Page 1 of 1

blender people needs help with character animation code!

Posted: Tue Nov 09, 2004 7:03 pm
by birras
after a fancy new interface and other things I hope that the character animation tools will get a bit more attention. every animator is longing for improvement. Im writing this specially because the developement of an interesting plug-in for crowd simulation got stuck, because of missing futures. just imagine battle szenes like in LOTR or Troy made with blender one day....

I limit myself to quote harkyman
the original can be seen at:
http://www.harkyman.com/bp.html
Tracking the development of the BlenderPeople script suite.
Thursday, November 04, 2004
New direction.

After banging my head against the armature code wall for a long time, I am semi-giving up. I have been unable to use the current Blender code to do what I need, and don't have the time or the knowledge to write it from scratch. What I needed was pretty straightforward: a function that, given an action, an armature, a frame number and a bone, would return the object space matrix of that bone, at the given frame of the action. If anyone reading this knows how to do this, then please drop me a line, because I have everything else in place.

What I'm going to do, though instead of giving up entirely, is try to push for this to be done by someone else. It was pointed out to me that a Character Animation cleanup has been mentioned in the weekly meetings

(note that it was only mentioned). What I'm going to do is to take all of my ideas, with even some pseudocode and c code that I've already written, along with animation examples, and make a good proposal page. I still believe that Blender's character animation system could be brought up to par with other systems with a few additions:

- pinning bones to align animations between NLA strips (pinBone)
- world pinning IPO (using on/off key to lock a bone in world space)
- floor constraints that freeze bones in world space when they pass below a floor threshold
- better access to rvk's, though an improved UI and clearer creation methodology

Right now, I'm on a quest to put all old projects to bed, as I have a new one that I want to begin. Those projects include finishing a children's novel (will finish by Sunday, only three chapters to go), finishing the Audition animation, and a couple of household projects. I'll be putting together my character animation proposal after the kid's book is done.

BlenderPeople is still in there. But it will have to wait until Blender's character animation tools have matured a little. In it's final form, I see BlenderPeople as a way to not only generate funky battles, but also indeterminate character animation. A crowd of people, each going about their business; a simple way for users to move their characters realistically over terrains; ants marching; tadpoles swimming; one, two or more background characters just doing stuff without you having to plan it out specifically.

So, the goal is there. And I finish what I begin, not out of some sense of slavish duty, but because the reasons that drew me to a project in the first place remain, and interest and excite me long after the initial creative rush has passed. To anyone who has been waiting for this project, or just following it and are feeling dissappointed - learn c, download the sources and start cracking away. You might be the next great coder and just not know it yet. You might be the one who adds the great character animation tools that makes Blender a ca powerhouse. That would be cool.

Until then, though, this is going to have to wait. I'll post a link to the proprosal when I have it up.
end of quote
would be sad to see such a promising project on the shelf due to the mentioned obstacles. hopefully this will wake a little attention to the left behind animation tools in blender.

:?: :?: [/quote]

Posted: Tue Nov 09, 2004 7:47 pm
by Bellorum
Character animation improvements are, as far as I know, definitely on the roadmap for 2.36.

Re: blender people needs help with character animation code!

Posted: Tue Nov 09, 2004 10:14 pm
by reimpell
birras wrote:just imagine battle szenes like in LOTR or Troy made with blender one day....
One problem you encounter when you reuse armature animations for several meshes is the order of the transformation: You can't move the mesh after parenting the armature without destroying the visual result, e.g. try this:
  1. Create a cube.
  2. Create an armature with a single bone.
  3. Parent the armature.
  4. Rotate the bone by 45 degrees.
  5. Insert a keyframe.
  6. Select the cube.
  7. Move the cube along on of it's axes.
You will see that the center of the mesh is moved along this cube axis, but the cube itself is moved along the corresponding bone axis.

Posted: Wed Nov 10, 2004 12:42 pm
by ton
The main issue of current armature code is that it has been too quickly expanded with features to get things working for releases and projects we did in NaN. The original code design has been stretched to far beyond the max of possibilities, blocking any new development with it.

My aim is to cleanup the core of routines, with design issues in mind that should solve the current needs, and hopefully opens up new development as well. I also won't be able to do this alone, so hopefully harkyman and others will take a look at this later on as well.

Its now release week, so can't start with this work soon, but will announce on maillists and here when there's a first proposal to discuss!

Posted: Mon Nov 15, 2004 2:39 am
by LetterRip
Ton,

I'll post an overview of the research literature that I've had a look at as well as an overview and comparison chart of the current commercial offerings.

LetterRip