AI Video Generator By Stable Diffusion Creators Is Now Available To Public
In a blink of an eye, generative artificial intelligence has leaped from imagery to video. And in another blink of an eye, you’ll turn into a moving claymation character.
Last month, video editing software Runway teased ‘Gen-1’, its tool to convincingly apply all kinds of styles and compositions to existing videos via text or image prompts. Now, the app is ready to be used by anyone who signs up for an account.
Runway, which has been developing video editors for about half a decade, was involved in the creation of the popular text-to-image creator Stable Diffusion, so its announcement about the novel AI video maker caused a stir.
The wait is over.
— Runway (@runwayml) March 27, 2023
Gen-1 is now available at https://t.co/ekldoIshdw pic.twitter.com/Wm2YVOvm26
With Gen-1, users can stylize videos in a desired aesthetic, generate imaginative renders out of ordinary objects, mask footage, autocomplete untextured renders, and edit these results.
In the demo video above, Runway transforms a man walking down a street into a clay character straight out of a children’s show (or nightmare). It also pastes a photo of an astronaut onto a clip of a person trudging through a snowy landscape to reimagine the earthling as a spaceman exploring an otherworldly colony.
With Gen-1 already out in the open, Runway is now working on Gen-2, which will allow users to dream up three-second videos from scratch using just text prompts and nothing else. In the company’s own words, the update is “no lights, no camera, [and] all action.” Here’s a sneak peek.
A basic account on Runway is free, and it will let you create three video projects and export them in 720p resolution. Those looking for a fuller experience will have to upgrade to the US$12 monthly Standard package, which supports unlimited projects plus 1080p, 4K, and Green Screen exports; or a US$28 Pro account, which enables “all video export” formats.
Notably, Runway isn’t the only AI video service that’s just entered the scene. According to Philipp Tsipman, the founder of a generative-AI video marketing platform, five such tools have been announced in the space of a week. Yep. This is our life now.
If you’re having trouble keeping up with Video AI😅, there have been 5 state-of-the-art generative video models released *in last 7 days*: 🤯😎🧵 pic.twitter.com/cki9SMfozr
— Philipp Tsipman (@ptsi) March 26, 2023
Picsart, too, is in the midst of building a Stable Diffusion-powered text-to-video app, and it has just published its research on GitHub.
Text2Video-Zero: Text-to-Image Diffusion Models are Zero-Shot Video Generators
— AK 🤗 in SF for the Open-Source AI meetup (@_akhaliq) March 24, 2023
abs: https://t.co/5xCsj4PNRj
github: https://t.co/BdSzlepGQG pic.twitter.com/XY4piH6j4v
[via Digital Trends, Gizmodo, Runway, cover image via Runway]
