Fri, Sep 20, 5:28 AM CDT

Renderosity Forums / Poser - OFFICIAL



Welcome to the Poser - OFFICIAL Forum

Forum Coordinators: RedPhantom

Poser - OFFICIAL F.A.Q (Last Updated: 2024 Sep 19 11:01 pm)



Subject: Mesh design theory?


chris1972 ( ) posted Tue, 29 April 2008 at 2:43 PM · edited Fri, 20 September 2024 at 3:39 AM

Has anyone come across or aware of articles pertaining to mesh design principles.
Specifically what I'm after is directing lines of stress and deformation along predetermined paths. Also isolating deformation to specific areas in order to lessen deformation where it is not wanted.


Paloth ( ) posted Tue, 29 April 2008 at 2:52 PM

Mesh deformations can be defined and isolated by falloff zones in Poser or weight mapping in most other applications. Typing “weight maps” into the Wikipedia search engine should yield pertinent information.

Download my free stuff here: http://www.renderosity.com/homepage.php?page=2&userid=323368


pjz99 ( ) posted Tue, 29 April 2008 at 3:12 PM

http://megamorph.gigacities.net/TopologyGuide/TopologyGuide.htm
http://www.renderosity.com/mod/forumpro/showforum.php?forum_id=12402

Since you're asking in the Poser forum, I'll assume that you want to use your models in Poser - if you plan to take advantage of Poser's REYES polygon smoothing, it is important to have some idea how it really works.  It is very different from subdivision (Catmull-Clark e.g.) and low-poly models behave VERY differently when using either of those techniques.  You will find that low-poly models that look fine under Catmull-Clark subdivision (like most modelers use) will go mad when brought into Poser and rendered with polygon smoothing enabled.
http://en.wikipedia.org/wiki/Reyes_rendering

My Freebies


chris1972 ( ) posted Tue, 29 April 2008 at 3:51 PM

I'm building clothing meshes for use in the cloth room. My prelimenary designs have shown that there is a relationship between how the mesh is designed and how well it performs under certain stretching and deformation sequences. As an example, lets say you have a piece of clothing with a strap that runs over the shoulder, you want the clothing item to fit well and conform to the body without the stap becoming distorted in the process. I have seen in my experiments that by interrupting lines with tesselated polygons so as to isolate the strap area from continuos lines runing across the chest you can dramatically reduce distortion. Note*Conforming clothing is not the solution I'm seeking.
As a corrolary to this idea automobile safety designs use engineered crumple zones to dissapate energy and reduce deformation in one area be increasing it in another.


patorak ( ) posted Tue, 29 April 2008 at 4:07 PM

Hi Chris1972

Dynamic clothes need edgelooping,  too.  Check out this link.    http://www.newtek.com/lightwave/tutorials/animation/cloth-fx/index.html

Cheers

Pat



Conniekat8 ( ) posted Tue, 29 April 2008 at 4:10 PM

You want to have higher density of polygons where you hope for more bends and wrinkles.
It's better to have longer facet edges follow the ridges and valleys of wrinkles.

But, a big but to this is that depending on the material properties, bends and folds will behave little differently.

I haven't seen anyone in poserdom document theese techniques. If I were looking for something to that effect, I'd be scouring univerity websites with papers on applied mathemathics.
Something like this: http://search.yahoo.com/search?p=mesh+topology+affecting+cloth+simulations&fr=yfp-t-501&toggle=1&cop=mss&ei=UTF-8

Hi, my namez: "NO, Bad Kitteh, NO!"  Whaz yurs?
BadKittehCo Store  BadKittehCo Freebies and product support


ockham ( ) posted Tue, 29 April 2008 at 4:19 PM

Alternatively, you could look into the 'engineering' of real clothing design.
Examine some Simplicity sewing patterns ... look at where the darts
are formed, especially look at which sections are cut on the bias.  I'll bet there
are some webpages that give the 'arcane secrets of the seamstress'
or something along that line.

My python page
My ShareCG freebies


patorak ( ) posted Tue, 29 April 2008 at 4:27 PM · edited Tue, 29 April 2008 at 4:28 PM

Check out these links,  too.    http://maxrovat.sns.hu/subdiv/subdivmodeling.htm
                                        http://forums.cgsociety.org/archive/index.php/t-26183.html
                                        http://forums.cgsociety.org/showthread.php?f=25&t=38469



Teyon ( ) posted Tue, 29 April 2008 at 7:17 PM

unfortunately, someone hacked subdivisonmodeling.com and caused the primer to be screwed up. However, you may want to check them out still, as it's likely they'll have it up soon.


chris1972 ( ) posted Tue, 29 April 2008 at 7:27 PM

Thank you all for your help, that gives me a lot to look at.
Chris


pjz99 ( ) posted Tue, 29 April 2008 at 7:32 PM

Something to keep in mind is that dynamics is not a one-size-fits-all solution.  When posing your figure, it is common for some parts of the character's mesh to intersect itself, typically the buttocks or the armpit area when the arms are brought fully down.  In cases like these, the vertices of your dynamic cloth will "pinch" and deform in very unexpected and undesirable ways, basically squirting away from the intersection in all directions, or worse, becoming "stuck" inside the mesh of the character.  For situations like these, let's say for example a tight fitting t-shirt, imo conforming clothing is A LOT easier to work with.

My Freebies


shedofjoy ( ) posted Wed, 30 April 2008 at 4:49 AM

ok the problem i keep running up against is when making a mesh that includes lots of quad-polygons i sometimes have to use triangular ones which when i import the mesh into poser i get ugly blemishes on the object.... how do you stop this??? 

Getting old and still making "art" without soiling myself, now that's success.


pjz99 ( ) posted Wed, 30 April 2008 at 5:09 AM

While it means polygon count may be higher, it is pretty much always possible to avoid triangles if you are careful.  Take a look at that toplogy guide:
http://megamorph.gigacities.net/TopologyGuide/TopologyGuide.htm

Although Poser doesn't really have a lot of problems with triangles, they tend to be easier to UVmap and to get more predictable deformations when animating - triangles cannot flex, but quads can flex as if they were cut diagonally both ways (which they are, internally).

My Freebies


chris1972 ( ) posted Wed, 30 April 2008 at 5:54 AM

Thank you pjz99, thats very informative. The concept I'm working on is essentially a "bodysuit or wetsuit" as a base that is srink wrapped (actually the inverse) to the body for accurate fit. By getting the mesh right and fitting with cloth density set to zero (.0001). Then posing with all or nearly all vertices constrained. By using a combination of transparency  and displacement maps I'm getting good results for swimsuit type applications. I have not tried other clothing type items as of yet. Your comment regarding armpits is something I will have to check out. I have all other areas including the buttock area worked out. What I find promising in all this is that you get a perfect fit if you do it right. I have never been satisfied with conforming clothing.


chris1972 ( ) posted Wed, 30 April 2008 at 6:05 AM

file_405142.jpg

This is an example


estherau ( ) posted Wed, 30 April 2008 at 6:08 AM

 From what I have noticed of conforming clothing vs dynamic (I'm not  a modeller though) is that on the whole the dynamic clothing looks very thin, and doesn't have modelled but textured seams so doesn't seem to look as real to me a lot of the time.  Conforming clothes has more modelled seams, cuffs, buttons and zips.
Love esther

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!


lesbentley ( ) posted Wed, 30 April 2008 at 6:45 AM

the swimsuit looks great!  😄


estherau ( ) posted Wed, 30 April 2008 at 6:51 AM

 awsome!

MY ONLINE COMIC IS NOW LIVE

I aim to update it about once a month.  Oh, and it's free!


pjz99 ( ) posted Wed, 30 April 2008 at 7:06 AM

That does look very good.  About how much time is required to run the simulation into a typical pose?  How many polygons is that swimsuit?

My Freebies


chris1972 ( ) posted Wed, 30 April 2008 at 7:23 AM

Around 12,000. fitting sim and posing sim each take around 5-7 minutes each. I often am working in photoshop with 300 to 500 meg files and if I go strait from there to run a sim in the poser cloth room it runs rather slow. If I run clothroom sims on a fresh boot it runs really pretty fast. I work on a 4 year old machine with 2 gig ram and a athlon64 dual core processor.


pjz99 ( ) posted Wed, 30 April 2008 at 7:41 AM

Ah, that's kind of brutal.  Great results though.  FYI cloth room activities are single-threaded in all versions of Poser up to and including P7 (presumably Poser Pro too although that wasn't specifically announced) - you could get some more speed from a bigger processor but not as much as you might hope for, it only hits one core.

Really nice results, that's obviously a good technique.  If you are constraining all/almost all your verts then the armpit won't be any problem.

My Freebies


chris1972 ( ) posted Wed, 30 April 2008 at 8:09 AM

Actually it's not that bad, because the fitting you only have to do once for each character. You can then save each fit as a morph target. Personally I think its much faster than messing around with magnets and a bunch of dials to try an get something to fit. And there is no poke through.


Conniekat8 ( ) posted Wed, 30 April 2008 at 12:09 PM

Quote - ok the problem i keep running up against is when making a mesh that includes lots of quad-polygons i sometimes have to use triangular ones which when i import the mesh into poser i get ugly blemishes on the object.... how do you stop this??? 

Quad polygons can cause blemishes too, if they're not positioned well or if the normals got reversed.  I couldn't tell you more without actually looking at the mesh.

Hi, my namez: "NO, Bad Kitteh, NO!"  Whaz yurs?
BadKittehCo Store  BadKittehCo Freebies and product support


Conniekat8 ( ) posted Wed, 30 April 2008 at 12:11 PM

Quote -  From what I have noticed of conforming clothing vs dynamic (I'm not  a modeller though) is that on the whole the dynamic clothing looks very thin, and doesn't have modelled but textured seams so doesn't seem to look as real to me a lot of the time.  Conforming clothes has more modelled seams, cuffs, buttons and zips.
Love esther

Meshes for Posers cloth simulations need to be a lot simpler then what you can do with conforming.
With a lot of detail, the simulation can run really really slow, or not run at all, or create strange deformations on the details.

Hi, my namez: "NO, Bad Kitteh, NO!"  Whaz yurs?
BadKittehCo Store  BadKittehCo Freebies and product support


svdl ( ) posted Wed, 30 April 2008 at 1:47 PM

Quote -  From what I have noticed of conforming clothing vs dynamic (I'm not  a modeller though) is that on the whole the dynamic clothing looks very thin, and doesn't have modelled but textured seams so doesn't seem to look as real to me a lot of the time.  Conforming clothes has more modelled seams, cuffs, buttons and zips.
Love esther

Doesn't have to be that way - see my freestuff dynamic clothes. It requires more attention when modeling the outfit, and some care when setting up the dynamic cloth groups, but it is very well possible to have seams and hems with thickness to them.

Problem is, there isn't that much expertise in modeling dynamic cloth yet, even five years after its introduction in the Poser world.

Conforming/dynamic hybrids tend to work pretty well. The tighter parts work easiest as conformers, the loose flowing parts work best as dynamic cloth.

The pen is mightier than the sword. But if you literally want to have some impact, use a typewriter

My gallery   My freestuff


pjz99 ( ) posted Wed, 30 April 2008 at 2:02 PM

Best thing about dynamics:  No JCMs!

My Freebies


jerr3d ( ) posted Fri, 09 May 2008 at 6:37 PM

Attached Link: Fundamentals

i was amazed at the depth this guy goes into


Privacy Notice

This site uses cookies to deliver the best experience. Our own cookies make user accounts and other features possible. Third-party cookies are used to display relevant ads and to analyze how Renderosity is used. By using our site, you acknowledge that you have read and understood our Terms of Service, including our Cookie Policy and our Privacy Policy.