Ive been trying to figure out how one would start doing something like this
If you didnt have a 3d scan of the model to work around, but only a 2d image of the person, how would you estimate distances etc? Is it all just moving things until it looks right?
And for shadows, I guess youd use a hidden plane as a shadow catcher, and angle it in such a way that it copies vaguely the shape/angle of the 2d image/person?
Thanks in advance