AI summary by TwinMind
The evolution of the movie backdrop
Vox
7m 20s
TikTok AI Film Reframing Trend
- Videos showing famous movies reframed for phone aspect ratios gaining popularity
- Created using Photoshop's new generative fill tool
- Tool generates imagery from simple text prompts
- Represents modern version of century-old filmmaking technique
Traditional Matte Painting History
- Technique used for 100+ years in filmmaking
- Examples: Wizard of Oz (Emerald City), Mary Poppins, Star Wars
- Dorothy/friends real, Emerald City illustrated
- Enabled impossible worlds at cheaper cost than full sets
- Original Wizard of Oz matte made with crayons
- Usually painted on glass
Max Burman Background
- Spent decade+ as matte painter in film/games
- Co-founder of Kash 3D
- Inspired by Alien, Blade Runner, Matrix films
- Career motivation: "I want to do that for a living"
Digital Matte Painting Evolution - Phase 1: Photoshop
- First chapter: transition from traditional to digital
- Photoshop enabled layer-based painting
- Revolutionary features:
- Work backwards/forwards
- Separate background/foreground layers
- Unlimited brush options
- Unlimited color palette
- First digital matte painting still felt like traditional painting
Phase 2: Photo Bashing
- Technique using real photographs as base material
- Max's work on film "Her" - future LA matte paintings
- Process: backpacked Southeast Asia 6 weeks with camera
- Used Bangkok photographs as starting point
- Modified architecture to resemble LA
- Senior matte painters initially resistant: "what the hell are you doing"
- Every technology era initially seen as "cheating" until commonplace
Phase 3: 2.5D Projection Mapping
- Pushing 2D layers into 3D space
- Max's Iron Man 3 work using Nuke compositing program
- Process:
- Painted clouds on separate layers
- Painted landscape and sky behind
- Set up as projection system
- Camera movement with Iron Man showed parallax
- Clouds moving, distance from shoreline increasing
- 2D approach cheaper/easier than 3D cloud simulation at the time
- Effective for expansive establishing shots
Modern 3D vs 2D Transition
- 3D tools have caught up significantly
- Currently as fast/faster to create 3D environments vs 2D
- Max's company Kash 3D provides pre-built drag-drop assets
- Manhattan kit example - same assets used in Doctor Strange films
- Technology improvements:
- Better rendering capabilities
- Cheaper computing power
- Reduced need for 2D artists
AI Integration Challenges
- Current AI tools not production-ready for Hollywood
- Gap between input and output requiring iterative process
- Multiple prompts + Photoshop tinkering needed
- AI lacks understanding of:
- Visual communication
- Composition principles
- Lighting techniques
- Color theory
- Tools like Nvidia AI Canvas bridging gap (brush strokes → photorealistic images)
Industry Adoption Barriers
- Legal concerns:
- Ownership of AI-generated work
- Training data ownership rights
- Quality not meeting production standards yet
- Tools thriving with individual creators, not Hollywood productions
Future Outlook
- AI will ultimately change everything
- "Human plus computer beats human and beats computer"
- Marriage of human creativity + AI technology = optimal solution
- Technology evolution inevitable
