AI video technology is rapidly evolving, exemplified by New York City-based Runway, a generative AI startup that enables users and businesses to create videos in various styles. Today, Runway launched an updated version of its Gen-2 foundation model featuring the new Multi Motion Brush tool. This innovative addition allows creators to apply multiple directions and types of motion to their AI video projects.
This capability is a groundbreaking advancement in the commercial AI video landscape. Unlike competitors, which typically apply motion to a single area or the entire image, Runway’s Multi Motion Brush allows for independent motion in multiple areas.
What is Multi Motion Brush?
Multi Motion Brush enhances control over AI-generated videos, enabling users to specify motion in selected regions. For example, users can animate a face or adjust cloud movements in the sky. To get started, users upload a still image and use a digital brush to "paint" motion onto it.
Once the image is prepared, users can customize the intensity and direction of the motion through slider controls on Runway's web interface. Each painted section can have independent settings for horizontal and vertical movement, with a range of -10 to +10. Users can easily reset adjustments by clicking the 'Clear' button.
Enhancements to Gen-2 Model
Multi Motion Brush builds on the Motion Brush feature launched in November 2023, which initially allowed only one motion type per video. The Gen-2 model, unveiled in March 2023, significantly advanced video capabilities by adding text, video, and image generation. Originally limited to four-second clips, users can now create videos up to 18 seconds long, thanks to updates in August.
New features like "Director Mode" allow users to control camera movement's direction and speed, along with options for various video styles, from 3D cartoon to cinematic. Runway competes with companies like Pika Labs, which recently introduced its Pika 1.0 video generation platform, and Stability AI's Stable Video Diffusion models.
Additionally, Runway offers a text-to-image tool that competes with Midjourney and DALL-E 3, although it's important to note that while these tools have advanced, their outputs can still sometimes be inconsistent, such as generating blurred or incomplete images and videos.
In summary, Runway's Multi Motion Brush represents a significant leap forward in AI video production, providing creators with enhanced tools for more dynamic and engaging video storytelling.