EZCloth Settings



  • @amethystpendant
    Having different stretch parameters does have an effect, but more complicated than one would expect.

    Below is the standard hgh density cloth pane, secured by the top row of vertices as 'choreographed'.
    The stripes in the cloth are parallel to v-direction.
    0_1491676364405_kleed0.jpg

    I tweaked the code so the u-stretch-stiffness is as specified (here default 50) vstretch-stiffness set to 0.1
    Result is non-symetry!!!
    0_1491676022752_kleed.jpg
    The sim was 300 frames and the end result was stable.

    I think the problem lies with the shear springs, the difference in stiffness and the fabric being ordered quads. Will see what happens with a Delaunay triangulation.



  • It was exactly that! When I use a random triangulation, the result is as expected!!
    Used vstretchstiffness 0.1 again.
    0_1491678244735_delaunay.jpg
    stretch and wrinkly contraction.

    Conclude:
    1 - it makes sense to explore further. I need to extend the code a little to include the other parameters.
    2 - there is sense in using irregular triangulated meshes for dynamic clothing.



  • Have you tried it swapping the stiffness U & V as I can't see why that would have an effect on a simple drop, the second picture you showed seems simply do demonstrate the left to right top to bottom calculation of the simulation engine. I'll look forward to your continued testing :)

    Amanda



  • @amethystpendant
    You are right! I was carried away. (and called away for other things so wanted to close too swiftly)

    0_1491679796854_Knipsel.JPG

    seeing the amount of sag (and some previous observations I did not share) I had assumed the U-stretch would be untouched.
    Unfortunately when I rotate the cloth 90 degrees and do the same hanging test with the u=1.0 area, I get the same sag.
    Darn!

    Sorry. Shouted too quickly.
    (hides in shame)



  • @F_Verbaas Nothing wrong with an experiment that went wrong, it proves something :) I really would have been surprised if they had suddenly decided to implement individual UV params when they never had before, altough from what I understand of the stitch API there is no logical reason why they wouldn't



  • Or not:
    there is a difference but I cannot exactly understand it:
    0_1491682451166_2curtains.jpg
    Stripes are parallel to v-axis:
    0_1491682596274_Knipsel.JPG

    Both cloths have the UStretchResistance set to 0.1 and the VStretchResistance set to 50.
    0_1491683397351_Knipsel.JPG

    It is not a misleading visual effect of the stripes:
    0_1491682477262_Knipsel.JPG

    I heve deleted the simulations, re-ran them, a few times but this remains to be the result.
    The cloth hung parallel to the u-axis wrinkles, the cloth hung parallel to the v-axis remains straight.

    Looks like I come to the inevitable conclusion that 'more research is necessary'.
    (but that will not be tonight)


  • Poser Ambassadors

    It's really cool to see this development being done.
    Seeing as the release of EZDome into the public domain has proven useful I have released the EZSkin 3 code too, in case it is also of any use.
    Get it at: http://snarlygribbly.org/snarlyspace/ezskin.html

    @amethystpendant Good luck with the rest of this project!



  • @amethystpendant said in EZCloth Settings:

    I've emailed PhilC to ask if he minded me including his material room preset values...

    From what I remember, the values weren't Phil's work anyway. They were posted on the RDNA forums by a member whose name escapes me for the present - I saved the images, but they have no name on them. However they say that the values were based on a table by Serge Marck at poserfashion (dot) net, which is long gone, although I haven't tried the Wayback Machine for it. Phil wasn't the only one to use these values in a script, either; svdl made something similar.

    If you manage to get the attributions sorted out I look forward to having this! I used to use svdl's script regularly in Poser 7.



  • @Snarlygribbly said in EZCloth Settings:

    It's really cool to see this development being done.
    Seeing as the release of EZDome into the public domain has proven useful I have released the EZSkin 3 code too, in case it is also of any use.
    Get it at: http://snarlygribbly.org/snarlyspace/ezskin.html

    @amethystpendant Good luck with the rest of this project!

    @Snarlygribbly WOW thanks, I will get it as soon as I get home! It took me longer to remove redundant stuff that I wasn't using without breaking anything, than it did to write the actual meat of the app, the access to the config files just worked, as a framework it was an absolute joy (I'm still trying to get to grips with the layout of controls but I wanted to get out the main functionality and leave the prettifying to later)

    I was planning on writing a plugin for EZSkin 3 to help use @bagginsbill's new hair shader once it's finalised, so having access to the base code will be awesome!

    Amanda



  • @englishbob said in EZCloth Settings:

    @amethystpendant said in EZCloth Settings:

    I've emailed PhilC to ask if he minded me including his material room preset values...

    From what I remember, the values weren't Phil's work anyway. They were posted on the RDNA forums by a member whose name escapes me for the present - I saved the images, but they have no name on them. However they say that the values were based on a table by Serge Marck at poserfashion (dot) net, which is long gone, although I haven't tried the Wayback Machine for it. Phil wasn't the only one to use these values in a script, either; svdl made something similar.

    If you manage to get the attributions sorted out I look forward to having this! I used to use svdl's script regularly in Poser 7.

    New version, with what PhilC used as defaults should be out tonight, including the splitting of stretch resistance in to U & V components



  • A new version has been uploaded including the defaults used by PhilC's scripts and the option to set different U & V stretch resistance values (If you modify the U version the V version will change to match but can then be changed manually)

    V1 Build 2

    I'm going to do more testing before exposing the other values that are not available in the Poser UI



  • Okay, look out for Build 3. I'm exposing the U & V scale parameters. This either shrinks or expands the mesh during the sim (1 = no change) The effect seems to be applied very early on and large deviations from 1 (eg 2 or 0.5) import a lot of energy, I ran some tests with the hanging cloths but I need to run it on some real clothes to see how useful it is, but at the moment I am really excited!!!!


  • Poser Ambassadors

    Thanks for figuring this out!



  • @wimvdb I just created a rudimentary ruched elasticated waist on shift dress! Let me play with this a bit more and I'll upload a new version, also want to test U & V fold and use edge springs


  • Poser Ambassadors

    @amethystpendant said in EZCloth Settings:

    @wimvdb I just created a rudimentary ruched elasticated waist on shift dress! Let me play with this a bit more and I'll upload a new version, also want to test U & V fold and use edge springs

    Is the U & V Scale a global setting for the simulation? I Or is it a setting per dynamic group?
    If it is the latter case, it could be used to get better fittings for the chest area as well. And that would be very welcome!



  • @amethystpendant
    That is great news!
    I understand it behaves similar to MD, so yes I can confirm scale variation should be done with care because the amount of energy introduced can be gigantic.
    Consider an elastic band. If you pull that to twice its length, you can shoot it. Good if that is what you look for but will not always be what you want.

    For those who ask what the difference is with geometry scaling:
    Think of jeans with this u v layout, with the circumference of the legs in u- direction and the length of the legs in v-direction:
    0_1491858939274_Knipsel.JPG

    If you x- and z- scale these jeans to 90% of its original size, the centres of the legs would move closer to the x=0.0 plane.
    If you u- scale it to 90% and simulate around the figure, the pant will shrink and become more skinny, vertical wrinkles will flatten.
    This scaling is best understood as 'shrinkage'.


  • Poser Ambassadors

    One minor bug to report:
    If you made a mistake and saved a preset before you created a sim, the resulting preset file can not be read and the next time ezcloth will generate an error. (this was the first time I used it before a preset file was generated).

    And one request - Can you make the default parameters on start up the same one as are currently in the sim? Some dynamic cloth already has parameters set up. Or make a button to copy them from the sim.



  • @wimvdb Thanks, I'll check that out, although I can't see why it would error just because you don't have a sim created but I can certainly try and recreate.

    I think a Copy From Current button would be best as it currently defaults to the poser default.

    Amanda



  • @wimvdb Build 3 is available with a "Copy Current" button and the U & V scale params

    @F_Verbaas thanks for the explanation I'm playing with it and can see some really useful applications for this

    Amanda


  • Poser Ambassadors

    @amethystpendant
    Build 3 works fine for me. Thanks for the addition of Copy Current!

    The U Scale and V Scale are really nice additions as well. It really allows you to tighten certain parts of the cloth. That is a big help. Just create a new dynamic group and add the vertices you want to shrink to the dynamic group. Then use U scale on that dynamic group.

    One other nice addition to the script would be to add the current actor in the EZCloth dialog box and - if possible - the current dynamic group name. If you have, you could also remove that warning dialog when applying the setings to the simulation.