Jump to content

Hair modelling


sibarrick

Recommended Posts

Well that's encourging, it's just something I'm tooling around with for kicks at the moment, it's about half way through as yet so I can't post anything .

I'm basing it on a Sig paper from 2002. As a concept it would allow you to define large scale bundles, with key strands that were modelled manually and then subdivided down to refine the level of detail where you wanted it. There will be some fairly simple parameters like twist and scale, and the ability to wrap key strands around each other or other dummy strands. From there it can be extended with anything that might make it better...

The idea is that it uses Houdini's proceduralism to create a work flow to model hair, rather than trying to simulate the whole thing with particles or some such.

Any comments about the suitability of this type of approach would be welcomed.. to me I like the control it gives you but obviously it involves much manual intervention than particle methods.

Link to comment
Share on other sites

Just curious Mario, is there anything you are not interested in, Houdini/CG wise at least? :lol:

Uhmmmm....

Nope. It's all good. :P

[EDIT]

Actually... change that. There's one thing I 'm categorically not interested in: modelling! (And by modelling I mean pulling points to make a shape; not coming up with a dynamic system for modelling hair... very different things!). I can do it well and all, but it mostly just bores me to tears. <_<

[/EDIT]

I'm basing it on a Sig paper from 2002. As a concept it would allow you to define large scale bundles, with key strands that were modelled manually and then subdivided down to refine the level of detail where you wanted it.

Ooooo... that sounds pretty involved. You don't think you'll need to resort to the HDK for something like that? (I'm thinking the dynamics/collision alone will need some kind of ODE solver at least, no?). But then I've been looking at some of the stuff you've been doing with off-the-shelf OPs and, if anyone can pull it off, it's you! Go Simon! :D

P.S. I think I have the 2002 proceedings lying around somewhere; but do you know if that paper is available online?

Cheers!

Link to comment
Share on other sites

Guess you're talking about this one.

http://graphics.usc.edu/~taeyong/papers_0291_final.pdf

I only skipped the paper some time ago, but if I rember it correctly this was mainly dealing the actual modeling / creation of the hair, compared to those comb/cut style modeling as found in shave and haircut or using only splines. The animation part is yet again a completly different issue that would follow and would definitly require and ODE solver that prefably could handle as well stiff ODEs ;)

Would be cool if this modeling approach could be done with Houdni's OP's :D and I'd love to see it

Link to comment
Share on other sites

Yes that's the paper, the idea would be to just create a modelling system that allows you to quickly define a "hair style" but the dynamics etc I'd leave til H7 dynamics show up. Although I can envisage a few fake dynamics being used.

The reason I was interested in it was that I've done similar "systems" for modelling vessels, whereby you basically make a bunch of tools that are linked up in designed manner don't try and solve the whole problem in one sop. Lots of attribute passing and so on to build up what you need, this allows for styling etc. then at the end the whole model appears magically.....

I think this gives the best balance of control for the artie types and eases the pain for people like me who get bored pressing to many button to get something done that they know they could automate.

Link to comment
Share on other sites

  • 4 weeks later...

Hey,

I am all for this. I know alot of people have been asking for it. I have been wanting it myself. So yeah go for it. Cloth, Fur, Muscles, and Insert ISO params, bandsaw, knife tool are probably the biggest complaints from people coming from other software packages. You would be addressing those concerns. Cloth has been taken care of by Syflex and Sidefx is taking care of the Muscles but it must not be done yet.

Cheers,

Nate Nesler

P.S. This would be a big help to the community.

:D

Link to comment
Share on other sites

I took a quick look at the paper. I like this part:

"However, in reality when hair moves, hair/hair interactions cause hair cluster to change dynamically, from large coherently moving clusters to independently moving strands, and vice versa. Simulation of such dynamic clustering effects is a challenging problem. The preservation of user-defined style raises another issue. Currently, we do not know of any animation technique that can handle this complexity."

Nice.

Anyone doing any work with longish hair will agree with the above statement regardless of how the hair was modeled. At a base level, the entire problem should be considered not just the construction of the hair.

The key, IMHO, is for Sidefx to continue to release a strong package which has a completely thought out tool set regardless what it is doing. Beit a better compositor, dynamics, cloth, hair etc. To release something which has not be fully developed will only sour the user's experience with the package.

-k

Link to comment
Share on other sites

Can't say I have any answers for that either. I'm about half way through implimenting a modelling, styling system. Naturally it will be procedural so it may be possible to add animation controls in at various levels. I think that it's a big issue and trying to solve both issues motion and modelling all at once is too big. My approach goes like this, find the cleanest quickest and most flexible modelling approach first. Once you understand the geometry and the requirements of the modelling process that might help inform the animation rigging dynamics process. If you hit a dead end go back to the beginning and make changes that remove the blockage.

Apart from that until dynamics show up I can't do much about it anyway. :(

Still I know for a fact that dynamics are well on the way, in H7.0.136 there are some big hints as to what is coming. :rolleyes:

Link to comment
Share on other sites

  • 1 month later...

Yeah it's getting there, I just need to do one more thing and I'll have a beta with limited functionality, I'll hopefully then work it up a bit and add some more functionality. I think it could be an interesting thing, it's quite an odd mix of otl and plugin, by which I mean in order to get it to work I kinda need to force the user to work a particular way - like a plugin - but just to get it to work obviously it needs to be a whole bunch of custom sops. I've done this sort of thing before at work, but then I'm always on hand to eplain how to use it. The interesting thing will be can anyone else but me use it without help. The other interesting thing will be whether anyone else thinks the whole exercise is worth while.... :rolleyes:

Link to comment
Share on other sites

  • 2 months later...
I took a quick look at the paper. I like this part:

"However, in reality when hair moves, hair/hair interactions cause hair cluster to change dynamically, from large coherently moving clusters to independently moving strands, and vice versa. Simulation of such dynamic clustering effects is a challenging problem. The preservation of user-defined style raises another issue. Currently, we do not know of any animation technique that can handle this complexity."

Nice.

Anyone doing any work with longish hair will agree with the above statement regardless of how the hair was modeled. At a base level, the entire problem should be considered not just the construction of the hair.

The key, IMHO, is for Sidefx to continue to release a strong package which has a completely thought out tool set regardless what it is doing. Beit a better compositor, dynamics, cloth, hair etc. To release something which has not be fully developed will only sour the user's experience with the package.

-k

12914[/snapback]

Sounds interesting...

The group I

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...