The method in the video looks elegant, except that the shadow plane remains visible when composited over a video image. I ended up doing something similar based on
viewtopic.php?f=22&t=56315, with an extra step of compositing a tree in the foreground to pass in front of the 3D text. (The foreground tree does not appear in this frame.)
The only problem with this is that I can't control the darkness of the shadow on the ground, so if you look between the "d" and the "T" in the attached image, you'll see the pedestrians are casting a shadow on top of the composited shadow. Adding a ColorCorrector node after Merge5 lets me darken the shadows from the 3D text, but it makes the overlapped shadow darker as well.
I'm also puzzled by the jagged edges on the shadows. Any idea what's causing that?
Resolve Studio 19.1.4 (build 11); MacBook Pro 14" Nov 2024; Apple M4 Pro; 24 GB Memory; Internal 2TB SSD; Sequoia 15.4.1