.0 (pretty effective) and Start/End frames (Good except for small lighting changes). You can create a video with one or the other, but not both. This was a problem for me because I wanted to seamlessly concatenate two clips with a character remaining consistent. The first of the two clips was straightforward. I just used a consistent character with Kling 2.6, a cheap but good model. Results were great, with the character remaining consistent throughout. The problem was with the second clip. Because the joint had to be invisible, I had to use the final frame from the first clip as the starting frame for the second clip. The character was, of course, in this end clip. The transition between the two clips was almost perfect, with just a very slight change in lighting. But by the end of the second clip, the character, which it could get only from the starting frame, had changed significantly, including completely different hair! This was a deal killer. I began with a naive fix. I used 'Change Face in video' and it was a disaster. The character's face kept flipping back and forth between old and new! And so began a complex and messy approach, but which ultimately worked. I took a snapshot of the final frame of the second clip. Everything was positioned exactly as I wanted it, but the face was all wrong. I used 'Change head in image'. Turns out you MUST use a face closeup as your reference; you cannot use a full body and trust the AI to find the face. This worked excellently, correctly changing the face (and hari!) and leaving everything else alone. The only small problem was that my face reference image had him looking slightly left, while in the video he must be looking straight ahead. When it swapped the face, it also changed the direction he was looking. Annoying.

Swordsman doing his thing