Master Runway's Expand Video: A quick guide to transform your videos
All the news from the AI world for creatives
Good day, creatives
In today’s Newsletter:
1️⃣ Runway’s Expand Video
2️⃣ Flux.1 gets new tools
3️⃣ Mistral quietly has become one of the best chatbots. And it’s free
4️⃣ Quick news you might’ve missed
5️⃣ Use this Prompt
6️⃣ Video of the Week
7️⃣ Steal this Midjourney sref
Let’s go.
Runway’s Expand Video
Runway’s Expand Video tool in for Gen-3 Alpha Turbo lets you adapt videos for multiple formats without compromising quality or creativity, empowering creators to expand their videos beyond the original frame.
You can now transform horizontal videos into vertical ones, or vice versa: this tool offers versatility and precision.
Take a look at these examples released by Runway:
What is Expand Video?
Expand Video is an advanced feature that extends the edges of your video, filling in missing details to reframe the content. It allows you to adjust your footage for different aspect ratios, ideal for creators repurposing content for various platforms.
How to Use Expand Video
Launch the Tool: Open Gen-3 Alpha Turbo, where the Expand Video feature is available.
Upload your video: If you upload a horizontal video, the tool automatically shifts to vertical expansion. You can also work with vertical videos to create horizontal formats. In this case, we will use a vertical video:
Choose your approach: After uploading, you have several options:
Generate without a Prompt: Allow the AI to interpret and expand based on the existing scene.
Provide a Text Prompt: Guide the model by describing what elements you want in the expanded areas.
Use an Outpainted Image: Upload an outpainted version of your video's first frame to give the model a visual reference.
This is the result without a Prompt:
Best Practices for Effective Expansion
To ensure optimal results, keep these tips in mind:
Position your subject: Place the primary focus near the center of the frame to simplify the expansion process.
Avoid graphics and text: Text or overlays can interfere with the AI's ability to extend the frame accurately.
Use Prompts thoughtfully: A detailed guiding text can greatly enhance the quality and relevance of the generated content.
Why this is important?
The Gen-3 Alpha Expand Video feature opens new doors for video creators, enabling them to repurpose content effortlessly for diverse platforms.
Flux.1 gets new tools
Black Forest Labs has released Flux.1 Tools, a suite of models to add control to the base text-to-image model Flux.1. These models allow inpainting, outpainting, structural guidance based on a depth map or on canny edges, and mix and recreation of input images.
Read more: Flux.1 just leveled up with four useful new tools
Mistral quietly has become one of the best chatbots. And it’s free
Mistral has updated le Chat, and you can access freely their powerful AI assistant, which now allows:
Web search with citations
Canvas for ideation, in-line editing, and export
Document and image understanding (thanks to Pixtral Large)
Image generation with Flux Pro
Quick news you might’ve missed:
KREA AI released a new AI trainer, allowing you to train your very own AI image generator model.
Freepik released its AI video generator. You can now animate any image with Luma, Runway, Mochi and Hailuo.
Suno v4 has been released, allowing you to upgrade your tracks in v4 quality, higher quality lyrics for your songwriting, and fresh cover designs to compliment your music’s vibe.
Google Gemini Advanced now has Memory. In other words, users can now save their info by telling Gemini what to remember.
Use this Prompt
Experimental modern organic building consisting of different variations of cubes which are stacked and aligned in different angles. Architecture is dancing with offsets and different views
Read more: Dive into mystic worlds with these 13 prompts for cosmic AI Art
Video of the Week
The Curse of Paganini by @isaachorror. Watch here
Steal this Midjourney sref
To use it, add --sref 787801915 at the end of your prompt in Midjourney. Your images will be generated with this style.
Read more: Use these 10+ Midjourney Sref codes for incredible AI Art