I'm of the same opinion as well... it's lame when a medium imitates another medium. 3D allows you to generate incredibly vivid imagery, so it's sad to see it do an unconvincing impression of cel animation and the like. 3D should find its own aesthetic that plays to its strengths and allows for maximum expression with minimum manpower, which is exactly what Japan has done with its 2D limited animation tradition. That will probably occur as its tools & workflow become less abstract, but at the moment it seems like truly great-looking, immersive 3D animation takes colossal manpower & production value. There's no "anime" analog for it yet.Leedar wrote:[...] but it feels disingenous to me to make CG animation that is trying to look like hand drawn animation.
Part of what makes hand-drawn look hand-drawn is the fact that humans are handling the so-called "transform" (rendering 3D objects on a 2D canvas). Some of the most interesting animators (Imaishi, Ohira, etc.) are able to put loads of human expression into the "transform" stage, by warping and distorting the space that the subjects & characters occupy. This dimension of expression gets thrown out the window in all those disgusting cel-shaded things. You feel it in your chest how rigid the space is in even the "best" anime-esque CG, because the transform is being handled by a deadly-accurate computer sim. This is beyond just making the lines wobbly-- it's about making human decisions on how to map 3D space into 2D.
I imagine a good future 2D-3D hybrid system would be something like an AI that can interpret a simple key drawing, something that looks like a rough silhouette with a face and many abstract hand-drawn markers & specialized symbols around it. With all that, it automatically fills in the nitty-gritty details of the character designs with lighting & shading, materials and so on handled by the renderer rather than by a person who has to meticulously illustrate it all. The artist can then correct & redraw to his heart's content, in 3D and 2D. This way, the expressive aspects like the pose & the almighty "transform" are handled by a person, while the busywork (character design details, lighting, coloring, some level of in-betweening) can be semi-automated by this Robo-AD/Colorist/Inbetweener of Justice, which blows the doors open to non-celpaint aesthetics, complex lighting and advanced materials.
Much of the real strength of 3D animation lies with its "simulations"... its simulations of light & materials, combined with sims for cloth, particles/fx, physics and etc. Back when I thought I wanted to be a 3D animator, I actually had way more fun playing with things like lighting & rendering & clothsim than slogging through the tedium of setting key poses on rigged characters. It would be magical to see those parts of 3D animation come together with 2D animation in a way that doesn't hamper either medium's inherent expressiveness.
The new Berserk movie trailers, if my eyes don't deceive me, seem to show 3D-animated characters with 2D-animated heads drawn over them. It kinda shows that there's some interest in this sort of 3D/2D hybrid future. The character designs of Berserk are hella elaborate after all. They were barely able to animate them in the old TV show.