Creative

Rigging a machine.

Things have been going absurdly slow lately. No commits to WIPUP. No new ThoughtScore models (though a few more seconds of video have been added). Nothing open-source related (except for trying out the Ubuntu beta1 on a box). Even schoolwork has slowed.

Because I fully emphathise that people with a grip of things wouldn’t give a rat’s ass about my life, I decided to show some pictures of the trouble I’ve been having trying to rig Taras, one of the main characters in The ThoughtScore Project. Here are two statically posed shots of Taras:

The left shows him in his unrigged pose. The pose he was modeled in. The right shows him "looking" down, slightly bent forward with his left arm reaching towards you. Disregarding the fact that the lighting is completely faked (what is that suit reflecting, anyway?), we have two other major problems to deal with.

Problem Number One: His arm was not built to be in that pose. Not was any other part of his anatomy. When standing straight his arms are abnormally squashed in order to look natural in that one pose… and when in a dark environment. In any other scenario you’d see two spindly arms sticking out of a hunk of metal. The way it was designed, his shoulder "ball and socket" joint is more of a "plank of wood stuck on a block of wood" joint. It doesn’t fit nicely like a joint should.

Put simply, all of his joints (legs included) will have to be remodelled in order so that you don’t have gaping holes or bits of the suit intersecting when limbs are moved in their extremeties. Not an easy task.

Problem Number Two: The torso. The torso is made up of several different meshes. Each part fits together nicely in one way and one way only. If you look at the picture, you’ll see that when he leans forward, the upper torso covers the middle torso, which largely remains stationary, the groin panel shifts outwards slightly, and the piping all has to move to accomodate this change and not randomly stick out where it shouldn’t. Think of it like the parts of a steam engine.

Long story short, it’s going to be a PITA to rig that guy just to bend over. Heck, I don’t think you can bend over in a suit like that.

Normally I stubbornly plod down the road of "create first, learn later, fix and redo even later", but this time I think I’d better buy some of Blender’s training DVDs before continuing on ThoughtScore.

Dion Moult

I've been developing software for well over 10 years, work as an architect (not the computer kind, the regular sort), and am classically trained as a pianist. I try to do the right thing when I get the chance in my field, such as through contributing to open-source communities and promoting sustainable living.

More Posts

Spread the love
Tagged: , , , , ,

2 Comments

  1. You’re actually creating a 3D movie on your own? WOW! I cannot believe how tough that must be. Creating a 2-d 4 panel comic takes me 4 hours or more. An animated 3D movie must take years!

    And yes, 3D character design must be awfully painful because of the various interactions in space with the character’s animated poses.

    Best of luck with this.

  2. The ThoughtScore Project has been in production for over 3 years now (on and off, juggling other projects like WIPUP as well). With my underpowered laptop I can probably only clock in about 10 hours a week at most not to mention overnight renderings. Not being very skilled at it either adds to the time as I have to learn from scratch.

    So yeah, it’s an ambitious project, but gosh am I learning a huge amount from it.

Leave a Reply

Your email address will not be published. Required fields are marked *