It’s the opening sequence of Cardcaptor Sakura, but done in Blender. What I loved about it is at the end, they show you how the scene plays out just with the meshes, then with textures, then the completed Blender project alongside the original traditional animation.
I searched the forums and didn’t see anyone do sort of a case study on it yet, and since I’m a total noob, I thought it was be a good practice to try to break down the animation, and open it up for anyone else to contribute.
I guess the first thing that stood out the most to me is the character mesh.
With the head, you can see everything is modeled–even the shape of the eyes/iris/pupil are all modeled. They didn’t use tris, but it looks like they still did textures.
Very impressive work! I remember watching a BTS of Guilty Gear Xrd where the devs said they wrote custom realtime shaders in Softimage that gave them explicit control of face (or was it vertex) normals. It was the only way for them to get the control over the cel shading they needed to make their 3d models look as good as hand painted cels, and damn it worked.
I wonder if these guys (guy? girl?) used similar methods. Do you get really explicit control over your cel shading in freestyle? I’ll have to dig into it one of these days…
OK, first up - that’s <bleep>ing amazing! Super impressed and I wish I understood Japanese so I could read the blog of his/hers (no idea). Some interesting looking videos & images but I cannot understand a word written about them
In regards to the Guilty Gear XRD technique (detailed in this video), it was vertex normals which could not be done in Blender at the time of the OP’s fanimation (it was done a year ago) and even now, as far as my understanding is concerned, you cannot tweak individual vertex normals in the way required using Blender.
