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

1246739

Comments

  • I thought I had posted this before, but just wanted to say thank you so much for putting this together. The examples of the different settings with images of the end results in particular are extremely helpful to troubleshoot when a simulation looks "wrong" and how to fix it.

  • CriosCrios Posts: 2,629

    Thank you a lot for the PDF, is really usefull.

  • Thank you so much! I haven't had the time to experiment that much with Dforce yet, just a couple of explosions and one kind of ok drape, so this PDF is gonna make it so much easier! Thank you all for sharing your knowledge!

  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    13. dForce Surface Parameter Tests. 

    To test out the influence of various surface parameters, I've created a test rig of five towels having the same influence weight map as I used in 11 and 12 above. This map has two small areas where the clothes pin hold the towel. Note that once you have the influence weight map, you no longer actually need any pins or clothesline, the towel will be held in mid-space as these upcoming images show. Default values are used unless noted in the description.

    a. Dynamic Strength. After Friction, this is the next surface parameter in the Simulation tab on the Surfaces pane. This parameter is meant to be applied to accessories like belts, buckles, buttons, etc. to keep them from distorting. A value of 1 allows the vertices to move freely and a value of 0 stops all movement. Intermediate values suppress movement but I have found it acts quickly. The image below shows 5 towels. From left to right:

    1. 1.0 (default)
    2. 0.96
    3. 0.92
    4. 0.88
    5. 0.84

    As you can see, by 0.88, most of the draping has been suppressed. At 0.8 (not shown) it all is.

    The above image doesn't show much since it was draped vertically with gravity alone. The next image has wind applied. As you can see, there's a lot more variation on towels 4 and 5 then what shows with the vertical draping. Since the folds that form in towel 1-3 are mostly suppressed, you wouldn't want to use these levels of dynamic strength for a towel, but you may want to for a piece of paper that was crumpled and then unfolded. If you take the Dynamic Strength down to 0.25 (not shown), the towel ends up flat and smooth.

     

    b. Stretch Stiffness. The higher the value, the less a mesh will stretch. In the following image, from left to right, values are:

    1. 1.0                      less stretch
    2. 0.8 (default)
    3. 0.6
    4. 0.4
    5. 0.2                      more stretch

    Without wind, the biggest changes are found in the lower corners, where there are fewer vertical folds, and along the lower center where there are more but less deep folds.

    With wind, most differences are gone although a little less wavy in the lower corners.

    The changes are subtle which is true for a number of these comparisons. 

    dForce dynamic strength by -4.jpg
    800 x 320 - 131K
    dForce dynamic strength by -4 wind.jpg
    800 x 320 - 143K
    dForce stretch stiffness by -0.2.jpg
    800 x 320 - 130K
    dForce stretch stiffness by -0.2 wind.jpg
    800 x 320 - 137K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    c. Shear Stiffness. Controls the amount of resistance to lateral shifting and twisting. The higher the value, the less a mesh will stretch. In the following image, from left to right, values are:

    1. 0.2 (default)                 
    2. 0.4
    3. 0.6
    4. 0.8
    5. 1.0                                    

    Without wind, you can see how the center drapes dramatically at the default value (left towel) and disappears as you go to the top value of 1.0 (right).

    With wind, you see a similar pattern.

     

    d. Bend Stiffness. Higher values will result in broader folds and shapes and lower values approximate thin and flexible materials. In the following image, from left to right, values are:

    1. 0.1                    thin material, fewer folds
    2. 0.3
    3. 0.5 (default) 
    4. 0.7
    5. 0.9                     thicker material, more folds

    Without wind, you can see the enhanced folding at higher values (right). At lower values, the center section becomes smooth and seems to be draping more.

    With wind, there’s a surface disturbance that develops and some of the folding is reduced.

    dForce shear stiffness by +0.2.jpg
    800 x 320 - 120K
    dForce shear stiffness by +0.2 wind.jpg
    800 x 320 - 130K
    dForce bend stiffness by +- 0.2.jpg
    800 x 320 - 129K
    dForce bend stiffness by +- 0.2 wind.jpg
    800 x 320 - 135K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    e. Buckling Stiffness. Controls the amount of resistance to compression. In the following image, from left to right, values are:

    1. 5% (default)                 
    2. 25%
    3. 50%
    4. 75%
    5. 95%                                  

    Without wind, it’s subtle but you’ll see the centers of the towels have somewhat more draping at higher values.

    With wind, it’s again subtle but you’ll see draping starts higher up the towel as the value increases.

     

    f. Buckling Ratio. Higher values should behave more like flexible silk while lower values behave more like denim. In the following image, from left to right, values are:

    1. 10%                        denim
    2. 30%
    3. 50%
    4. 70% (default)
    5. 90%                        silk

    Without wind, not a lot of variation…

    …nor is there with wind.

    dForce buck stiffness by +- 25.jpg
    800 x 320 - 134K
    dForce buck stiffness by +- 25 wind.jpg
    800 x 320 - 138K
    dForce buck ratio by +- 25.jpg
    800 x 320 - 130K
    dForce buck ratio by +- 25 wind.jpg
    800 x 320 - 136K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    g. Density (GSM). Lower values represent lighter fabrics while higher values represent heavy fabrics. In the following image, from left to right, values are:

    1. 90                          lace
    2. 180 (default)
    3. 270   
    4. 360
    5. 450                        canvas

    As with Buckling Ratio, without wind there’s not much variation.

    With wind, you see folds forming somewhat closer to the center.

     

    h. Contraction-Expansion Ratio.  Values below 1 will cause the mesh to contract/shrink while values above 1.0 will cause the mesh to expand/grow. If it grows too much, it can interact with itself or other objects and crash. In the following image, from left to right, values are:

    1. 80%                       shrink
    2. 90%
    3. 100% (default)
    4. 110%
    5. 120%                      expand

    I used more extreme values to show the effect. Without wind, you can see how the left towel shrinks in on itself while the right towel has expanded. The middle towel is the default size.

    With wind, some of the folds are being blown smooth.

     

    dForce density by 90.jpg
    800 x 320 - 132K
    dForce density by 90 wind.jpg
    800 x 320 - 137K
    dForce contr-exp by 10.jpg
    800 x 320 - 134K
    dForce contr-exp by 10 wind.jpg
    800 x 320 - 135K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    i. Combined Effects. The above image series were single-variable tests. Obviously you can combine these parameters in many, many different ways. I went through and chose settings that tended to cause more folds in the towel. The image on the left shows the default dForce surface values while the one on the right shows the combination (note: in the PDF I have the image descriptors flipped in this paragraph, the words here and in the table are the correct order). You can see that the settings I chose did indeed enhance the amount of folding. The settings used are:

     

    Left Towel

    Right Towel

    Dynamic Strength

    1.0

    1.0

    Stretch Stiffness

    0.5

    0.2

    Shear Stiffness

    0.2

    0.2

    Bend Stiffness

    0.5

    0.9

    Buckling Stiffness

    5%

    95%

    Buckling Ratio

    70%

    50%

    Density

    180

    180

    Contraction-Expansion Ratio

    100%

    105%

     

     

     

    j. Note regarding these tests: The towel images above are static figures after simulation is complete. If you watch the objects while they are undergoing simulation, you can see more variation for some surface properties than in the finished results. As these are also free-hanging towels that are not interacting with other portions of clothing or other objects, they may not reflect the results you get while draping clothes on a figure.

    dForce default vs folding settings.jpg
    800 x 800 - 268K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    If you spot any errors in the PDF, please let me know either by PM or a note here. I'm sure there will be places were I mistyped or miswrote something. There are also some areas of the PDF that have revised wording versus what I posted in this thread. I do not plan to re-edit everything I posted here but will get the numbering to match at least. Use the PDF for the most up-to-date information. Glad it's being of help to so many - I'm learning a lot myself!

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

    Attached is an updated PDF. I went through and reformatted the entire document to have it read and look more consistent. I also added a hyperlink index on page 1 of the PDF so you can more quickly jump to the section you want to read or review. The topics to date are: 

    1. Creating and Draping a Blanket Over a Figure
    2. dForce Global Environment and Duration Settings (Gravity, Air Resistance, Stabilization Time)
    3. Cast Off Blanket
    4. Adding a Smoothing Modifier
    5. Using dForce Surface Smoothing Settings (Velocity Smoothing and Velocity Smoothing Iterations)
    6. Draping Clothes on Furniture
    7. Friction
    8. Creating a Clothes Pile
    9. Various Methods to Correct Poke Through (including Push Modifiers)
    10. Getting a Figure’s Head to Sink into a Pillow
    11. Hanging Towel (introduces use of dForce influence weight map)
    12. Wind Node
    13. dForce Surface Parameter Tests

    All except the last topic is posted to this thread. I'll get the rest of #13 added in the next day or two. And more topics are on the way!

    Bless you, bless you, bless you!  I know I'm not alone when I say thank you so much for this extremely helpful, easy to understand, and well written tutorial.  I appreciate it so much!

  • MelanieLMelanieL Posts: 7,070

    This is great! Thanks for the latest update and improvements.

  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    I updated the numbering and a few paragraphs and images this morning, so what's posted in this thread and in the PDF is very similar.

    I also added links to the menu on the very first post.

    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    14. Creating a Foot Stool.

    This is an example showing how you can use a primitive cube and negative gravity to create a foot stool which will react to objects and figures sitting on it.

    a. Create a primitive cube and scale it to be squat. You need enough polygons in the cube so it can distort, so set Divisions to 100 when creating the cube (Create/New Primitive). Select the cube in the Scene pane and then from the main menu select Edit/Object/Geometry/Add dForce Modifier: Dynamic Surface.

    b. Create a smaller primitive sphere and position it partially embedded in the middle of the cube. EDIT: Create a primitive plane to act as the floor. It needs to be at least as large as the cube. This acts as an anchor to hold the cube in place so it expands upward. See image below. Neither the sphere nor the plane receive a dForce modifier.

      

    c. Go to the Simulation Settings pane and set Environment/Gravity to -0.1. During simulation, this will cause the cube to expand instead of collapsing in on itself.

    d. Run the simulation and you will get something similar to this image. The negative gravity causes the fabric to push up and around the ball. It also causes wrinkling along the edges of the cube and pulls the material in from the corners. This gives the appearance of a plush foot stool being pushed down on by the ball.

    Without the ball, it looks like this:

    After rendering:

    This is a render without the ball. You can see that it’s partially puffed but without all the surface disturbance caused by the ball.

     

    e. Replace the ball with a pair of feet partially embedded into the top of the stool.

    Here’s the finished render showing a comfortable foot stool for tired feet!

     

     

    dForce ball start side.jpg
    415 x 161 - 7K
    dForce ball start.jpg
    592 x 399 - 16K
    dForce ball after.jpg
    582 x 401 - 25K
    dForce ball after ball hid.jpg
    598 x 398 - 22K
    dForce ball render.jpg
    800 x 600 - 417K
    dForce no ball render.jpg
    800 x 600 - 385K
    dForce stool feet start side.jpg
    502 x 255 - 15K
    dForce stool feet render.jpg
    800 x 600 - 374K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    f. Replace the feet with a young girl sitting on the stool. In this case, the figure is hovering over the stool before simulation, but the growth of the cube with negative gravity will push up into the figure.

     

    g. In this example, we move the girl further into the stool.  Now you see more surface interaction then above, but you also see that the extended leg does not depress the edge of the stool as you might expect; instead, it’s like the leg cuts into the edge.

    Closeup:

     

    h. I didn’t find a way to use dForce to solve the problem but I could use a D-Former to push the edge down. I used a very small weak field as you can see in the right image.

     

    Here’s the resulting render:

    In this closeup, you can now see some space between the leg and the fabric which looks more like you would expect in the real world:

     

    When you use a D-Former or Push Modifier to modify a simulation, remember that you are moving away from a physics-based outcome. So use judiciously, more to nudge things rather than make dramatic changes.

    dForce stool sitting start side.jpg
    616 x 322 - 19K
    dForce stool sitting.jpg
    800 x 600 - 391K
    dForce stool sitting start side 2.jpg
    640 x 310 - 18K
    dForce stool sitting 2 wo dform.jpg
    800 x 600 - 483K
    dForce stool sitting 2 wo dform closeup.jpg
    444 x 444 - 186K
    dForce stool dform side view.jpg
    413 x 350 - 23K
    dForce stool dform field.jpg
    460 x 439 - 36K
    dForce stool sitting 2 with dform.jpg
    800 x 600 - 370K
    dForce stool sitting 2 with dform closeup.jpg
    444 x 444 - 181K
    Post edited by RGcincy on
  • RGcincy said:
    f. In this example, we move the girl further into the stool.  Now you see more surface interaction then above, but you also see that the extended leg does not depress the edge of the stool as you might expect; instead, it’s like the leg cuts into the edge.

    Closeup:

     

    g. I didn’t find a way to use dForce to solve the problem but I could use a D-Former to push the edge down. I used a very small weak field as you can see in the right image.

    Animated drape, or drape from memorised, with the leg initially raised?

  • sapatsapat Posts: 1,735
    edited November 2017

    I'm putting the sphere on the flattened cube and running the sim, but my cube isn't doing anything.  I have a plane, then the flattened cube, and the sphere partially submerged in the cube.  Then I add a dforce dynamic modifier to the cube because even though you didn't mention it in the scenario, the sim won't run without it.  I set the gravity to -0.1 as you said.  When the sim is done it pretty much looks like it started out.  No dent and no wrinkles.  Am I doing something wrong?

    footstool before sim.JPG
    746 x 546 - 23K
    Post edited by sapat on
  • RGcincyRGcincy Posts: 2,804
    edited November 2017
    sapat said:

    I'm putting the sphere on the flattened cube and running the sim, but my cube isn't doing anything.  I have a plane, then the flattened cube, and the sphere partially submerged in the cube ... I set the gravity to -0.1 as you said.  When the sim is done it pretty much looks like it started out.  No dent and no wrinkles.  Am I doing something wrong?

    The only thing that comes to mind is how many divisions does your cube have? It should have something like 50-100 to give it enough polygons to work with. If you don't recall what it had when made, set your viewport to wire shaded to check). I'll add this to the directions.

    sapat said:

    Then I add a dforce dynamic modifier to the cube because even though you didn't mention it in the scenario, the sim won't run without it.  

    Good point, I keep forgetting to mention that step. I'll add this in too.

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

    I'm putting the sphere on the flattened cube and running the sim, but my cube isn't doing anything.  I have a plane, then the flattened cube, and the sphere partially submerged in the cube ... I set the gravity to -0.1 as you said.  When the sim is done it pretty much looks like it started out.  No dent and no wrinkles.  Am I doing something wrong?

    The only thing that comes to mind is how many divisions does your cube have? It should have something like 50-100 to give it enough polygons to work with. If you don't recall what it had when made, set your viewport to wire shaded to check).

    sapat said:

    Then I add a dforce dynamic modifier to the cube because even though you didn't mention it in the scenario, the sim won't run without it.  

    Good point, I keep forgetting to mention that step. I'll add in it.

    Yikes, my cube only had 24 divisions.  I'll really have to amp it up, thanks for the tip.  Oh, and yeah, that pesky dforce modifier, LOL!

  • RGcincyRGcincy Posts: 2,804

    Animated drape, or drape from memorised, with the leg initially raised?

    It was a static pose with the cube and the figure in the final position. Draping the figure from a memorized pose (T-pose) causes a crash as the feet catch and wrap the cube around the figure. Because of your astute question, however, I tried an animated pose with the leg initially raised and it works much better. I'll be updating the writeup to reflect animation. Thanks!

     

  • sapatsapat Posts: 1,735

    Ok, I ramped up the cube divisions to 100 and it made all the difference.  Great spot for kitty to curl up and sleep!

     

    111517-dforce used on cube to create indent.jpg
    1198 x 888 - 104K
  • RGcincyRGcincy Posts: 2,804
    edited November 2017
    sapat said:

    Great spot for kitty to curl up and sleep!

    Cute!! Glad you got it to work. I updated the write-up to cover the two points your question raised.

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

    Ok, I ramped up the cube divisions to 100 and it made all the difference.  Great spot for kitty to curl up and sleep!

     

    RGcincy said:
    sapat said:

    Great spot for kitty to curl up and sleep!

    Cute!! Glad you got it to work and I updated the write-up.

    laughThanks!

  • RGcincyRGcincy Posts: 2,804
    edited November 2017

    15. Animated Posing for the Foot Stool.

    An astute question by Richard Haseltine led me to try a better way to bend the edge of the cushion: animate the extended leg so it comes down onto the cushion rather than starting within the cushion.

    a. Set everything up the same as in section 14. Be sure Start Bones From Memorized Pose is turned OFF in the Simulations Settings.

    b. If not already open, go to Windows/Panes (Tabs) and select Timeline.

    c. Set the total number of frames to 31.

    d. On the timeline, click on frame 10 then click on the + Key icon in the lower right of the pane. This sets the pose we’ve been using as the final pose.

    e. On the timeline, click on frame 0 then click on the + Key icon again. Next select the figure’s shin and bend the leg so it just clears the edge of the cushion.

    f. In the Simulation pane, under the Duration tab, select Animated (Use Timeline Play Range) from the drop down box.

    g. I found that 10 seconds was enough time for the leg to move down and depress the edge of the cushion in a natural fashion. However, if you only have 10 frames in the timeline, that was not enough time for the cushion to fully expand so the figure did not look like it was sitting into the cushion.

    The final render:

    h. This was the first time I’ve used an animated figure pose. Besides the advantage of removing interference between the extended leg and the edge of the cube, you can use the timeline to see how the simulation is progressing and choose an intermediate time for a render.

    Here is the 10th frame:   

    15th frame:

     20th frame:

    30th (final) frame:

     

    i. There’s another way to do an animated pose (set Start Bones From Memorized Pose to ON) but in this situation it will just cause a crash. The figure will start from the T-pose, dive down into the cube as it moves into the desired pose, and wrap the cube around its feet in the process. As it wraps, all goes haywire and Daz Studio will shut down.

    [Edited to Add]image You can change the Memorized Pose to one that will work. In this case, move the extended leg as in step e. so it clears the edge of the cushion. From the main menu, select Edit/Figure/Memorize/Memorize Figure Pose. Now move the extended leg back down into the cushion where it was before. In the Simulation Settings pane set Start Bones From Memorized Pose to ON and run your simulation. I found this to be slower than setting up the animated scene but others may find this a better approach. Here's a render, which is only slightly different from the animated result:

     

    dForce timeline.jpg
    738 x 149 - 20K
    dForce stool raised leg.jpg
    508 x 233 - 15K
    dForce animated duration.jpg
    495 x 196 - 33K
    dForce stool animated.jpg
    800 x 600 - 368K
    dForce stool animated closeup.jpg
    444 x 444 - 184K
    dForce stool animated frame 10.jpg
    545 x 394 - 23K
    dForce stool animated frame 15.jpg
    568 x 388 - 23K
    dForce stool animated frame 20.jpg
    547 x 391 - 24K
    dForce stool animated frame 30.jpg
    559 x 392 - 24K
    dForce stool memorized pose.jpg
    800 x 600 - 366K
    Post edited by RGcincy on
  • RGcincy said:
    i. There’s another way to do an animated pose (set Start Bones From Memorized Pose to ON) but in this situation it will just cause a crash. The figure will start from the T-pose, dive down into the cube as it moves into the desired pose, and wrap the cube around its feet in the process. AS it wraps, all goes haywire and Daz Studio will shut down.

    You can use Edit>Figure>Memorise>Memorise Figure Pose to start from a different pose, just like starting from a non-zero pose in an animated drape.

  • RGcincyRGcincy Posts: 2,804

    You can use Edit>Figure>Memorise>Memorise Figure Pose to start from a different pose, just like starting from a non-zero pose in an animated drape.

    Good point Richard. I've added a paragraph to the post above. Thanks

  • RGcincyRGcincy Posts: 2,804
    edited October 2018

    16. Head Scarf and Telephone Wires. This shows two other objects you can make with primitives and dForce. These are pretty straightforward if you’ve performed the steps in the previous sections: position the object, apply the dForce modifier, run the simulation.

    a. The scarf in this image is a square plane positioned over the head to match the rotation angles of the head. The center point of the plane is vertically above the center point of the head so it's centered when gravity pulls it down.

     

    b. The wires in the next image were made from primitive cylinders. After applying the dForce modifier, I then added a dForce weight node and removed the influence from just the tips (paint while holding down the Alt key to remove weighting, see section 11 for details). I found a Gravity of 2.0 and a Stabilization Time of 0.3 allowed the wires to sag without the tube going flat. 

    dForce scarf placement.jpg
    377 x 283 - 43K
    dForce lace head scarf.jpg
    600 x 800 - 389K
    dForce wires.jpg
    723 x 578 - 218K
    Post edited by RGcincy on
  • RGcincyRGcincy Posts: 2,804

    Just wanted to bring another thread to your attention. It's an on-going tutorial on using dForce with clothes written by L'Adair. First topic is setting up all the material zones and weight maps for the PC+ item Medieval Cloaks. Well-written and illustrated. You can find it here: dForce, dWeight and dWardrobe: Chronicles of...

  • sapatsapat Posts: 1,735
    RGcincy said:

    16. This shows two other objects you can make with primitives and dForce. These are pretty straightforward if you’ve performed the steps in the previous sections: position the object, apply the dForce modifier, run the simulation.

    a. The scarf in this image is a square plane positioned over the head to match the rotation angles of the head. The center point of the plane is vertically above the center point of the head so it's centered when gravity pulls it down.

    b. The wires in the next image were made from primitive cylinders. After applying the dForce modifier, I then added a dForce weight node and removed the influence from just the tips (paint while holding down the Alt key to remove weighting, see section 11 for details). I found a Gravity of 2.0 and a Stabilization Time of 0.3 allowed the wires to sag without the tube going flat. 

    Wow, the possibilites are endless.  Pretty amazing stuff! 

  • RGcincyRGcincy Posts: 2,804
    sapat said:

    Wow, the possibilites are endless.  Pretty amazing stuff! 

    I agree. Another good example I've seen is a Mongolian style tent (image below) made by DaWaterRat. You can read about it here.

  • sapatsapat Posts: 1,735
    RGcincy said:
    sapat said:

    Wow, the possibilites are endless.  Pretty amazing stuff! 

    I agree. Another good example I've seen is a Mongolian style tent (image below) made by DaWaterRat. You can read about it here.

    It's inspiring to see all the creative ideas that have nothing to do with clothes!  Thanks for the link.

Sign In or Register to comment.