Animations
WIP instructions ported from the old readme
upsampled_AndroidChairs2.0001_2.mov
The plugin will add two new types of tracks to the Sequencer which can be added using the +Track button, a Stable Diffusion Options track and Stable Diffusion Prompt tracks.
To create an animation, create a single Options track with a section that will span the length of your animation. You can right-click the section and go to properties to modify parameters that will stay consistent over the course of the animation such as the model options and output frame size.
The options track has 3 animatable channels, Iterations, Seed and Strength. All of these can be keyframed across the length of your animation.
To add prompts, you can create multiple prompt tracks containing multiple prompt sections. To set the prompt text, right-click on a prompt section, go to properties, and set the prompt text as well as the sentiment. A positive sentiment will tell Stable Diffusion that this prompt is something we would like to see in our generated output, whilst negative sentiments will instruct it to exclude unwanted prompts from the output.
Prompt tracks have a single animatable parameters, Weight. Increasing or decreasing this value will increase or decrease the influence of the prompt in your animation.
To render your animation, click the Render movie button to open the movie pipeline panel for your sequence. Add a SDMoviePipelineConfig config asset to instruct the pipeline to export a Stable Diffusion image sequence. Rendered frames by default will show up in YourProjectFolder/Saved/MovieRenders.
Note: Currently the Stable Diffusion Movie Pipeline will only export floating point images so I recommend sticking to OpenEXR as the output file type.
Click Render (local) to render your stable diffusion animation. If you haven't already initialised a model, then the pipeline will use the model options set in your Options track to initialise a model at the start of the render which may cause a white screen.