Stability AI’s new text-to-animation tool is closer to the community-driven tool Deforum, but can grow exponentially.
Supporting groundbreaking toolkits Stable Diffusion 2.0 and XL, the Stable Animation SDK from Stability AI will generate animations from text, images, or videos. Content creators and those in the animation industry will be able to cut costs and save time, but it also begs the question of how artistic creativity and originality be preserved.
There is no limit on the length of output, which is one of the biggest differentiators and USPs of this service. Users are already generating small animations using the tool.
The tool can process data in three forms: text, text with image, and text on video. You can access the SDK through a paid API.
At a quite primitive stage (and closer to Deforum than what one would hope), the tool can introduce small motions at best to give a sense of motion. A lot of the output looks more like stop-motion animation than full-fledged animated scenes. That’s to be expected. What’s also to be expected is that this tool will also grow at an exponential pace and the current drawbacks will be a thing of old within a month or two.
I’m not saying that this tool will make many Disney-like studios crop up, but it’s certainly going to be a gamechanger in how we tell stories. No matter how easy it has been to write, design, or even produce–It’s always been very difficult to animate your thoughts and the requirement of professionals has always been felt.
It’s an open source SDK. You can indeed install and run it locally but the GPU compute happens remotely on Stability AI’s servers.
Read the official announcement post here.
Below are 5 animations users have created using the Stable Animation SDK.
You’ll note that the output is close to Stable Diffusion’s free community-driven tool Deforum that also produces image transitions. The catch here is that Stable Animation can grow exponentially after it collects enough data and feedback and become vastly superior to Deforum or any other text-to-video or image-to-video AI tool out there.
