Video to Anime in AI Stable Diffusion

hello everyone in this video I will show you how to make animation from any video in stable diffusion AI to begin with we will need a stable diffusion installed I showed how to do it earlier it's also good to have a control netscript installed I also have a video about it all links in the description starting with splitting your video into frames for this I will use a Premiere Pro but any video editing software will work open the video go to export medium format PNG select the folder where to save frame rate is a number of frames per second we will process each frame separately so the more frames the more time it will take in my experience 15 frames per second is enough to make the video look good click expert at the end of the export we have each frames from the video in the folder we specified go to stable diffusion the image to image stop throw any frame from the video select parameters and type runs of what you want since we will most likely have a HD or full HD frame resolution you should not specify it exactly the same resolution in the parameters I use maximum 1000 pixels it's better to upskill the result right in the stable diffusion later this will be much faster than creating frames in the full resolution you can use control net it can help to achieve more permanent results but not always sometimes it turn out better without it everything depends on the original video and the idea you have so you will need to experiment it's also important to remember that the control net saves more details so in order to achieve significant changes you need to raise the generation straight and the more this parameter is the more time it will take to generate in my case with a weak GPU every second in hundreds of frames of the video plays a big role so time for generating one image can be seen here under the resulting picture when you have achieved the desired result copy the seed and try a few more frames out of the video with the same parameters if everything looks good you can proceed to the next step go to the badge tab all the parameters of the previous generation should be saved here but just in case better to check input directory specifies the path to the folder where you store all the frames from the video while using the right folder you can simply click on the paths and copy it output directory is where all Our Generations frames will be saved you can insert the same paths and add something stable diffusion will create this folder automatically click generate and you can go get yourself some team because this process is not fast when the process is complete you may want to increase the resolution of the resulting frames this can also be done in extras batch from directory tab the process is identical here we specify folders from and where to save images I use those parameters to upscale now when everything ready we return to the video editing software and do reverse process file import open the folder with the generated frames click on the first one and make sure that the image sequence check mark is pressed click open an important point if you have upscaled your images through the extra tab then this changes the file names and it might be that the Premiere Pro does not allow you to click the image sequence check mark we need to change the names of the files for this I found a free program called Advanced renamer I can say it's super convenient but it does the work done in the project files you have a video made of our frames I drag it into the timeline and it turns out that it's a much shorter than the original this is not surprising since we have a source of 30 frames per second and the generated images are made for 15 frames it's easy to fix just stretch the video in the Premiere Pro press R key just click and stretch the video to the desired plate then you can export it as a video this is the final result we get that's all for today see you next time bye

Channel: AI Drawing

Views: 38,199

Keywords: ai, ai art, art, midjourney, artificial intelligence, artificial intelligence art, ai generated art, ai art generator, digital art, dall-e, midjourney ai, ai image generator, ai painting, stable diffusion, openai, ai drawing program, ai images, ai drawing, stable diffusion tutorial, how to install, controlnet, control net, stable diffusion model, video to animation, ai tutorial, ai tutorial for beginners

Id: khnrj1_zxMQ

Length: 5min 1sec (301 seconds)

Published: Tue Mar 07 2023