ProofLab: AI Pre-Vis: Sketch to Image to Video
This isn’t just a demo - it’s a glimpse into the evolving alchemy of art and technology.
NeoPangea’s experimental R&D streaming series, Proof Lab, returns for another hands-on dive into the frontier of AI-powered creativity. Join us as we explore the future of AI image generation in real time.
There’s a lot of noise (and garbage) out there about everyone’s love and hate for AI but but very few actual examples of how to integrate AI tools into your concept design work flow.
Guiding the session are NeoPangea’s Founder and Chief Experience Honcho, Brett Bagenstose, and our inventive Creative AI Technologist, Aaron Randall. Together, they’ll conjure digital magic, transforming rough sketches into polished static images, and then into cinematic videos, in less time than it takes to choose an Instagram filter (all powered by cutting-edge AI software).
To truly test the system, we’ve invited Latoya Charisse Flowers, Senior Multimedia Creative at The Field Museum, to take the reins. With Latoya in the pilot seat, her prompts will steer the creative process live, as we bring her vision to life, from napkin sketch to pre-viz illustration, to full-blown animation, right before your eyes.
If you’re a concept designer for location based experiences, a filmmaker, an architect, or a crappy illustrator like me, you can use AI to interpret your sketches, turn them into pretty renders, and produce fully-rendered pre-visualizations and smooth animations.
Here’s the work flow...
Create a Rough Sketch:
Start in Procreate, Photoshop, or your preferred drawing tool. Adding some shading helps guide the next steps.
Generate the Text Prompt:
Upload the sketch to ChatGPT and ask it to help you describe the tone, style, and finish that you’re aiming for. Copy the resulting image-to-text prompt.
Refine the Image with Vizcom:
Upload the sketch to Vizcom. Paste the prompt from ChatGPT and set the Drawing Influence to 80% — this tends to be the sweet spot.
Click Generate, play with the text to make adjustments and generations, and when you’re happy with the results, download your preferred output.
Create the Video Prompt:
Upload the Vizcom image to ChatGPT and ask it to describe the image, let ChatGPT help you generate a matching image-to-video prompt to describe the scene and any camera motion. Copy the result.
Generate the Video:
Upload the image to your video tool of choice like Runway. Paste in the prompt and hit Generate.
If you want to see Latoya, Senior Multimedia Creative at the Field Museum, Aaron, creative AI Technologist, and me (Brett!) do this live, check us out in the Proof Lab: https://lnkd.in/eiVnpfcu
Runway:
https://runwayml.com/
Nano Banana:
https://lnkd.in/eA7iue7e
Vizcom:
https://www.vizcom.ai/