Dynamic cloth under stress



  • It is a common phenonenon. Let me try to explain.
    The simulation starts without strain in the garment so there is no pressure of the garment on the figure, so also no friction. There is however gravity pulling the garment down. Initially there is no force to counteract gravity, so gravity does its work and the garment begins an acceleration in the downward direction. In the second step, there is a velocity in the negative-y direction. In the third step there is a displacement in the -y direction, so strain may begin to build up, friction builds up, and the two begin a battle against gravity. It takes some time first to build up counter force against gravity and then to reduce velocity until the garment can begin the way back up, that is, if the garment is not caught in the lowest position (max pressure so max friction)!!. In the mean while substantial sag has built up, which will have made the 'cup' in the garment slide down off the breast, leading to a void below the breast point.
    The fabric dome spanning the void is likely to collapse under the load, and that is what you see: a mis-aIigned collapsed preform 'cup' in the fabric.
    In real life the fitter would lift and 'pluck' the garment a little to make the cup return to its position. Poser clothroom does not have such a 'plucking' possibility.
    The solutions? Not one in general that is perfect but a few things that will help:

    • Make gravity come in slowly in the sim.

    • Use constrained vertices

    • Use flat clothing and build breast shape in the course of the sim.

    Have fun.



  • @fverbaas Thanks.



  • This is working well. Following with interest.



  • @englishbob said in Dynamic cloth under stress:

    This is working well. Following with interest.

    You and me both :)



  • I think I have it now. Dynamic trousers under heavy stress:
    0_1531817229460_Trousers-A18.png

    No constrained or choreographed groups at all!
    0_1531817247304_Trousers-A15.png

    Because there is nothing constrained, in this pose the waistband is pulled down at the back. Like in real. And note: The waistband keeps it's form!
    0_1531817270671_Trousers-A14.png



  • There are intersections in the figures. On thigh/shin and tigh/hip/abdomen. The figure is a standard Roxie, no morphs used, just posed.

    0_1531818822654_Bildschirmfoto vom 2018-07-17 11-08-32a.png
    0_1531818833553_Bildschirmfoto vom 2018-07-17 11-10-26a.png



  • Before someone asks: No postwork, no morphbrush used, simple shader without bumpmap or displacement.



  • @adp said in Dynamic cloth under stress:

    There are intersections in the figures. On thigh/shin and tigh/hip/abdomen. The figure is a standard Roxie, no morphs used, just posed.

    Not necessarily. Remember that there is a minimum distance of space between the cloth and the figure and that space is measured in centimeters. The only way to know for sure if there is intersection is to push the camera inside the figure from the other side to see.



  • @eclark1849 said in Dynamic cloth under stress:

    @adp said in Dynamic cloth under stress:

    There are intersections in the figures. On thigh/shin and tigh/hip/abdomen. The figure is a standard Roxie, no morphs used, just posed.

    Not necessarily. Remember that there is a minimum distance of space between the cloth and the figure and that space is measured in centimeters. The only way to know for sure if there is intersection is to push the camera inside the figure from the other side to see.

    You are right. I exported the posed figure to my modeller and explored it there. No intersections at the knees. But the thigh intersects deeply with the hip.



  • Trousers uploaded to ShareCG:
    https://sharecg.com/pf/full_uploads.php?pf_user_name=adp__

    It's the version I used to render the last images. Not optimial because I made some cuts here and there just to try. Means: Has some more Polys as needed...



  • @adp said in Dynamic cloth under stress:

    @eclark1849 said in Dynamic cloth under stress:

    @adp said in Dynamic cloth under stress:

    There are intersections in the figures. On thigh/shin and tigh/hip/abdomen. The figure is a standard Roxie, no morphs used, just posed.

    Not necessarily. Remember that there is a minimum distance of space between the cloth and the figure and that space is measured in centimeters. The only way to know for sure if there is intersection is to push the camera inside the figure from the other side to see.

    You are right. I exported the posed figure to my modeller and explored it there. No intersections at the knees. But the thigh intersects deeply with the hip.

    Just for the sake of my own curiosity, how was your Cloth Collision set up. The default is Cloth Polygon to Object vertex, I think. But did you enable Cloth polygon to object polygon?



  • @eclark1849 can't speak for @adp but I always set all 3 options in the simulation settings. I used to go with the default unless it needed changing, which it nearly always did so now I set them by default. (would have loved to be able to programmatically create the sim, but those options aren't exposed in the Python API :(



  • @eclark1849 said in Dynamic cloth under stress:

    Just for the sake of my own curiosity, how was your Cloth Collision set up. The default is Cloth Polygon to Object vertex, I think. But did you enable Cloth polygon to object polygon?

    As @amethystpendant said: All 3. And don't forget to set "collision Offset" low enough (I mostly use 0.2 or 0.15).



  • I tried hours with a low-poly model. No chance. Even not if denser on the critical parts. Because to mutch different poly-sizes results in deformation, poke-trough or randomly exploding vertices even on spots not under stress.I assume the reason is that the clothpiece is very tight.



  • @adp
    Did you take a sufficient number of steps for the sim? That is key importance and not always understood by users. Unless you are a ninja fighter changing pose in real life easily takes a few seconds. Take a stopwatch (every smartphone has one today) and check for yourself. Move as usual. Multiply the seconds with the frame rate to get a minimal amount of frames in the sim. If you have trouble, add more steps.
    The faster the move, the more energy is created that needs to be dissipated and the larger the chance the sim becomes un-stable, cloth tangles up or the vertices jump through the collision margins.
    Poser assumes linear behaviour of the fabric so lots of elastic energy can be stored quite easily. The more energy there is, the more damage can occur.



  • @fverbaas Yes, more than enough I think. First 5 frames the figure stays in the standard pose, then 50 frames to kneel down, then 20 frames to settle. And I tried "steps between frames". Interresting effect: using 4 frames here (insteed of 2) needs just a few seconds more time for the whole animation. At least for this model.



  • @adp said in Dynamic cloth under stress:

    I tried hours with a low-poly model. No chance. Even not if denser on the critical parts. Because to mutch different poly-sizes results in deformation, poke-trough or randomly exploding vertices even on spots not under stress.I assume the reason is that the clothpiece is very tight.

    A problem with increasing the density in certain areas is that unless you create a separate cloth group for the denser area, it skews the settings, those denser areas becoming effectively heavier and more "stretchy", which of course can be an advantage but needs to be taken into account.



  • @amethystpendant Yes, I know. I tried with smooth transitions between areas. But from low to high density is not enough room.
    Now I think I have a good compromise. It needs a reasonable combi of mesh construction, density and parameters. I just made another dress to test. Time between frames: 2 secs. Time needed to fix flaws after using the cloth-room: 0.0 secs :)



  • Here is the fresh dress I made. Without finetuning.
    0_1531845109625_Dress-A2.png
    0_1531845125551_Dress-A1.png

    My conclusion: Clothmakers can output several Clothpieces a week if they become familar with the clothroom and have a good workflow (and enough ideas).



  • What about the material settings like friction, inline stretchiness, bending, etc.