Blender 3.1 "Tooth fairy" test scene

My attempt at realistic integration of a 3d character into a video footage (actually picture).
Conclusion:
Blender compositor is very useful as always.
Blender 3.1 speed is very good.
Blender shadowcatcher is bad ( or i don't understand how it works).
Information about scene.
Render: CyclesX
Resolution: 1920х1080
Render speed: 10 sec per frame.
Graphic card: RTX 2060
Model with animation was taken from sketchfab: sketchfab.com/3d-models/mythc...
Music was taken from Yutube audiolibrary and named: The Blue Pearl - Jesse Gallagher

[@rr24creations]

*_Hope this helps:_* Blender 3.0's shadow catcher is actually very much improved over the predecessor. If you give the shadow catcher the correct material, with a decent roughness map and normal map, it actually catches not only shadows, but any emission from your CG objects and also coloured reflections (unlike the previous shadow catchers which only caught shadows)
I typically like to U -> Project from View (from the Camera View) to project one of the frames of the footage/image onto the shadow catcher object (with the shadow catcher checkbox turned off at first, so I can see what I'm doing). In your video example, basically the table image would be projected onto the shadow catcher object. Then I plug the same image into the roughness and tweak it with a color ramp node. I sometimes also use the image to drive the normals (but this may or may not give weird results, sometimes it works well, sometimes it doesn't). It gives some really convincing results 90% of the time if you manage to recreate the reflectiveness and the characteristics of the material well enough. Then of course, the last step is to turn on the shadow catcher checkbox, so that the object itself isn't rendered, only the shadows and reflections
Another point is that for some reason the shadow catcher isn't automatically set up correctly in the compositor. To correctly use the improved shadow catcher, you have to manually enable the shadow catcher render pass, and use the Multiply option on the Mix node to mix the shadow catcher pass under the CG render image pass but above the background image/footage
I'm currently working on a short CG clip that uses the technique as well (which I will post soon), but with a camera tracked video as the background instead of a still video or still image

[@Rivental]

Hello! Thank you for explenations. Yes, the shadow catcher was improved but when i was creating this video i didn't realize how it actually works. Now i would to add realistic emission influence and reflections with shadow catcher pass. (by the way i think it is not so obvious way. In corona you can turn one checkbox and you have reflection, emission, bump, without any passes that you have to add manually). Only one problem that i see with new shadow catcher it is incorrect GI influence from shadow catcher plane to object. For example i have video footage where floor has multicolored pattern. For light i use only hdri and my object is just sphere. Like you said for plane i can use project from view but how does it solve problem with incorrect light? For solving this problem my plane must to give GI like emissive material but also catch shadows and ambient occlusion. That how it works in corona.

[@uzaykisi1863]

beautiful

[@siberiancggeek5152]

Классно!

[@add7389]

процветания тебе, мир

[@add7389]

круто делаешь, жаль мало контента

[@Rivental]

Спасибо. Делаю когда есть вдохновение.

[@rembrat847]

сделай урок по созданию анимации воды в стиле гибли