Welcome to today’s tutorial, where we’re delving into the fascinating world of animation enhancements with the SD-CN-Animation extension for Stable Diffusions. We’ll also be addressing some compatibility issues you might encounter with the latest Automatic 1111 version 1.6.
All Google Colab links here : https://thefuturethinker.org/stable-diffusion-google-colab-ipynb-list/
SD CN Animation extension: https://github.com/volotat/SD-CN-Animation
Many creators have used it to elevate their animations, but be warned, it can introduce some flickering in videos. In this tutorial, we’ll experiment with different settings to address this issue.
To get started, copy the extension’s GitHub link from the video description and paste it into the “install from URL” box in the web UI. Once installed, apply the extension, and restart the UI. Now, let’s gather some stock videos and trim them down for our experiments.
With the groundwork set, we’ll explore various ControlNet models to see which plays best with SD CN Animation. Our first example uses the ControlNet Tile model type. After setting up, we’ll generate a video. In case of flickering or inconsistencies, we’ll adjust extra settings to reduce blur and flickering.
For our second example, we’ll try another video without prompts, focusing on a consistent character output. We’ll tweak extra settings to achieve a more stable display. Each experiment may require different settings, but that’s all part of the learning process. We’ll also introduce Openpose to enhance character output.
In the end, we’ll preview our results and discuss the performance of the SD CN Animation extension compared to other tools. So, join us on this journey of experimentation and animation enhancement!