Sim Cloth Testing

[steve392]

That looks like it work's good ,I can't see any passthrough of the legs at all

[Paul Forwood]

Well, it is actually harder to see the cloth penetrations when the dress is made of grass.

 

Actually, Steve, if you look carefully, the legs do pass through the skirt in example E as she walks away from the camera. This seems to be the result of increasing the mass value and the air drag value. On a denser mesh this might not cause problems. I hope to find that out soon.

[jpappas]

Paul,

 

Nice work! Looking at the animations, I think "B" looks the best.

 

A: looks a bit too springy when she twists in the beginning

B: looks very natural overall

C: fans out in the back too much as she walks, like something's holding the back up

D: also fans out in the back like C

E: unlike C or D, it seems to wrap around the back of her legs as she walks

F: fans out a bit too much in the back like C and D

 

Thank you for posting your settings. I've never used cloth but I appreciate your hard work and I'm eager to give it a first try now.

 

-Jim

[Paul Forwood]

Thank you for posting your settings. I've never used cloth but I appreciate your hard work and I'm eager to give it a first try now.

You're very welcome, Jim. Remember to allow some pre-roll time in your choreographies to let the cloth settle befor the action starts.

[NancyGormezano]

This is so great Paul - Thanks for posting your work, its extremely helpful. The tartan skirt looks terrific.

 

(and thank you Holmes as well for the written tutorial)

[HomeSlice]

Hey Paul, that skirt look great. What method did you use to attach the skirt to the figure so it wouldn't fall down? Whenever I tried to attach cloth, either by omitting a spline ring from the cloth group or by using the special "Attatched" setting in the cloth properties, I got CPs flying all over the place. I'm using AM v13s.

[Paul Forwood]

Thanks, Nancy and Holmes.

 

Holmes, I'm not using the "attached" setting and sub groups and all that. I couldn't get my head around it.

I don't know if I'm doing things the right way, just testing stuff, but here's how I did it anyway:

-------------------------------------------------

 

The skirt in these examples is actually part of the character model, though not attached. All the CPs that make up the skirt are assigned to the pelvis bone. Ensure that all of the "cloth" CPs are moved away from the pelvic area and away from the legs. Try to imagine where the CPs will fall and arrange them so that they will collide close to a CP on the underlying body. This is just to reduce the chance of a CP colliding in the middle of a patch and being missed by the collision detection. I haven't tested this enough but I think it does help.

 

1) Create two simcloth materials and change one to "deflector". Set deflector to 16 subdivisions. (This may not be neccesary but my thinking is that more subdivisions means more testing for collisions across a surface and therefore less chance of cloth penetration). Set the other material to "cloth" and set up it's properties.

 

2) In the model window of your character, create a group for all the patches that will use this cloth material and call it something obvious like,... "MyCloth". I included all the points in the skirt except for those that form the waistband.

 

3) Create groups of all the patches that will act as deflectors. Call it "MyDeflector" or whatever you like.

 

4) Drag-drop the cloth material onto the MyCloth group and drag-drop the deflector material onto the MyDeflector group.

 

5) Start a new choreography and edit the ground object so that it is made from a grid of smaller squares instead of one large patch. (I went for 1 metre spacing). Make a group for the ground and drag the deflector material into this also.

 

6) Drop your character into the choreography, add some actions and slide them along the timeline so that you have at least a second of lead-in time before the first action starts. (This is just to allow the cloth to settle. The demands of each shot will dictate how much time you will need for this "pre-roll". You may not need any.

 

7) Mumbo-Jumbo and superstition: When you have your animation set up click in the choreography window and then tap the space key a couple of times. This is something that I do every time before running the cloth simulator, and other stuff, just to make sure that the choreography updates and includes any changes that you just made to your settings. (One day all the update issues will have been fixed and I will be stuck with this habit).

 

8) Right click in a clear space in the choreography window, or on the choreography icon in the PWS, and click on Plug-ins\SimCloth Simulate. The plug-in will start calculating all the motion of the cloth and may take a few seconds or a few minutes depending on your settings and the complexity of the interactions. If you see a problem while it is constructing the simulation you can stop it by pressing "escape". When the simulation has finished play the animation to see if there are any problems. If all is good then render your choreography else go back and find what caused the problem.

--------------------------------------------------

 

Sorry. This is not very clear, is it? I will try to write up a complete tutorial sometime but I must get on with the testing and then back to my project as quickly as possible.

[MattWBradbury]

Very good results. I tried getting results like these when I was working on TWO, but I never seemed to get it down. I did probably fifty tests to see what some good results would be. I'd write down the best ones, go to a different sequence, and they would be terrible. I really couldn't understand it.

 

You wouldn't happen to have any techniques for creating cloth movement and applying it to a piece of geometry like a thick jacket and making it move realistically would you?

[Paul Forwood]

Nothing particularly good to report on that one yet, Matt.

I did spend quite some time trying to get cloth to work well as a shirt or jacket material but my best results came from using dynamic bones with a very reduced angle. I do intend to revisit this problem again though so maybe I will have better luck next time around.

[entity]

Take solver resolution down to .01 and I see you finally reduced the friction down to 1... good choices.

Cloth is not like bones where you can repeat settings precisely. EVERY simulation will vary with just slight changes to anything. There are NO MAGIC settings. We have to learn how to recognise what simcloth needs to keep away from the deflector... sometimes, that may require a longer dressing pose.

[robcat2075]

I was just wondering... how would you attach a non-flexible item like a button, or a medal to some cloth?

 

It would be of low mass so it wouldn't need to affect the simulaiton of the cloth, but it would need to travel properly with the part of the cloth it was attached to.

 

Is there some sort of spring system tool that does that?

 

Well, it is actually harder to see the cloth penetrations when the dress is made of grass.

I know you're doing basic R&D now, but would hair be a better choice for a grass skirt?

[MattWBradbury]

I would use a Newton constraint from the cloth to the medal where you wanted it, and then simulate newton physics with the cloth as a deflector. It would be really nice if we could simulate both newton physics and SimCloth at the same time. The physics dynamics should really be under one system. I'd love to see rope constraints added to A:M. It would make this problem fairly easy to solve.

[itsjustme]

I was just wondering... how would you attach a non-flexible item like a button, or a medal to some cloth?

 

It would be of low mass so it wouldn't need to affect the simulaiton of the cloth, but it would need to travel properly with the part of the cloth it was attached to.

 

Is there some sort of spring system tool that does that?

 

 

Maybe a bone for each button with a dynamic constraint (collisions turned on) on them for the springiness and a path constraint to anchor the button bones to the material (a short spline sticking up off of the surface of the material)? I haven't tried it, but it works in my head.

[Paul Forwood]

a path constraint to anchor the button bones to the material

 

That sounds well worth trying. Good one, David!

[entity]

That's a wonderful idea... and sounds like it would work. Maybe changing the orientation of the bone itself as an offset after the constraint is applied would be better than adding splines to the cloth. And is it possible to make a pose that could constrain the bones to the path/cloth spline?

[robcat2075]

Maybe a bone for each button with a dynamic constraint (collisions turned on) on them for the springiness and a path constraint to anchor the button bones to the material (a short spline sticking up off of the surface of the material)? I haven't tried it, but it works in my head.

 

This is a great idea.

 

I recall it used to be possible to select a spline on model as a path constraint target, but now I can only select a spline made in the chor with the path tool. What's gone wrong here?

[Bendytoons]

Maybe a bone for each button with a dynamic constraint (collisions turned on) on them for the springiness and a path constraint to anchor the button bones to the material (a short spline sticking up off of the surface of the material)? I haven't tried it, but it works in my head.

 

This is a great idea.

 

I recall it used to be possible to select a spline on model as a path constraint target, but now I can only select a spline made in the chor with the path tool. What's gone wrong here?

Hold down the shift key when choosing the spline, according to this thread:

Creating a path, can it be done in a model?

[robcat2075]

Hold down the shift key when choosing the spline, according to this thread:

 

Ok that works, thanks. Trying to make the constraint in a pose gets a crash in V13t. I've made an AMreport of that.

 

Here's a test of a button on a "sleeve" via a path constraint made in the chor.

 

sleevewithbuttonmp4.mov

 

 

It seems to be lagging the mesh of the cloth. Unfortunately there's no negative lag possible to try to offset that. And the button isn't maintaining it's orientation to the surface of the cloth. But it's promising. THere may be work arounds.

[mtpeak2]

What about adding a dynamic constraint to the button (its own bone, not model bone) with a fairly stiff setting and low angle limit with collision on.

 

BTW, nice test Robert.

[Paul Forwood]

Nice cloth there, Robert, and the button test was quite a success. If the dynamic bone works that will be quite a step forward.

[entity]

I did a test also, but I guess another example won't hurt. You have to set the ease to get the attached object to get to the place on that spline where you want it, but other than that... it stays with the cloth perfectly. Make sure the base of the bone associated with the object is at the place where you want the object to connect to the cloth.

 

I guess to push this idea further... if you wanted to add an oblong object that bends with the cloth, you could have several bones constrained to different splines and eases on the splines so that the cps stay with the cloth's splines also. I'll try that later.

CapePathConstraintX2.mov

[robcat2075]

You have to set the ease to get the attached object to get to the place on that spline where you want it, but other than that... it stays with the cloth perfectly. Make sure the base of the bone associated with the object is at the place where you want the object to connect to the cloth.

 

Mine has all of that, but mine is always drifting behind the motion of the cloth. I wonder why yours works so much better?

[entity]

Mine has all of that, but mine is always drifting behind the motion of the cloth. I wonder why yours works so much better?

No... your button is staying with the spline... it's large enough to penetrate where the other splines are bending/wrinkling above it. You could get it to ease to another less active spot on the cloth or get it to average between two bones positions on the spline (button bone -"translate to" two "path constrained" bones?), to keep it out of the fold/curve.

Interesting problem there. Maybe offsetting the button bone from the spline a little to keep it outside the cloth would be a quick way to fix.

Mine uses a sphere so the connecting "side" of the sphere is against the cloth, not quite touching the cloth.

[Mohammad]

Hey is simcloth in SimA:M or is it something in the newer versions of A:M sorry if you already posted that, come on there are five pages now

[entity]

AM-v13 and v14 have "new" simcloth. It's in the manuals. I believe AM also has "old" simcloth in version 12... not sure about that, but this thread applies to any cloth in AM. Only thing is - simcloth is more simple than ever to use! You can use 5 point patches without too many problems but hooks are a no-no.

 

Anyway, I have new examples with dynamic constraint applied to a cone attached to the cloth. I tried a flat button having the button bone "aim at" another bone that is path constrained to another paralell spline to the button. It keeps the bone's angle changing with the cloth. I also did a final render. Sometimes when you see penetration, it's only penetrating because you are looking at a lower res version of the patches and actually there is no penetration, so it's best to look at a final render to see if it is true at the best resolution before you abondon it. (4 poly per patch or 16?) Most people have the default setting for patch "quick render" mode: 4 p per patch.

 

I better get back to my cloth tut, now. Thanks guys!

CapePathConstraint3dynwX2.mov

Cloth_attached_parts12fr.mov

[Paul Forwood]

Well, it looks like dynamic bones on a path constraint work very nicely!

Thanks for the input, Rich.

[mtpeak2]

Nice tests Rich, they look very promising.

[HomeSlice]

Excellent tests Rich! Those look awesome.

Did you apply a deflector material to the button and the dangly thing?

[entity]

No- I have been using a simulation that was exported from the chor as an action. The button and dynamic floppy cone were path constrained to the cape. The cape is penetrating the leg because I lost the action for the legs that go with the cape simulation action. If I had "baked" in the chor then the legs actions would have been exported with the simulation data. Anyway, the point was to see how you could attach objects onto cloth... read the whole thread because there is information on how to layer cloth on cloth, too.

Lot's of good ideas going on here.

[Ramón López]

Ey, Robcat2075 (HI!) I've been working/investigating latelly a lot with that "Path Constraints" inside Distortion Boxes and other similar situations with or without SimCloth with some crashes here and there too, cause of this I'm very interested in the progression of this issue, so... well, do you think/know if your 0004655 *AM:Report* will be solved in the next A:M v14.0 Beta 6 too? I hope/suppose it'll be, but as actually it only appears in the 13.0 u "Change Log" and not in the v14.0 "one", I only wanted to be sure just in case another *AM:Report* for v14.0 must be reported... Well, and that's all for now, THANKS!

[robcat2075]

Ey, Robcat2075 (HI!) ... well, do you think/know if your 0004655 *AM:Report* will be solved in the next A:M v14.0 Beta 6 too?

 

A note from Noel says "Resolved: You should be able to test this in v14 beta 6 to be released soon. "

 

give it a try when beta 6 comes out. I reported it in V13 because that was what I had.

[Ramón López]

Ah, THANKS for the answer! It was only that I was not sure what really happens when you report a bug in v13 that "surelly" will exists too in v14, I thought that probably the bug would be searched for the programmers in both versions to kill him definitely at the same time, but... it seems that the best thing in that cases is report it for both versions if possible?I say it cause I did it in that way for another issue but I was not sure if I was doing the correct thing... Well, the good news is just now I've seen v14 Beta 6 is available so we can know it inmediately , Let's try it!

[Ramón López]

And finally it works!!! And well, I can't be more happy cause I'm putting a lot of hope in all that marvellous relationship possibilities and his mixing and combination between other features, sometimes (although I'm sure I'm not doing nothing new here) up to unsuspected limits... (I think )

[entity]

Okay guys... a quick test... Say you want to add someobject that has the properties of cloth, but is made up of variable shapes and parts... like say you want to make a chain or rope that can act like cloth. What do you do?

 

Well since cloth is made up of a flat mesh, you can add objects to the surface via path constraint to the cloth's splines, to allow the abjects to stay with the cloth... well! This is done with constraints/bones, so why couldn't you do it also with bones in an object like a rope or a backpack (Woot).

:: You can create a cloth grid and set up as you would normal cloth on a character, then create some bones at each of the cp's of the cloth. Now you have a grid of bones (don't parent them... keep them all seperate). Create a pose (ON/OFF) and start constraining those bones to the splines (path constraint). You will have to use "ease" to get the bones into the exact position on the spline, because the bones will attach to the spline at "0% ease" which is the begining of the spline. Now, you need to keep these bones handels pointing to the bone above in the grid, so use "aim roll at" to do that. Now, you got all these bones attached to the cloth when you turn on the pose in an action... wait... theres more!

Next: you can create that belt or chain or backpack and rig it like your cloth grid (these bones need to be controlling the bulk object... only use the amount of bones you really need to make it work: the bones nearest the cloth will be the main bones.). Next, make a new pose (ON/OFF). This pose will constrain your object bones to the closest cloth bones using "translate to" and "orient like". MAKE SURE CORRESPONDING BONES ARE ORIENTED AND IN THE SAME POSITION! That is a basic rigging rule... even if your not orienting like it's always a good idea to do it so your rig looks clean (it doesn't have to be perfect, just neat). ... Now you have all these object bones in the exact position and orientation of the cloth bones.

Now hide (H)your object and simulate your cloth. This is just a test... doesn't need to be perfect or long lasting!

Now, unhide your object and make sure your two poses are ON. Make the cloth transparent 100%, so it doesn't show, unless you want it to show.

 

Look at it.

There is an object distorting and bouncing around your character, something that can't be done as well with dynamic bones. If your object is sinking into your character... try setting your collision radius larger to keep the object from penetrating your character too much.

Hope you can understand this... I have a sample animation and project (2.62mb-prj^v14.0...and mov files zipped) I did of a simple test, but the object could almost be anything that dangles... hehehe (get your mind out of the gutter!)... well you will see.

I'm going off to play with this idea a little more... it might even be good for making characters rigs react to a gun blast. Hmmmmmm. I'll be back later!

 

::basicly, this technique is a way to have cloth deform a bulky object, just like it does with a button except using a bunch of bones instead of one.

 

EDIT:I noticed later that the ease changed during the animation from 0% @ frm0 to 100% @ frm120 in the chor (I set the poses to ON in frame 0 of the chor). I'm going to look into this, but I think you have to set a value for ease in each (path)constraint in the pose so it doesn't change over time.

 

Edit#2::It's true. For the path constraints, the"ease" must have a keyframe set so it doesn't automaticly animate to 100% over the length of the animation in the chor. (I was using 0% to position the bone on the spline: but I must enter "0" into the "ease" field to create a keyframe in the pose, otherwise A:M will animate the ease automaticly) I fixed the pose in the proj I uploaded here so you can get the new one. Sorry.

....Now I'm off to test this on something else!

Cloth_Path_control_mesh.zip

Edited August 15, 2007 by entity

[HomeSlice]

Create a pose (ON/OFF) and start constraining those bones to the splines (path constraint)

 

Me and Steve Sappington were trying to get the path constraint to work in an ON/OFF pose a couple weeks ago but I couldn't get it to work. I tried turning the constraint ON AFTER the simulation in the chor and also BEFORE the simulation, but I could never get the pose to actually constrain the bone to the spline unless I defined the constraint in the chor.

 

How did you get the Path Constraint Pose to work in a chor?

[mtpeak2]

I just did a quick test with constraining to a path in a pose and it works fine.

 

Did the path constraint work when you created the ON/OFF pose?

[entity]

Create a pose (ON/OFF) and start constraining those bones to the splines (path constraint)

 

Me and Steve Sappington were trying to get the path constraint to work in an ON/OFF pose a couple weeks ago but I couldn't get it to work. I tried turning the constraint ON AFTER the simulation in the chor and also BEFORE the simulation, but I could never get the pose to actually constrain the bone to the spline unless I defined the constraint in the chor.

 

How did you get the Path Constraint Pose to work in a chor?

Well it's a good habit to close the chor and action windows that are using the pose as well... that could be it. But I was using v14.0 when I did this... never tried it in the betas or v13.

 

Did you download the project I uploaded? I didn't have any trouble...

... a work around would be to do this in an action and copy the keyframes to the pose.

[entity]

...

I recall it used to be possible to select a spline on model as a path constraint target, but now I can only select a spline made in the chor with the path tool. What's gone wrong here?

Hold down the shift key when choosing the spline, according to this thread:

Creating a path, can it be done in a model?

Here above is explained how to select a spline in chor.