Muscle Skin Deformations

[nighthwk1]

wow.. thats some amazing stuff.

the only thing i can think of is doing it the way you would in maya.. by making each muscle an influence object to the original model .. but i don't know if there is a way to get the skin to accurately deform & slide over those muscles.

[wmendez]

I posted a link to that site on XSI Base today. I think a how to is in order. He explains what the muscle rig is doing but I too would like to see how it was setup.

[skello]

whats the influence object equivalent in XSI?

[skello]

80 views no still no helpful reply:hmm: ...bump!

[MorBioS]

is there another tutorials about muscle skin deform??


my english is not good. I need read to understand.:shrug:

thanks guys

[wmendez]

What does the influence do in Maya?

[nighthwk1]

I've tried doing it with Volume Deform objects, and Deforming by Cage (using cubes parented to the bones) and neither method worked very well... I have no idea how initialprariedog bound the skin to those muscles.

[skello]

wmendez...In maya you can convert any object (modeled or primitives) into influence objects.

"If the influence object is a NURBS or polygonal surface, the skin points can be influenced by the shape of the surface. The surface can push or pull skin points that are in its vicinity, creating deformation effects that reflect the surface's shape. When placed near the surface of the skin, these polygonal influence objects can be very useful for creating deformations that indicate the effects of moving veins, bones, tendons, or muscles. You can also create interesting effects with NURBS curves."( quoted from maya help files)

[fhdatlist]

in xsi you could but any object as a deformer inside the mesh
(not only bones) so i belive if you would like to mimic the effect
mentioned in the site above maybe you could do the following:

1- made a (creature mesh).
2- made a small meshes which represent the muscle ( anyway you like)
3-apply SoftBody Deformer on it.
4- loacte it inside you creature mesh in proper hirarchiey
5-pick the soft muscles as deformers (bones)

maybe this help,but iencourage you to email those
guy who done the movie and ask them..

if they respond to you please let us know;) :p

[fhdatlist]

Originally posted by skello
wmendez...In maya you can convert any object (modeled or primitives) into influence objects.

"If the influence object is a NURBS or polygonal surface, the skin points can be influenced by the shape of the surface. The surface can push or pull skin points that are in its vicinity, creating deformation effects that reflect the surface's shape. When placed near the surface of the skin, these polygonal influence objects can be very useful for creating deformations that indicate the effects of moving veins, bones, tendons, or muscles. You can also create interesting effects with NURBS curves."( quoted from maya help files)

BTW this is the XSI deafulat behavior.

[ggg]

Originally posted by fhdatlist
BTW this is the XSI deafulat behavior.

???

[nighthwk1]

There is a similar thread on XSIbase:
http://www.xsibase.com/forum/index.php?board=11;action=display;threadid=4804;start=15

[fhdatlist]

Originally posted by fhdatlist
BTW this is the XSI deafulat behavior.

I Mean Picking any object as a Deformer (Influence)
Is The deafault option in XSI..

Common friend is it about my english or the information?:wavey: :hmm:

[ggg]

Hi fhdatlist, Sorry, I understood your response to skello that it is the default behaviour, but it is not the same as the description in the quoted paragraph.

[Atyss]

In XSI, when you use a deformer that is not a bone, you are very limited. Deformers - bones or not - in XSI only deform the envelope based on their SRT, not their shape. For example, if you assign a sphere as a deformer to an envelope, scaling the sphere will deform the envelop, but moving around its vertices won't give any results. This is why there are Cage Deformers.

So basically this
"If the influence object is a NURBS or polygonal surface, the skin points can be influenced by the shape of the surface. The surface can push or pull skin points that are in its vicinity, creating deformation effects that reflect the surface's shape. When placed near the surface of the skin, these polygonal influence objects can be very useful for creating deformations that indicate the effects of moving veins, bones, tendons, or muscles. You can also create interesting effects with NURBS curves."( quoted from maya help files)"
doesn't apply to XSI.


Salutations - Cheers
Bernard Lebel

[skello]

AAARGH!!!!:surprised ....there must be an easier way!!!!.
hmm. Maybe the skin object can be made into a cloth object and draped over the muscles

BTW. here's the reply i got from Brett Hartson.:shrug: I hope he finds time to explain how he did it:hmm: ...i may just bug him till he does :D


"Hello Isaac,

I am sorry, but i am in Nepal for four months, it is hard for me to answer any questions.
All i can recommend is you read the muscle rig, and face rig doc on my site, download the demo
scenes, and play around with it.

good luck,"

[Atyss]

Hum... very helpful. :shrug:

[ggg]

no mention of how he skinned it...

[skello]

From the video his skin seems to react really well to the muscle deformations......i'm gonna bug him like a fly till he replies....MUHAHAHAHA!!!:insane:

[skello]

I got a reply from Brett here's what he said (similar to some of your explanations)

that's simple, it is just a normal envelope.

the only difference from normal setups is that the bones are not used in the envelope, just the
muscles.

i have tried using cloth and softbody to bind skin mesh to the muscles, but that is much too slow,
and does not look much better.


Its still a bit tricky though.

[PANZER]

i don't get it. what do he mean when he says "only the muscle is enveloped... not the bone.".......... this is reallly confusing..:shrug:
a

[nighthwk1]

he's talking about binding the skin to the muscle objects instead of bones. the bones are only there for controlling the muscles.
in xsi, you can bind skin to any object.

--

anyway, to quote the_jaco over at xsibase:

the basic principle behind muscle simulation is having deformers that DO NOT pull points like current weighting does.
the system devised by IPD is nice and everything, but all it does is making the default weighting look good (in exchange for a MASSIVE setup time overhead).

Muscle simulation instead relies on having driving objects that will NOT move verices directly, but rather slide under the skin and change overall shapes WITHOUT streching the skin (and possibly translating the final vertices (not just a subDcage control) only along their normals)

[skello]

Muscle simulation instead relies on having driving objects that will NOT move verices directly

I do not agree... In order for the skin mesh to react realistically, its vertices WILL have to be moved by the underlying muscles(deformer) by all means!.....at least thats how it works in real life anyway

[ggg]

Originally posted by skello
I do not agree... In order for the skin mesh to react realistically, its vertices WILL have to be moved by the underlying muscles(deformer) by all means!.....at least thats how it works in real life anyway
Hi Skello, but if that were true then a muscle rig using the standard envelope deformer would work and the responses to this thread wouldn't exist.

The Jaco is of course saying that vertices will move as a result of underlying muscle shape or transformation but not directly, or somewhat directly with limitations.

[ThE_JacO]

Originally posted by skello
I do not agree... In order for the skin mesh to react realistically, its vertices WILL have to be moved by the underlying muscles(deformer) by all means!.....at least thats how it works in real life anyway

I can give you for granted that the way normal skinning affects the enveloped object has little to share with muscle simulation.
IPD's rigging is just a normal weighting with deforming objects instead of bones; the movement of each point is calculated by an averaged influence extrapolated from the transformations of one or more components, that means you'll have to fight surface stretching etc. etc.

obviously enough points need to be moved, if they didn't move you wouldn't have deformations at all, but how they are moved in relation to the skeletal movment differs a lot between muscular simulation and normal straight-forward rigging.

the real difference that gives away a normally enveloped mesh compared to one with an underlying muscular simulation only really shows when you see how differently lighting and texturing are affected by the 2.

I already wrote a lot on XSIbase about this, please bear with me if i won't type a redundant post here too

[skello]

Thats true.. JacO.
Don't worry!... Any light shed on this method is always welcome....I posted the same thread at XSI base also because i; and obviously some others wanted to learn more about this topic.

...I think next version of xsi should introduce a deformer that deforms based on its surface shape or topology and not by weightmaps and envelopes alone....

[ThE_JacO]

altho I understand your request, and looking at it from an abstract point of view it could seem of simple implementation, it's a damn hard thing to do.

to give up weightmaps (to be precise to give up averaged transforms in general, weightmaps are only a visual friendly representation of influent elements) you need to resort to a simulation method, and that's by far more complex then it sounds.

Not to mention that creating a human interface that you can hand to an artist to take advantage of such a method would be an impressive feat by itself.

P.S.
don't even hope that something of that magnitude, even only for a mildly complex character, could be realtime on a workstation right now, it would take A LOT of computation.
Said this it's something lots of companies are looking into, and it's something that's going to happen relatively soon, just don't expect a full-fledged real time muscular system anytime in the next months :)

[skello]

On the contrary Jac0. I was messing with a similar method in maya before my student licence expired.....Everything is realtime and has the artist friendly interface(apart from a few set driven keys..there wasn't much to it other than modeling a good looking muscular system.)...You just have to use a workflow that suits you best.

It was sort of like IPD's method but the muscles are not used as envelopes but as deformers, which influence the skin mesh based on their shape and topology( maya's influence objects).
The skin is bound to the skeleton which drive the deformation/animation of the influence objects which in turn deform the skin mesh based on their topology.
It is realy easy(will inform you if i post some wips in the next couple of months when i get back to using maya)...feedback is almost realtime.......ehem!!..but then again maybe because i use a monster of a computer:D

I was doing this from the perspective of an artist not a scripter.
Even though it may be easy for me a non-artist/bad modeler will find it demanding.

[skello]

BTW...i forgot to explain my motif.
It would have been easier to do in XSI because of the way sub-dees work in xsi.
Thats why i was hoping xsi had a deformation systen similar to influence objects.....i have fallen in love with xsi :love: ...but it just broke my heart....sob!!...snifff!!:cry:...i still like it though:buttrock:

[ThE_JacO]

I know of that method in maya, and managed to come up with something similar for several body parts in XSI as well, but it's still not simulative, while the results are nice it's no muscle simulation yet.
It relates to what I was telling you on XSIbase about using deformers to make your life easier :)
deformers and some underlying object with reduced movment will stretch skin more realistically then direct weighting, but still does't get there.

the real thing requires a skin with a thickness factor, some volume constraints and relations between muscles, pull zones for tendons, a softbody-like or cloth-like simulation that will also consider adipous layers aderhence and some other things.

I'd still suggest you to dig into weighted partial muscles (either cages or null covered surfaces) some constraints setups (IPD approach to up-vector constraints is pretty interesting but will take you lots of practice to figure out in what order to operate) and last but not least an extensive use of the push op.

[wmendez]

Take a look at this thread and the examples shown, The same can be done in XSI with envelopes.

http://www.cgtalk.com/showthread.php?s=&threadid=28922

[skello]

JacO....i'm yet to try out the "null covered surface method" and study the push operator.....soon as i get my muscles system done

Wmendez... thats a cool link....I know that can be done in xsi. but as i've said earlier i wanted a method that will eliminate messing with envelopes as well as modeling muscle features as part of the mesh..(ie. i can use a simple cylinder for the arm skin and let the deformation of the undelying anatomically correct muscles system determin the final look of the arm. I'm talking of the undelying musculature influencing the cylinder mesh on their own without the need for assigning each muscle with a corresponding weight map(or covering them with nulls)
Thus if i had a smiley face as a muscle, that smiley face will be visible when it bulges against the skin.
With the influence objects, eack point on the surface of the deformer is considered as are the points on the shin mesh...The deformation is not determined by weight maps but by the positions of the skin vertices and the various positions of the deformers(muscle) surface vertices....i hope this explanation is a little clearer :hmm:

[ThE_JacO]

it is clear, and that is what commercial muscle simulation will be about in a few years from now.
now that you so clearly depicted your idea I'm sure you can understand why you'd need to:
"dress" your muscular system with a rough surface.
shrink it over the system.
possibly specify parameters like a push lenght that will represent adipous tissues and a crease ingore threshold
manage the skin like if it was a softbody or a cloth.

[skello]

But the jaco....its already the way maya's influence objects work...it can be done already..just no one has done it yet..i guess its because of the amt of modeling involved..and as we all know maya's sub dees are not as great.

[skello]

Look even this guy created something similar using wrap skinning(wrap deformers in maya)....what i said is possible ...and can be done with either influence objects or wrap deformation...no waiting a few years for that :)
Until i show you mine look at a slightly different approach
http://www.highend3d.com/maya/tutorials/wrapskin/

"With wrap deformers, you can shape deformable objects with NURBS or polygonal objects. The shapes of the NURBS or polygonal objects you use provide the shapes of the deformation."...quoted from maya docs under Character setup>Deformers>using wrap deformers

[ThE_JacO]

If you feel like going that route you can also consider shrink wrapping in XSI, that's the closest thing you can get, in either package, to subskin deformers.

You will still miss a lot of indispensable parameters, and lots of tweaking will be needed, but in some body parts it can work (the abdomen being one); said this in lots of other regions it won't work.

[frumpy_bunyin]

Shrink wrap will give you some very good effects. You can even use it to create dynamic adjustments on meshes, such as pushing out muscles. In a simmilar way to sculpt deformers in maya.

Influence objects in maya are explicitly bound are they not? Or at least they were when I was rigging in maya. In xsi you can use any object as part of your bind. You are not restricted to joints, and then having to add additional influences.

For muscles and sliding skin in xsi you can do many things. Shrink wrap, strechy bones, skin to nulls on paths, use latices (set up properly you can slide around small lattices, scale them and what not and habe adaptive deformations, - just like muscles sliding under the skin.

I had some luck a while back with spine deformes that were curve deformed. Though I think a change made a couple of years back stopped that from working so well.

Pick and choose what will work best until we get a full muscle system. I'd still like to see a couple of requests come though that they have, to add more to this.

Ho hum.

FB

[skello]

no influence objects are not bound... Its just like the character in a cloth simulation...where the cloth is the skin and the character/colision objects are "influence objects.

[adam|zeliasz]

Has anyone gotten a response on how initialprairiedog.com set up their muscle system?

I looked into it and it seems that shrink wraping the skin around the muscles isn't practical. Using a cloth system dosen't hold the shape.

Any suggestions?

[frumpy_bunyin]

I thought they were just weighting to the muscle objects.

Did anyone download the scene?

FB

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.