Uploaded January 2026 | Updated September 2026, 1 week ago
A NOTE ON REALITY, AI, AND CHOICE
For this project, we start with real photographs.
That base of reality almost always gives you the strongest and most believable starting point.
You see the real images first.
They anchor everything that follows.
BUT THIS ISN’T A REQUIREMENT
You don’t actually need real photos to create compelling AI work.
What you do need is:
• High-quality source material
• Strong visual continuity
• Realistic structure, lighting, and detail
• Consistency frame to frame
If you can generate clean, realistic, high-resolution frames, 8K and beyond, you can build incredible sequences entirely inside AI.
The starting point matters less than the quality and continuity of the materials.
WHY WE MIX METHODS
In this workflow:
• Real photography provides the foundation
• AI transitions create motion between moments
• AI-generated images occasionally step in, clearly labeled
• Everything flows together through continuity and intent
Character consistency plays a huge role here, and models like Gemini Pro already handle this surprisingly well. Future models will only push that further.
WHY I PERSONALLY LIKE A BASE REALITY
Starting from reality gives me something solid to bend, stretch, and evolve.
It makes the morphing feel grounded, even when things get abstract.
That said, this is just one approach.
Not the ceiling.
Not the rule.
THE BIGGER PICTURE
We’re still at the beginning of this kind of creation.
The tools will improve.
The workflows will get cleaner.
The line between capture and generation will keep shifting.
This is one way through that space.
More explorations ahead.
Created with Dream Machine by Luma AI
A NOTE ON REALITY, AI, AND CHOICE
For this project, we start with real photographs.
That base of reality almost always gives you the strongest and most believable starting point.
You see the real images first.
They anchor everything that follows.
BUT THIS ISN’T A REQUIREMENT
You don’t actually need real photos to create compelling AI work.
What you do need is:
• High-quality source material
• Strong visual continuity
• Realistic structure, lighting, and detail
• Consistency frame to frame
If you can generate clean, realistic, high-resolution frames, 8K and beyond, you can build incredible sequences entirely inside AI.
The starting point matters less than the quality and continuity of the materials.
WHY WE MIX METHODS
In this workflow:
• Real photography provides the foundation
• AI transitions create motion between moments
• AI-generated images occasionally step in, clearly labeled
• Everything flows together through continuity and intent
Character consistency plays a huge role here, and models like Gemini Pro already handle this surprisingly well. Future models will only push that further.
WHY I PERSONALLY LIKE A BASE REALITY
Starting from reality gives me something solid to bend, stretch, and evolve.
It makes the morphing feel grounded, even when things get abstract.
That said, this is just one approach.
Not the ceiling.
Not the rule.
THE BIGGER PICTURE
We’re still at the beginning of this kind of creation.
The tools will improve.
The workflows will get cleaner.
The line between capture and generation will keep shifting.
This is one way through that space.
More explorations ahead.
Created with Dream Machine by Luma AI
