Perspective "tool" WIP

[Víctor Paredes]

Hi, I have been trying and failing with different approaches to animate some objects in perspective.
Doing that, I discovered we can use the tools of software to distort images and vectors only moving some bones.


I'm excited about the current results, but the creation process is tricky, so I'm thinking in a way to do it automatically by using a script, maybe.
If someone wants to take a look, check this file:
https://www.dropbox.com/s/6r2nfw4ghsysd ... e.zip?dl=0
I can't describe the creation process right now, but I hope to have more time soon. You can comment and ask if you want

[synthsin75]

Awesome technique, Victor.

For those who don't get how it works...show all hidden bones, and it's a complicated version of Victor's earlier technique: viewtopic.php?p=149734#p149734

[synthsin75]

Playing with this technique.

[Víctor Paredes]

Haha, that is so much better than my boring box example! Thanks for sharing it and for pointing the worm rig. I forgot that one is actually related.
Last night I was thinking about how to automatize the process and have an idea I will share later, when I have time.
Anyway, the basic idea is I'm thinking we could draw vector shapes and then a script could create the "bone grid" based on that.

[synthsin75]

Thanks, Victor. You're box shows fewer bad deformations, so you can do a lot more with it. I had to be pretty careful how far I tried to twist that car image. My rig probably could have used more bones in the grid, but it was already getting pretty confusing.


EDIT: Added more bones:

[synthsin75]

Man, I can't stop playing with this technique.

[slowtiger]

Last one reminds me of my own experiments with spherical panorama: it's convincing only in a very small area. It's still the same dilemma after all: if I need real 3D, I must learn 3D (and use 3D software).

[jahnocli]

That room has some expressive possibilities if you're not too hung up on reality though!

[Víctor Paredes]

Last one reminds me of my own experiments with spherical panorama: it's convincing only in a very small area. It's still the same dilemma after all: if I need real 3D, I must learn 3D (and use 3D software).

I disagree. For me the idea is to get good results in the easiest and fastest way posible, I think that is the principle of all software (and all tools). Of course, there are specific software for specific taks (I wouldn't recmmend Moho for composition, for instance), but I think in this case Moho bones and smart warp are helpful and easy enough to setup.
I thought about this idea preciselly because we need to animate png images quickly, interacting with the characters and with a rig the animation team can deal without issues.
Of course, the result is no always the best, but it can always be improved and adjusted.

[synthsin75]

Last one reminds me of my own experiments with spherical panorama: it's convincing only in a very small area. It's still the same dilemma after all: if I need real 3D, I must learn 3D (and use 3D software).

If you broke out the furniture as separate elements (or only animated to exaggerate the existing furniture angle) and added a few more controlling pin bones to handle the architectural details, I think this would be feasible.

Here's an outdoor scene:

[synthsin75]

Put them together:

[jahnocli]

That woodland trail is more convincing, especially if it was part of a sequence. Would be difficult to tell it wasn't 3D in passing. The car in that setting doesn't seem to maintain its volume very well though. Really interesting ideas.