ABOUT THE LIGHT: try to include some instructions in the prompt to change the light type and direction....  (Like: hard light coming from the left)

My only qualm is you didnt say how you got the deph map in magnific

11:25 was an amazing result, just got to remove that lighting umbrella that stayed in the BG lol 😂 BUT MAN! I can't thank you enough for this amazing tutorial! I've always wanted to learn how to transfer real acting & camera motion into an AI cinematic generation! As my workflow was 100% AI, never Hybrid!. However, now thanks to this tutorial you helped me start creating amazing generations, that match my original filmmaking style.

Cool video! 🎉
Are you planning to do a video on invisible AI VFX? For example, when a scene is shot in a crowded place, in the wrong weather conditions, or when unwanted visual elements are left in the frame. I think these are the scenarios where filmmakers can immediately save budget. Of course, the output quality is crucial.😅

you can also render the entire scene in blender or unreal engine via MCP. sick accurate.

It makes sense to see AI become part of the production pipeline. Cantina could add another option for character-driven scenes.

That's just crazy 😮 the integration and relighting if the actors is amazing

Wow. Talk about stripping the joy out of filmmaking. One of the things we love the most about this art and craft is the opportunity to explore our world directly, not through a prompt. And any actor worth their salt will tell you how helpful it is to be in a real environment. It affects their choices, fills them with life. And any filmmaker actually there on the Irish cliffs will tell you that THAT informs how they fill the scenes. Obviously, I know that most people can't afford to shoot on location or to build out elaborate worlds. And I'm not saying that people won't create meaningful stories using A.I. tools. What I am saying is that there is a lot of us for whom the real, the tactile, that actual place, is part of the thing which opens up the possibility of art, and that the idea of "gray-boxing" our movie is like embalming life itself.

Great! Thanks for the tutorial!

Very nice, thanks!

depending on the imagegeneration AI for the First Frame, seedance (or higgsfield) give me an error: Seedance ‘Input Image May Contain a Real Person’. it looks like I couldnt use flux.3 for generating the first frame, gpt image 2.5 seems to work... any ideas where this comes from and how to work around this issue? thanks!

Do the actresses find the final output weird or uncanny or do they feel it's still them acting? How much does it modify their performances? That's the biggest issue with this workflow I think, as an actor, I wouldn't wanna be replaced almost entirely by a representation of myself.

Nice video after long time from curious refuge

Super good examples!!!

I love your content, man! Best I ever found on AI film stuff

That's crazy. Some of this wouldn't be attempted in tradigital VFX.  Thanks.

The cliff acting was better suited for a surprise cliff, not a obvious cliff. She had a hard stop implying immediate effect of falling had she taken one more step. The seaside cliff doesn't sell it.   Maybe a scene in a dark cave or a 2012 disaster type scene where the earth opens up right in front of her?

A new generation of films is nigh, where studios will have showcases to find product to distribute, and since they didn't have to budget it, they could lease the film from the makers, establish distribution in theaters and/or streaming. 2028 will likely be the year all studios move in this direction. Commercialized AI Video is "INEVITABLE".

I really like the vibe of this video

I need to try this and with more control