Video-to-Video Editing with Google Omni
Google Omni
Demo summary
The demonstration shows Google Omni performing video-to-video tasks, such as replacing a physical object with a laptop, turning a piece of paper into a Pokémon card, and changing a character's clothing or species while maintaining motion.
Step-by-step
- Upload your source footage to Google Omni
- Enter a text prompt to specify the desired change (e.g., 'change to a laptop' or 'put him in swimwear')
- Review the generated video to ensure motion and likeness are maintained
Options
- Replace physical objects with different items
- Change character clothing or remove equipment (e.g., removing a parachute)
- Transform characters into different species (e.g., turning a person into a dog)
- Modify the entire environment while keeping the original motion
Watch out for
- Users in Europe may experience connectivity issues and may require a VPN to access the tool effectively
- Complex scene changes can occasionally 'break' or produce visual glitches
- Omitting a reference image can lead to imperfections in specific details like logos
Tips
- Add a reference image alongside your prompt to improve the accuracy of specific objects
- Use the tool to mix AI-generated footage with real footage for higher customizability in client work
Highlights
“allows you as an editor to have a way more customizability when it comes to editing videos”
All demos from “I Tried AI Video Editing for 8 Days - Here’s what DOES work”
- 1:202:35Using Higgsfield Plugin in Premiere ProThe creator demonstrates the Higgsfield plugin integrated directly into Premiere Pro, showing how to reframe clips for social media, upscale footage to 4K, and generate new scenes without leaving the editing timeline.Adobe Audition· AI Video Editor
- 3:552:16Generating B-roll and Assets with Higgsfield SupercomputerThe video shows how to use Higgsfield Supercomputer to create a consistent character from a reference image and generate an 'explainer video toolkit' of animated motion graphics and B-roll from text instructions.Higgsfield· AI Stock Video Generator
- 6:272:27Creating Motion Graphics with Claude and HyperFramesThe creator demonstrates using Claude with the HyperFrames extension to generate complex motion graphics, UI animations, and text pop-ups by simply describing the desired visual changes in natural language.Claude· Text to Video
- 8:581:51Video-to-Video Editing with Google OmniCurrentThe demonstration shows Google Omni performing video-to-video tasks, such as replacing a physical object with a laptop, turning a piece of paper into a Pokémon card, and changing a character's clothing or species while maintaining motion.Google Omni· Video to Video
- Watch “I Tried AI Video Editing for 8 Days - Here’s what DOES work” →
Video to Video
5:561:11Simulating physics and terrain in OmniThe creator demonstrates how changing the terrain prompt (e.g., to ice or dust) affects the physical behavior and visual effects of the vehicle in the generated video.TheAIGRID- 8:340:31Applying custom logos to video objectsThe creator uploads a custom logo and prompts the model to apply it as a theme across the car in an existing video clip while keeping the camera angle consistent.TheAIGRID
6:271:22Apply visual style presets in Gemini OmniThe user demonstrates various style templates including Noir, 80s music video, Origami, and Pixelate to transform the look of generated video clips.AI Master- 8:581:51Video-to-Video Editing with Google OmniCurrentThe demonstration shows Google Omni performing video-to-video tasks, such as replacing a physical object with a laptop, turning a piece of paper into a Pokémon card, and changing a character's clothing or species while maintaining motion.Dan Kieft