How to Use dForce: Creating a Blanket, Draping Clothes on Furniture, and Much More [Commercial]

1568101139

Comments

  • sberlinssberlins Posts: 128

    Thanks for this thread!

  • RGcincyRGcincy Posts: 2,805

    Thanks sberlins

    Just wanted to point out a thread showing how to get a dress to drape across a figure and a horse. It cleverly uses animation of a horse/figure that starts upside down and then flips into place. It illustrates that with dForce you sometimes have to be creative in how you set it up to get the results you want. 

  • RGcincyRGcincy Posts: 2,805
    edited January 2018

    22. Banner Flag

    Flags are a natural for dForce. Since there’s a test scene of conventional flags in dForce Starter Essentials (which comes with Daz Studio 4.10), I decided to test a longer banner flag that is found in FirstBastion’s Ridge Walk Mountains. That flag is already configured by the PA to look like it’s blowing in the wind but you can also transform it with dForce. After running a few tests with his flag, I replaced it with a primitive plane of similar size to see how that looks. Here are the results.

    a. This is what the flag prop looks like. It has the flag, a pole, and a set of rocks.

    b. Select the flag prop and from the main menu select Edit/Object/ Geometry/Add dForce Modifier: Dynamic Surface.

    c. Run the dForce simulation. What you will see is the whole prop begins to sag:

    d. Since the pole and rocks have separate surfaces, you can quickly turn them off by going to the Surfaces pane and set Visible in Simulation to off.

    e. Now you’ll find that the flag slips down the pole. To correct, you need to add a weight map to keep the portion wrapped around the pole in place.

    f. Add a weight map as normal by having the flag prop selected then go to the main menu and choosing Create/New dForce Modifier Weight Node.  Select the new node in the Scene pane, go to the Tool Settings pane and select Node Weight Map Brush, then click the Add Map button. Using the brush and holding down the Alt key, remove color from the area around the pole as shown below (area in gray).

    g. Now when you run the simulation, the flag stays at the top of the pole while the flaps hang down.

    h. Add a wind node and set it to blow at an angle to the flag (since the prop was built to have the look of wind blowing, blowing it in a different direction lets you see the effect wind is having). I set Wind Speed to 5 mph which will make the flag stream horizontally.

    Side View: 

    Top View at Start:

    Top View midway: 

    Top View at end:

    The flag after the simulation is complete:

    dForce banner prop.jpg
    705 x 522 - 166K
    dForce banner prop whole deformed.jpg
    693 x 522 - 163K
    dForce banner prop 3 surfaces off.jpg
    402 x 131 - 25K
    dForce banner slides down.jpg
    703 x 522 - 164K
    dForce banner weight map.jpg
    687 x 513 - 130K
    dForce banner after map.jpg
    698 x 520 - 165K
    dForce banner wind side view.jpg
    566 x 446 - 32K
    dForce banner wind node.jpg
    569 x 494 - 84K
    dForce banner wind blowing flag.jpg
    554 x 481 - 84K
    dForce banner wind blowing flag end.jpg
    574 x 504 - 91K
    dForce banner with wind.jpg
    702 x 517 - 168K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,805
    edited January 2018

    i. Next I added a primitive pane with 50 divisions and sized and positioned it to match the previous banner.

    j. The primitive plane won’t wrap around the pole, so I kept that part of the original banner. To do this, select the flag prop, go to the Tool Settings pane and select the Geometry Editor, then using the tool, select all the polygons of the flag that project from the pole and assign them to a new surface. Next go to the Surfaces pane and set that surface to an opacity of 0 and it will be hidden from view.

    k. Here is the new flag at 3 wind speeds.

        0.1 mph:

        

        5 mph:

        

        10 mph. At this higher speed, the end of the flag gets compressed into a tiny strand:

        

     

    l. Depending upon the look you desire, there are some surface parameters you can change. These all use the higher wind speed of 10 mph.

    • Influence is the quickest way to cut back on flag movement:

              1.0 (default):

              

             0.88:

             

     

    • Stretch Stiffness has slight effect, mostly changing how the end of the flag wraps in the wind. Note that as with some other dForce surface parameters, small changes can make a difference:

               0.8 (default)

               

              0.9

              

             0.98

               

           1.0

               

     

    m. Mesh resolution has a big effect, but this is harder to achieve as you have to make a new plane (I tested using sub-D on a low resolution mesh but it wrapped the same as the original low mesh).

         50 by 50 mesh (used in images above):

         

         20 by 20 mesh:

         

        10 by 10 mesh:

        

    dForce banner new.jpg
    695 x 516 - 155K
    dForce banner rectangular 50 0.1mph.jpg
    1200 x 900 - 834K
    dForce banner rectangular 50.jpg
    1200 x 900 - 836K
    dForce banner rectangular 50 10mph.jpg
    1200 x 900 - 826K
    dForce banner rectangular 50 10mph 1.0 influence.jpg
    1200 x 900 - 823K
    dForce banner rectangular 50 10mph 0.88 influence.jpg
    1200 x 900 - 835K
    dForce banner rectangular 50 grid 0.80 stretch stiffness.jpg
    1200 x 900 - 828K
    dForce banner rectangular 50 grid 0.90 stretch stiffness.jpg
    1200 x 900 - 828K
    dForce banner rectangular 50 grid 0.98 stretch stiffness.jpg
    1200 x 900 - 830K
    dForce banner rectangular 50 grid 1.0 stretch stiffness.jpg
    1200 x 900 - 831K
    dForce banner rectangular 50 grid.jpg
    1200 x 900 - 823K
    dForce banner rectangular 20 grid.jpg
    1200 x 900 - 826K
    dForce banner rectangular 10 grid.jpg
    1200 x 900 - 842K
    Post edited by RGcincy on
  • sapatsapat Posts: 1,735

    Thank you for this section on the banner.  Much appreciated for content and your attention to detail.  

  • I don't think I understand. I've tried to follow this yet can't seem to replicate the results. I tried a simple scene with a plane and a sphere yet struggled to get the plane to conform to the sphere correctly.

    First, I converted the plane to subd and gave it a few subdivisions. I forgot how to make it keep its square shape, but whatever.

    https://drive.google.com/open?id=1jdzFSyZLOLJzjAJAY1XbrhOO3L06lnjQ (Not sure how to properly post images on here)

    then, I opened the simulation pane and simulated it with the default settings, causing this

    https://drive.google.com/open?id=1i8LFEobvyBkK15ZCG7Rgdz1gJhoUqTdD

    I tried fiddling with different settings and tried placing the plane under the sphere but they ultimately lead to similar flops.

    What exactly am I doing wrong?

  • How many divisions did you give the plane when you created it? It needs those, not the virtual divisions from SubD. (Keeping corners and edges in SubD is controlled by the Edge Interpolation setting under Mesh Resolution in the Parameters pane.)

  • sapatsapat Posts: 1,735

    Yes, RGcincy told us that a primitive plane might need upwards of 50-100 divisions when you create it.  I had a problem with a plane reacting properly and that's what he told me to do.

  • RGcincyRGcincy Posts: 2,805

    I don't think I understand. I've tried to follow this yet can't seem to replicate the results. I tried a simple scene with a plane and a sphere yet struggled to get the plane to conform to the sphere correctly.

    ... What exactly am I doing wrong?

    What's the resolution (how many polygons) on your plane? You probably need something like 50 or more so there are places for the mesh to bend. It looks like there's a lot of polygons but if they are from sub-D that won't help

  • Any advice on dynamic settings for chainmail?  It's heavy, but made of a lot of joints.

  • RGcincyRGcincy Posts: 2,805
    edited January 2018

    Any advice on dynamic settings for chainmail?  It's heavy, but made of a lot of joints.

    My sense is that chainmail if fairly flexible because of all the joints but it's weight would make it resistant to wind and slower to change position when moving about.. I would also think weight would also cause draping to be relatively vertical. A high air resistance might help. Have you tried a render? If so, post it and say what doesn't look right about the draping.

    I ran a test on a chainmail shirt (from Knight Guard for G2M). It draped fine but I only used the default dForce values. No weight map either. Here's the result:

    Knight Guard no displacement.jpg
    960 x 1200 - 1M
    Post edited by RGcincy on
  • I'm using the Voyager for M6 tunic that I've retextured as chainmail.  When I try to dforce, it goes sideways instead of down.

    This is without dForceimage

    And this is dForce with a weight map locking tight everything bust-up.image

    no dforce.png
    667K
    dforce default.png
    635K
  • RGcincy said:
    What's the resolution (how many polygons) on your plane? You probably need something like 50 or more so there are places for the mesh to bend. It looks like there's a lot of polygons but if they are from sub-D that won't help

    Err, I'm not sure how to look that up or change that. But I did try this on an actual blanket model with its own set of polygons and it did the same thing. It kinda falls down a bit into the shape of a Pringle's chip as if it acts like the model underneath doesn't exist.

  • RGcincyRGcincy Posts: 2,805

    Err, I'm not sure how to look that up or change that. But I did try this on an actual blanket model with its own set of polygons and it did the same thing. It kinda falls down a bit into the shape of a Pringle's chip as if it acts like the model underneath doesn't exist.

    That's odd. Not sure what's going on. If those are primitive models, you could post a scene.duf file and I could take a look.

    You can find the number of polygons by selecting the object, going to the Tool Settings pane and select Geometry Editor. You'll see a category called Face Groups and a column called Total. Total is the number of polys for each Face Group (there may be only one or there may be dozens depending upon how the mesh was made).

  • RGcincyRGcincy Posts: 2,805
    edited January 2018

    I'm using the Voyager for M6 tunic that I've retextured as chainmail.  When I try to dforce, it goes sideways instead of down.

    I have that item so I took a look. After suffering through a half dozen system crashes, I finally got it to simulate by turning off Visible in Simulation for everything except the tunic, setting Self Collide to off for the tunic, and hiding the big belt polygons in the Geometry Editor (not sure that's needed). I used default dForce surface parameters, gravity, and air resistance. I set the dForce simulation to animated using the timeline and went from a zero pose at frame 0 to the shown pose in frame 15. I also added a chainmail shader (different than yours). 

    Here's the weight map:

    After the simulation render. A little tight across the back with some indications of mesh stretching:

     After turning the belt back on. It's a little loose in this view; from the other side, it has chainmail poke through.

    No unusual lifting. I had all the zbrush and adjust fit morphs turned off. You don't by chance have one of those on? 

    dForce voyager chainmail weight map.jpg
    413 x 591 - 110K
    DForce voyager chainmail.jpg
    960 x 1200 - 1M
    DForce voyager chainmail w belt.jpg
    960 x 1200 - 1M
    Post edited by RGcincy on
  • I had some fit control morphs dialed in.  It looks like turning off self-collide was the big thing - I've got it working now.  The belt's giving me a bit of trouble because I want to drape it a bit differently than the rest, so I'll probably just hide it and put a different belt on. 

  • RGcincy said:

    That's odd. Not sure what's going on. If those are primitive models, you could post a scene.duf file and I could take a look.You can find the number of polygons by selecting the object, going to the Tool Settings pane and select Geometry Editor. You'll see a category called Face Groups and a column called Total. Total is the number of polys for each Face Group (there may be only one or there may be dozens depending upon how the mesh was made).

    Well, here's the scene

    https://drive.google.com/open?id=1PECtz85yo-cl_pvksKfJKSMWZ7Ec2o_4

    and here's the screenshot 

    https://drive.google.com/open?id=1jlgJyAcltpK_P6W2j9jONG_Y3a69J8nk

    I didn't see "total", only "count", and the number was 1

  • RGcincyRGcincy Posts: 2,805

    I didn't see "total", only "count", and the number was 1

    The means it has only 1 polygon. That won't drape as there needs to be multiple polygons. 

  • nicsttnicstt Posts: 11,714

    I notice you used the default cubes in a tutorial; I found those helpful and interesting btw, thank you.

    I sub-divided a cube in Blender, did a basic unwrap and added to Studio - available on my deviant art page: http://fav.me/dbz5kwj

  • RGcincy said:

    I'm using the Voyager for M6 tunic that I've retextured as chainmail.  When I try to dforce, it goes sideways instead of down.

    I have that item so I took a look. After suffering through a half dozen system crashes, I finally got it to simulate by turning off Visible in Simulation for everything except the tunic, setting Self Collide to off for the tunic, and hiding the big belt polygons in the Geometry Editor (not sure that's needed). I used default dForce surface parameters, gravity, and air resistance. I set the dForce simulation to animated using the timeline and went from a zero pose at frame 0 to the shown pose in frame 15. I also added a chainmail shader (different than yours). 

    Here's the weight map:

    After the simulation render. A little tight across the back with some indications of mesh stretching:

     After turning the belt back on. It's a little loose in this view; from the other side, it has chainmail poke through.

    No unusual lifting. I had all the zbrush and adjust fit morphs turned off. You don't by chance have one of those on? 

     

    Did you have the belt's collision layer set 1 higher than the shirt? I've found that really helps. Also Collision Offset. I've had to go as high as 0.6 on some items.

  • RGcincy said:

    I didn't see "total", only "count", and the number was 1

    The means it has only 1 polygon. That won't drape as there needs to be multiple polygons. 

    Hmm. Not sure how to "increase" polygons beyond subd, as I didn't see it being done in the tutorial. 

    But I also had that same chip-shaped, object-ignoring problem with an actual bed cover prop with multiple polygons. I tried yet another bed like this

    https://drive.google.com/open?id=1ivNEoYJPD7WXaRuA5hqfTp5cWfUvK37H

    but it resulted in this mess

    https://drive.google.com/open?id=1xWEG_OJ0kZJW_pK1LkM_DNzr68csVxOk

  • RGcincyRGcincy Posts: 2,805

    Hmm. Not sure how to "increase" polygons beyond subd, as I didn't see it being done in the tutorial. 

    But I also had that same chip-shaped, object-ignoring problem with an actual bed cover prop with multiple polygons. I tried yet another bed like this

    You can only increase the number by taking the object into a modeling program where you can add more polygons. If you are following what I wrote, a square primitive plane with about 50 divisions will work (I have not had success with circular meshes I have available).

    The bed prop has an irregular polygon pattern with a lot of triangular polygons. My limited experience with them show they can be harder to drape as you are finding. In terms of the pictures you linked to, those long triangular strands come from where the bed cover mesh is touching another object. Those places of touch act to secure the meshes together and you'll see this elongated stretching if the rest of the mesh rolls up or explodes. Since the mesh is bunching up, I'd suggest that in the surface pane you (1) turn Self Collide to off (2) reduce Dynamic Strength to something between 0.8 and 0.9 and see if that helps. 

  • RGcincyRGcincy Posts: 2,805
    edited January 2018

    Did you have the belt's collision layer set 1 higher than the shirt? I've found that really helps. Also Collision Offset. I've had to go as high as 0.6 on some items.

    No I didn't. I tried your suggestion and the belt still had lots of poke through and mangling. Just for learning sake, I tried some more things and got it to work:

    1. The belt on the tunic has separate polygons but they sit right on top of the underlying tunic material so they get frozen to it and take on it's draping.

    2. To eliminate that, I used two copies of the tunic (not instances). Tunic 1 had the belt polygons selected in the Geometry Editor then:

    • 1st: right click in viewport and select Geometry Selection/Select By/Surfaces/Big Belt
    • 2nd: right click in viewport and select Geometry Visibility/Hide Selected Polygon(s)
    • 3rd: right click in viewport and select Geometry Editing/Delete Hidden Polygon(s)

          Tunic 2 was similar except choose Hide Unselected Polygon(s) in 2nd step.

    3. Tunic 1 used the weight map I showed in a previous post. Tunic 2 had a weight map that was filled to a constant 0.4.

    4. Tunic 2 was assigned Collision Layer 2 and Collision Offset of 0.5 in the surfaces parameters.

    5. Tunic 2 was parented to the figured (not fitted). It was also moved up the Y-axis off the hips so the polygons didn't touch.

    6. On the animation timeline, the belt moved down to the hips as it proceeded.

     

    Here's the front view and you can see the belt has no poke through and looks smooth across the hips and waist.

     

    Here's the back view.

     

    Note polygon stretching is causing stretching in the chainmail texture just above the belt (you see similar stretching on the lower hem). That probably means the belt touched a little early and locked the polys together, so some fine tuning could be done. The angle between the legs is fairly extreme compared to normal posing, so that puts more strain on the fabric and the simulation

    dForce voyager chainmail tunic take 2.jpg
    960 x 1200 - 1M
    dForce voyager chainmail tunic take 2 back.jpg
    960 x 1200 - 1M
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,805
    edited January 2018

    I've started to take a more detailed look at the scenes Daz provided in dForce Starter Essentials to further my understanding of dForce behavior.  Here is the first one I've looked at.

    23. Effect of Surface Properties Using Simple Sheet Drop - Part 1. The simple sheet drop is a scene in the dForce Starter Essentials, which comes with Daz Studio 4.10. The sheet has a number of simulation surface parameters that are different than the dForce defaults. In this section, I’ll compare the original scene settings to the default surface properties, then show what happens when you change an original setting back to the defaults one by one. The only changes to the scene I made was to move the sheet so it doesn’t cover the figure’s head and use different shaders. It's easiest to see differences if you click on the images which will open them in a new tab and you can flip through them.

    a. Original scene settings and figure’s pose. With the knee up, there’s more room for draping than what I had in my original blanket tests.

     

     

    b. Original scene settings (left or top, depending upon how your browser displays them) vs. dForce defaults (right or bottom). The scene settings result in a less stiff draping that more closely hugs the figure and bed. The fabric is more rounded over the knees and foot and there is more dipping between high points. The defaults might be better for a stiffer blanket while the scene settings would be better for a sheet.

     

     

    c. One by one replacements. Because the scene settings result in a drape much different from the original, keeping 5 of the 6 scene settings lets you see how one set back to default is affecting the drape. For each parameter change, I show an image of the resulting sheet (blue) plus a second image with the original sheet (yellow) also visible. Whichever color is visible indicates a higher elevation. 

    d. Stretch Stiffness at default (0.8 vs. original scene at 0.01)

      

     

    e. Shear Stiffness at default (0.2 vs. original scene at 0.01)

     

     

    f. Bend Stiffness at default (0.5 vs. original scene at 0.01)

     

     

    g. Buckling Stiffness at default (5% vs. original scene at 2%). The intermingled colors on the overlay image shows buckling stiffness had little effect.

      

     

    h. Buckling Ratio at default (70% vs. original scene at 98%)

     

     

    i. Density at default (180 vs. original scene at 80)

     

     

    j. Conclusions. It’s easiest to see differences if you open each image in a separate browser window. You can then flip through them to see what is happening.

    • Resetting Stretch Stiffness to default gave more and longer folds.
    • Resetting Shear Stiffness to default pulled the fabric in closer to the figure and bed and had sharper folds.
    • Resetting Bend Stiffness to default gave steeper valleys than the original combination or stretch stiffness at default.
    • Resetting Buckling Stiffness to default had no discernable effect.
    • Resetting Buckling Ratio tot default gave more folds than the original. It also gave more rounded folds than when shear or bend stiffness were at defaults.  
    • Resetting Density to default suppressed almost all the folds. It also caused the sheet to hang vertically down the side of the bed. It pulled the fabric closer to figure and bed than when shear stiffness was at default. Without the lower density, the folding effect of the other parameters is largely missing.
    • dForce default settings give a shape most similar to bend stiffness at default but with very stiff folds and little droop between the high points. Looks more like a stiff rather than silky fabric. Interesting that although the density was high, it did not suppress folds (see previous bullet). 
    dForce sheet test parameters.jpg
    266 x 381 - 36K
    dForce sheet test body pose.jpg
    1200 x 900 - 397K
    dForce sheet test original settings.jpg
    1200 x 900 - 313K
    dForce sheet test default settings.jpg
    1200 x 900 - 314K
    dForce sheet test original settings ex. default stretch stiffness.jpg
    1200 x 900 - 316K
    dForce sheet test original overlay ex. default stretch stiffness.jpg
    1200 x 900 - 479K
    dForce sheet test original settings ex. default shear stiffness.jpg
    1200 x 900 - 317K
    dForce sheet test original overlay ex. default shear stiffness.jpg
    1200 x 900 - 461K
    dForce sheet test original settings ex. default bend stiffness.jpg
    1200 x 900 - 322K
    dForce sheet test original overlay ex. default bend stiffness.jpg
    1200 x 900 - 473K
    dForce sheet test original settings ex. default buckling stiffness.jpg
    1200 x 900 - 313K
    dForce sheet test original overlay ex. buckling stiffness.jpg
    1200 x 900 - 680K
    dForce sheet test original settings ex. default buckling ratio.jpg
    1200 x 900 - 321K
    dForce sheet test original overlay ex. buckling ratio.jpg
    1200 x 900 - 481K
    dForce sheet test original settings ex. default density.jpg
    1200 x 900 - 305K
    dForce sheet test original overlay only density.jpg
    1200 x 900 - 481K
    Post edited by RGcincy on
  • sapatsapat Posts: 1,735

    Interesting results.  Lots of odd little ripples in most every one though.  Any cure for that?

  • RGcincyRGcincy Posts: 2,805
    edited January 2018

    I don't know. Depending upon where the ripples are, lower friction on the sheet or the bed/floor might help (but too low and things can slide right off).  Also, I didn't resize the sheet when I moved if off the head, so more is on the floor at the foot of the bed (the sheet lying on the floor at the sides of the bed was in the original scene). Making the sheet fit so it doesn't hit the floor might help.  

    Post edited by RGcincy on
  • sapatsapat Posts: 1,735
    RGcincy said:

    I don't know. Depending upon where the ripples are, lower friction on the sheet or the bed/floor might help (but too low and things can slide right off).  Also, I didn't resize the sheet when I moved if off the head, so more is on the floor at the foot of the bed (the sheet lying on the floor at the sides of the bed was in the original scene). Making the sheet fit so it doesn't hit the floor might help.  

    Yes, maybe it's as simple as resizing the sheet.  I'm not well enough to concentrate on this or I'd give it a go.  Need to wait til I'm better.

  • RGcincyRGcincy Posts: 2,805

    Sorry to hear you are ill. Hope you feel better soon.

  • sapatsapat Posts: 1,735
    RGcincy said:

    Sorry to hear you are ill. Hope you feel better soon.

    Thanks so much :)

  • wsgentrywsgentry Posts: 572
    edited January 2018

    Just thought I'd add that dForce can be used on some jewelry items like farah bangles to make them slide naturally up and down arms.

    Scott

    Post edited by wsgentry on
Sign In or Register to comment.