Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
Just ask. I can make a trial version that does a limited number of frames pretty quickly.
I have just installed your plugin MDD in c8, but I do not find any trace of it in Carrara.
Where are the options of import/export ?
It is a file exporter. Choose the object to export, click File | Export..., then choose "Lightwave MDD (*.mdd)" from the "Save as type..." drop-down box.
OK, I found it.
For the importation in Marvelous Designer, the animation is OK.
When I import the dress animation.MDD (Standard or Maya/3DS) in Carrara, I don't manage to adapt clothing on the character (V3).
My tweaners are however on “Linear”…
I use preset (M.D.) for the import and export, I don't understand well how to do that.
What's the exact process ?
Thanks.
I use the inches scaling preset in Marvelous Designer on both import and export, which gives consistent scaling between the two apps. I also check the three option buttons when exporting from Marvelous Designer, I can't offhand remember what they are but select them all. You are correct to use the linear tweeners, others can cause issues. On import of the clothing back into Carrara, disable the Auto-position and check “Disable Auto-Scaling” then it should load exactly in the right position to match the animated character which you exported. You shouldn't have to do anything else, it should be in position and correctly scaled. You mentioned "Standard or Maya/3DS" - this isn't something that I use.
Ahhh, okay...here's the correct thread....
Anyway, as I was saying... so I’m curious…what happened to the interest in this? Seems like a very nice and innovative solution to the current limitations with Carrara cloth.
What I’m doing is exporting the character mesh (say, V4) as an .obj, then breaking it up into sections that correspond to the arm bones, leg bones, etc., and attach them via softbody attach to the corresponding character bones.
Does anyone have any results so far, or has all interest dwindled? :) :)
I started working on a tutorial outlining what seems to be the best approach so far but free time has been my enemy recently. Hoping I can finish it this weekend. I've been able to get usable results on long sleeve long skirt meshes.
Y'know, it just occurred to me...
Instead of making separate proxy objects that are softbody attached to the character, why not forget all of that and make the character itself one big softbody object that is softbody attached to itself or something? That would be ideal if it worked.
Since cloth doesn't collide with moving, non physics meshes, then make the moving mesh a physics object...
No idea if that would work though...but since we learned that rigged objects can also be cloth objects, then maybe there's something there we can use....hmmm.....
I was thinking the same thing Joe, though I have had zero chance to test this idea. I'm assuming others must have tried before me and there must be something wrong with this approach.
I've been using a low-poly proxy figure for my dynamic hair simulations with conforms to V4 like clothing, which is how I'm getting super fast hair simulations that give great realistic hair movement, and I couldn't help wondering, what if I made the V4 proxy a soft body (I have collisions turned off for the V4 itself), wouldn't then I have a low-poly mesh (which makes for faster calculations anyway that both the hair could grow out of, collide with, and I could put soft-body modified clothing on that would also collide and behave dynamically the way it should?
I figure there must be something obvious that's wrong with this approach, since it seems so simple, but have had no time to test it yet. If this won't work then I'll probably adopt your excellent prior approach of taking a character mesh, breaking it into sections and using the pieces as softbody 'armor' on the character to collide with the cloth.
I know from your prior experiments that you've mentioned that boned/rigged conforming clothing can be made into softbody same as anything else. Does it retain the bones and rigging though? If it doesn't, then it may be why this 'universal body' approach might not work, as how would it know how to stay with the character as it moves... I'm probably not thinking this through clearly though, need to do some real tests...
This was one of the things I tried. You loose the ability to animate or morph the character. It basically becomes a Macy's parade balloon!
Just tried a very simple test with a vertex cube that I extruded a little, then added some bones to it for some rigging, then after that added a soft body attach. The bones are still there but basically do nothing after the soft body has been attached, so in essence you lose all rigging once you've made it soft body. Also it seems to freeze the primitive in place and I can't move it using translate or rotate or anything, it only reacts to running the physics simulation (drops to the ground and shivers, etc).
I'm guessing this is why we probably can't do an all-over body solution for this (sadly), and need to drag and drop various sections of the body on the relevant character bones so that the soft body 'over-armor' covering will move with the character in animation. Maybe I'm doing something wrong though...
Even if the entire mesh is "attached" to the rigging ? I'm thinking select ALL of the mesh and attached it to the rigging tree? I *think* that all mesh points that are under SB attach influence still are influenced by bones, aren't they? Maybe not...
Okay, I did a very quick simulation.
The cylinder is a vertex mesh object, and is rotating under the influence of a single bone. The cylinder also has both SoftBody and Softbody Attach applied. For the Attach, I have not only selected ALL the vertices to be under the SB attach influence, I also selected the bone as the "tree" of influence.
As you can see, the cylinder maintains it's shape and rotates with the bone during the simulation even though it's under the influence of not only a bone, but also softbody and attach. Which is good.
However, as you can see, the collision isn't quite what we want. It seems to collide to some degree, but even with the Simulation Accuracy increased the cloth falls thru.
Now, if we can just figure out why the collision isn't faring too well we might have something..
http://youtu.be/u8Wa0LZKFJk
Have only had time to test on an infrequent basis, but have not dropped it by any means. For my own convenience I made myself low polygon envelope meshes for V4, M4, Genesis, and the Genesis 2 figures. These are perfect for slicing and attaching bodypart envelopes to the bones. What I am finding is that the whole part rotates with the bone with no blending at the joints. This is not optimal. It is still getting OK results, but I would like to test a few more ideas related to the joints before I post any screen shots or anything.
EDIT: think we were typing at the same time, J2K. Thanks for posting the test.
Hey, I GOT IT !! You need to crank up the collision distance to some high value like 40% and it works fine !!!
Maybe there's something different about cloth vs. SB attach collisions that needs a high distance. Normally I put it around 10%, but not for this....
COOL !!!
So it seems the concept, at least, of a boned and SB attached mesh colliding with a separate cloth object while maintaining its shape and operating under the bone influence is do-able.
Now the question is, can it be done with a character mesh like V4.....
Here's the link: http://youtu.be/BCobh5RDdvU
Looks great J2K, but there is still the problem of multiple bodyparts. In your test, you only have one cylinder moving in one direction. There is still the problem of joints where bodyparts rotate in different directions so the envelope meshes get separated if there is no blending.
But great progress. Kudos to you. One step at a time.
I'm not sure I follow, diomede....
We're talking about a single mesh V4, and whether we can designate that as a softbody and softbody attach mesh so it will collide, while moving under animated bone influence, with a separate cloth object.
Are you thinking we're talking about the "proxy parts" scenario?
It looks fantastic Joe, I think you're onto something here. I can't understand why my own test trying nearly the same thing failed though. As soon as I put the softbody on it, the whole object froze in place, bones would move and rotate but had no effect on the mesh any longer, and when I did the physics all it would do is fall to the floor and quiver a bit.
Wait, I just read that you had softbody attach also applied to your cylinder, maybe that's the difference...
Yeah, that's the key. Softbody attach kinda converts it from a softbody object to a rigid object that can be under the influence of bones, even though the remainder of the object is indeed a sofbody mesh. That's why you can "attach" a cape to a rigged object and have it move under the influence of the bone motion, while the majority of the cape mesh is dynamic cloth.
The cylinder doesn't appear to be rigged with bones moving in different directions so I'm missing how it is a test for a rigged figure like V4 or M4. I think I'm the person who came in the middle of a conversation. Looks great, though.
Yeah, that's a good point. Next step I suppose is seeing if you can designate the entire rigging "tree" as a softbody attach for the entire mesh...or something like that... :) :)
Definitely worth further tests along those lines, but I think Stringtheory suggested some potential problems. We will get there if we each keep experimenting and sharing the results. I will try to post results of an experiment I have in mind if I get a chance later this weekend.
This was one of the things I tried. You loose the ability to animate or morph the character. It basically becomes a Macy's parade balloon!
My biggest concern is whether the Bullet simulator will choke on the polygons in the V4 mesh. It definitely chokes on many of the conforming clothing meshes I've tried to cloth-ify....maybe I'll have to apply subdivision or something...
I'm thinking I'll just apply SB attach to the V4 mesh, then select all vertices in the mesh as under the attach influence, and forget about selecting a "tree".
We'll see...
Well, it looks like as soon as you apply softbody to the V4 mesh you get the wonderful Carrara error message "An error has occurred while executing" when you start the sim. I even tried subdivision, but no joy.
Maybe the V4 mesh just isn't cut out to be part of a Bullet sim as a softbody object.
Maybe I'll take the mesh into another app and analyze and see what the problem might be....
The V4 mesh has over 64,000 polygons so maybe that has something to do with it. One of the benefits of the envelopes was that they could be lower resolution and simplify the calculations. Maybe you could export the mesh and then reduce the resolution and bring it back in and then attach the reduced resolution mesh to the skeleton?
Just brainstorming.
Now that I'm trying using soft attach, I can definitely get the bone to stay in place and it seems to work great for a single bone, but with multiple bones it only seems to be affecting the bone at the 'root' and the other bones no longer seem to have any effect on the mesh... is it something I'm doing incorrectly?
In order to get soft bodies moving in different directions, I've attached separate objects to separate bones. Unfortunately, they separate at the joints because they don't blend due to weighting.
If nothing else, at least we've come to the determination that you can in fact have bullet cloth collide with a moving, rigged object, provided that both objects are acceptable to the Bullet simulator. And, of course, you have the appropriate settings. Right now that excludes a V4, but anything else with the correct mesh should work.
It would be interesting to see how the newer figures respond. I don't have them, so I'll leave it to someone else to try that.
I'm not sure I followed that, you mean you had entirely new mesh for each bone? That would be pretty equivalent to the approach of doing a different softbody overcovering dropped on each of V4's bone then, right?
Or is there a way to have multiple bones within the same object still able to deform that object, at the same time it's a soft body object?
Yes, because I don't think J2K's tests have been with boned or rigged cylinders, unless I missed something. I have had good results for single moving objects like the cylinder (or a single object attached to a single bone), but as far as I know we still have the problem that a rigged figure will have its parts move in different directions, which I don't think the cylinder tests simulate.
I could have missed something, and everything is worth continuing to experiment with further adjustments.