Quote - Okay, that's a starting point: so did I. Didn't have much success with those default settings, did you? What did you usually end up with (in terms of settings), msg?

I did ok with the default settings when using cloth for draping figures or as sheets trailing a figure.

As for clothing, I never took note of the values and didn't save any of the sims  :-(
Maybe I can  squeeze in a quick play with dynamic cloth, taking note of the values.

Have a go with the values suggested to me by KobaltKween, msg - I've had tremendous, almost uniform success with them, with pokethrough only on body bits with extreme bends...

-detailed mesh

-color,bump,displacement

feedback

Do you have a mesh (wireframe) preview you could show? Looks quite good, to my view, ice-boy.

The long lines that kind of look like looking at monitor on tv, that go across the sweater, look a bit odd. The hem on the neck looks really good, but the bottom and sleeve don't looks as good. I think if the neck hem didn't look so good, I wouldn't have noticed and issue with the sleeves or bottom. Over all I think it looks pretty good the folds are nice and not squared off.

quote]Do you have a mesh (wireframe) preview you could show? Looks quite good, to my view, ice-boy. i have a low version. and a high res version . both will be realesed.

Thank you: strong work, ice-boy. I'm no expert: just looking at a lot of mesh these days to see what people are doing.

i agree that it looks strange. but its funny that this is the exact same pattern like on my sweater. i took photos and projected them on the mesh in blender. then in photoshop i made a bump map.  

here is a close-up of the texture. 4k.

I didn't realize I could enlarge the images. I take back what I said, I think you did a fantastic job. The detail you got with the maps is wonderful.

i watched a lot of DVD's about dynamic clothing. WETA and ILM and every studio doing CGI are using dynamic clothing.  the clothing and human body is very detailed. they say that they dont use the final body(render) as collision. because the mesh is to detailed. and it true. you dont need every vein as a collision for clothes.

so here i made a very low poly M4 collision body. its meant to be used for dynamic clothing. so  set collide against  .  i think more then  70% times you can use this. but there are times when you should use M4 for collision. if your clothing is very dense and has a lot of details. if you want small foldes its maybe good to use M4 since the low body figure is not smooth enough.

-the figure is conforming. so conform to M4. that way you only pose M4

-before you render turn the low body invisible.

http://www.renderosity.com/mod/freestuff/details.php?item_id=64256

That's a brilliant way of speeding up calculation times!  How did you make the low poly figure?

blender. there is an option called shrinkwrap. you model on M4. you keep the vertices low. then you add body groups and use a blank M4 .cr2.

Ah!  Thank you.

Brilliant! Thank you for the research and the figure.

lmk

did anyone try it out?

I just did a hasty check, using a dynamic robe for M4.  It's time for bed, so my eyes are crossing, but it looks like the low-poly version took about half the time to calculate.

@ice-boy - asking for feedback

Looks promising. My note in general is that you are using quads. This will make the thing behave like rubber, leather or something else non-woven. It may cause troubles when resolutions become too low. When using tri's instead, it will behave more like linen, as a long-sleeved T-shirt. The thicker edges of a T-shirt can be created by giving it a low(er) resolution at the sleeve-, neck- and hip ends.

All the best

i now also made the same swetshirt with tri's. will ealse this week.

Robyn's corset that she made had thickness at the seams that she modelled in.  It looked great and she said it worked okay.

Love esther

The low-poly M4 proxy figure seems to be a success.  Will you be creating a proxy for V4 as well?

for 2 years i have been searching for a technique how to get similar shoulders. nothing. i always thought that i have to model the jacket with some special loops. but no.

so what did i do? i modeled 2 simple shapes for the shoulders. they look similar to the pads under the jacket. then i was thinking, ok how will i now parent this to the body. ohhhhh yeah we have conforming clothes. this is perfect.

this is how this looks. those are conforming shoulder pads. the groups are only for the M4 chest. that way they dont bend when you move M4's shoulder.

http://www.renderosity.com/mod/freestuff/index.php?user_id=497653

dynamic pants and dynamic jacket plus conforming shirt.ok so just we understand. you dont render the shoulder pads. you only use the shoulder pads when calculating the dynamic clothing. so in thecloth options youselect the shoulder pads as collision. when its finished you make them invisible. they are under the jacket and give the jacket the right shape.

very nice attempt, really. I examined my own jackets, since emulating the real thing is usually the way to go. I found that the "shape keeping" effect mostly results from

a) the very thick and non-flexible stitching of the sleeves to the jacket-body

b) a filling in the jacket-body from the top-breast (say; the breats pocket) up to slightly over the shoulder.

senior business suits have a thick filler, frivole summer jackets hardly have a filler at all. Your pads serve as those fillers, but they appear a bit (too?) stiff to me, at least in the upright standing position. The black suit on the posed guy looks pretty good, though.

I'm still under the impression (nothing more or less) that the effect also can be obtained by creating zones in the jacket with different cloth properties (high fold resistence etc), and perhaps some modelling adjustments (larger polys make stiffer cloth). Since jackets are woven cloth, I should prefer tri's over quads.

Quote - The low-poly M4 proxy figure seems to be a success.  Will you be creating a proxy for V4 as well?

I'm guessing that the answer is "no".

Nice work, IceBoy. My own efforts have been considerably less auspicious: still working with combining dynamic with constrained and decorative. It is slow going.

Quote - The low-poly M4 proxy figure seems to be a success.  Will you be creating a proxy for V4 as well? I'm guessing that the answer is "no".

If you have Wardrobe Wizard support for M4, you might be able to get something usable. I don't, or I'd give it a go. You'd probably have to remove the hands to get the conversion to work since WW doesn't do gloves.

OT: This low-poly proxy approach is also very useful for dynamic hair. 

that looks awesome. I would buy that if it was in the MP.

Love esther

Looks great, ice-boy!

for all, and especially for RobynsVeil,

just kicked off writing a serius tutorial on all this Cloth Room stuff. So if you do have particular issues to be discussed, just state them here, I'll do my best.

• on the user interface as well as the sim engine itself, I cannot change things but I do present a decent guided tour at various levels of expertise
• on setting the proper sim parameters, and avoid crashing, exploding calculation times, meshes going wild and/or collapsing results: I've got that now, see next post (as I don't let you wait for the publishing of the tut for that :) )
• on crumbling, weird folds and 'crystal ridges', got that one too, it's a mesh thing. Mesh structure in general determines the behavior of the cloth in simulation (folding, stretching, etc). In short:
   - quads make the cloth behave like non-woven (leather, rubber, fleece, ...) and tight knits, and they are vulnerable to crumbling and ridging
   - mono tris (all diagonals in the same direction) behave like woven (linen, silk, ...), are less vulnerable to crumbling but do show a-symmetric behavior
   - mixed tris (diagonals in alternating directions) are woven as well, not vulnerable to crumbling and show a neat symmetric behavior like real cloth
   - hex meshes are great for loose knits like thich winter sweaters.
• I'll do mesh details, turning conforming stuff into dynamic, the physics / cloth detail parameters, handling parts of cloth elements (grouping) and handling multi-cloth situations, saving and loading settings.
• more ... just tell me

tucking shirts into jeans thanks.

Love esther

okay, the Magic Bullet on sim parameters.

It's not a bullet but a process, and it's not magic because some parameters (Offset and Depth) have practical upper and lower limits depending on the scene and animation itself. Which ones, why the process works and more details will be in the tutorial. Anyway:

Collision Offset: start with real world values, and visual requirements on the end result. Offset creates some distance between the body and the cloth, suggesting some thickness. I would suggest: 0.25 for lingerie and fine clothing, 0.5 for sweaters, 1.0 for thick cloth over cars and statues, 2.0 for serious winter overcoats.
Changing the offset does not affect system resources (the offset iself does, but its the same for all values).

Collison Depth: start with about the Offset value.

When the sim goes wild, then check the "object vertex against cloth poly" option first. If not good enough, then check the "object polu against cloth poly" option too. This will increase calculation time. If still not good enough, double the Depth and eventually, double again. This will increase calculation time even more.

If still not good enough, start doubling Steps per frame and go on doing so till the sim runs fine. Actually this IS the magic bullet, but the price is high: each doubling also doubles calculation time. This is why it's better to try the other options first.
This setting refines the sim from 60 steps per second (30 fps * 2 steps per frame) to 1,000,000 steps per second (30 fps * 33333 steps per frame). Which takes over 16,000 as much time. Pray you don't need to go that far.

Quote - ...tucking shirts into jeans...

More generally, anything relating to simulations in which something presses against the outside of the cloth (e.g. belts). The cloth room seems biased towards collisions occurring against the "inside" of the mesh, probably with good reason. I've had only moderate success so far, and that was by continually going round this loop: simulate until it breaks; export mesh; repeat until fed up. :)

Also, your thoughts (if any) on the merits of Delaunay triangulation for cloth meshes. I've heard it's pretty good, and had read an article on Delaunay-ising an existing mesh using Sculptris, I think it was - however that was at the old DAZ forums, I believe, so no link at present. :(

@estherau:

do you mean - how to make (a still result of) a figure wearing a shirt and jeans (over the shirt) and a jacket (over the shirt and jeans) and so on, or

do mean mean - how to animate a figure tucking a shirt into the already present jeans without unbuttoning / unzipping them, like my mom always told me not to do. That will create a body - shirt - fingers - shirt - jeans stack at the waist.

the former.

Love esther

Whilst I applaud you, aRtBee, on your steadfast and faithful attention to these questions that I addressed at the outset of this thread, I've taken a bit of a wander into non-poser terrain and am exploring Cycles materials for Blender atm. Poser 2012 remains unopened for the past weeks as I strive to nut out this incredible material set still under development. Mind you, what's available now is really quite striking and makes me hope for new toys (materials based on an advanced - non-Reyes - renderer) for Poser's future.

@EnglishBob

as far as the belt-thing goes:

1. the collision offset should be less than half the (least) distance between the belt and the body

2. if the belt is two-sided from a mesh perspective (so it had vertices, edges, polys at the inner as well as the outer side) do try to reduce the collision depth to (preferably half) the thickness of the belt - then the cloth cannot 'see'the outside. A trick can be to split the belt into an inner and an out group, and make the cloth collide to the inner group only.

3. follow the reipe I just posted. I guess you will need quite a large Steps per frame setting (in a recent test i needed about 3000 to prevent a lowres cloth to fall through a hires box with a thickness of 0,005% = 0,013mm).

4. to my understanding there is no algorithmic preference between in and out. But... belt, cloth and body might have very different mesh densities. Most cloth items I see passing by have a vertex distance of say 2 to 5 cm (1-2 inch), while Vicky has an average vertex distance of say 0.5cm (1/6th inch) only. Cloth agains body means: medium / lowres mesh against very hires mesh. Belt against dress means: lowres against medium. People tend to forget the effects and differences of that.

as far as Delauny goes: the math theory on advanced cloth sims state that the most regular meshes are the least fit for purpose in this area, so that's promising for Delauny 9and bad news for quads). For further reading into 3D cloth sim, I can recommend http://www.artbeeweb.nl/researchpaper_manasilib.pdf (10Mb).

have fun.

Thanks, for that and for your previous "recipe" post which came in as I was typing. 

Quote - I guess you will need quite a large Steps per frame setting (in a recent test i needed about 3000...

Eek. I may have to apologise to my computer in advance. :-) 

you might be surprised. Without anomalities the sim calculates pretty fast. It's handling the cloth going wild that takes the time. And... just start at 2, than go on doubling.

And just give that PC something to do you have paid it for. Its not a fax... Sleepy times are over :-)

@EnglishBob

As i don't have that much issues with dresses and belts, I started an extreme test case. I'll make a full and illustrated description later, but just for short:

• PPro2012, using Andy2 and the torus prop for belt, positioned at the waist level, shrinking from 120% to 35% (frame 1-30) and hold so (frame 31-60). Since Andy2 hardly has a waist, any dress has to make sharp bends over the chest, between the waist and the belt, over the hips, down.
• First I took clothified the capsule prop for a dress, which was a disaster at all settings. The simple reason: mesh density aka poly-size. Since polys cannot make sharp bends arond an object themselves (you need more polys, and they can make angles relative to each other. Strong bends within a single poly around an object will be interpreted as poke-throuugh) this 'dress'' could not support the required deformation.
  Note that Andy2 and the belt in minimal position have a vertex-distance of about 1 cm (twice that of Vicky), the capsule has about 15cm per vertex.
• So i gave Andy2 a 'normal' evening gown (vertex distance 1 cm as well), and used some scale-dials to establish a fit since Andy2 has quite a wide chest. Offset 0.25, Depth 0.5 and please note that EACH collision object has its own settings, so these had to be set for Andy2, the belt and the ground separately. To support the dress for making the sharp bends, I reduced the folding etc and eliminated other frictions and damping. As the dress itself came in pieces, these had to be set for each piece separately.
• first run at say 2.5 sec/frame, gave a bit of poke-through on the body parts.
• checking "object vertex against cloth poly" solved this, it doubled the calculation time (5 sec/frame) but I was not very happy with the belt area, crumbled and rough.
• checking "object poly against cloth poly" resolved that to a great extend, and quadrupled calculation time (20 sec/frame). The remaining crumbling around the belt is due to the dress-mesh itself: quads again. Actually, we're asking the dress to do the impossible. Quads behave like leather, and I don't know real world leather dresses that can make such sharp bends and folds as needed this time around Andy2's waist.
• just for the sake of it, I'm running now with the cloth-self-collision checked, and the steps doubled to 4. Will take to noon tomorrow, Poser reports 1200 sec/frame. Good I did not include self-collision in the recipe :). Perhaps crumbling can be resolved in rendering, by increasing the crease angle for the dress.

please comment.

Perhaps I should apologise to your computer. :) And thank you for going to all this trouble, of course!

I have something unrelated running at the moment, but I'll retry my belt versus tabard simulation again, taking the things you've said into account. That project was some time ago, and certainly wasn't run anything like your suggestions. I have an idea for making the belt to use during simulation using Cage's Mr. Skinner script, so it should end its "shrink" a consistent distance from the figure surface; and I can also define it to be higher poly, which sounds like the right thing to do.

I should add that I'm using Poser 7, so I hope there haven't been considerable improvements in the cloth simulation. Come to think of it, the original tabard simulation would have been run in Poser 6. 

my PC should not complain, just a few hours on a single thread is nothing compared to a 48 hour render at 100% CPU load.

To avoid flooding on this forum thread I put things in a preliminary Case Study (pdf attached, loose the txt). Things work out well, with some lessons and side nodes. Saves you the wait till the final thing.

all the best

Anyway, here's a quick summary of the points. Some of these are repeated from the advice previously given by aRtBee and others. Apologies if it seems that I'm trying to take credit for that advice.

• Make the belt used in the simulation follow the figure's surface as consistently as possible. This is because the cloth will be squashed between belt and figure as the simulation proceeds, and the simulator doesn't like it if that's overdone. I used Cage's Mr. Skinner script to make the belt, with an offset of 0.001 or 0.0025 (see later discussion). Bear in mind that you can change the sim-belt for something more detailed afterwards.
• Have lots of triangular polys in the cloth. I triangulated my original quad mesh using UVMapper tools > facets > triangulate (at polygon centre) to give 11952 polys.
• In simulation settings, check both 'Object vertex against cloth polygon'
  and 'Object polygon against cloth polygon' options.
• For both figure and belt: set collision offset to minimum (0.1)
  and collision depth to 0.001.
• Increase the steps per frame as required - for this picture I went to 16 (giving ~16s per frame for 30 frames). However I wasn't wholly scientific about it, and changed more than one parameter at once. I need to try reducing this figure again to see what happens.

The render attached was made after a simulation with the belt offset being 0.001. You can see Antonia poking through the cloth a little, presumably as a result of having such a close approach of the belt to the body.

I then increased the belt offset to 0.0025, on the assumption that it ought really to be twice the collision depth. However the cloth then started poking through the belt, for some reason. As I said, inconsistent. I have to put this aside for now, but I'll carry on when I'm able.