Stable Video Diffusion (SDV 1.1)
Unfortunately, Stable Video Diffusion (SDV 1.1) is a legacy video model that has been sunset and is no longer available on NightCafe. We recommend trying one of the newer models instead. Read more
An image-to-video model by Stability AI.
Stable Video Diffusion (SVD) is an early-stage technology, and is a bit less reliable than image generation models. Sometimes the resulting animations can be amazing, but they can be disappointing at times too.
It may take a few attempts to get a good result. SVD has limited controls, all it does is turn an image into a video (it doesn't even take a prompt) however we've found that some types of images are more likely to return good results than others.
Supported Dimensions
- 1024x576
- 576x1024
- 768x768
Inputs will be resized to match the nearest resolution.
Usage Guide
Here are our guidelines to get the most out of SVD.
- Realistic images tend to work better than cartoon or artistic styles.
- Faces, scenery and interiors tend to have a higher rate of good results.
- If you can look at an image and easily imagine how it might look as an animation, it's more likely you'll get a good one. We've found that our own intuition is a good guide for what will work well.
It's a wonderful tool.
Is it because or despite the total absence of control we have on it ?
Here's an experiment that worked
I called it an Earthquake.
Comments