It's worse than that. Again, because I have a scenario specifically designed to reproduce the quad problems, here I can show you them at will although you may not see them yourself in real use.

Here I'm using twice the density, 2m[20mm]Quad. Instead of reducing the crystal ridge, it's still there, but now you get the second artifact - the spiral ridge! (Look in the lower left quarter.)

You can subdivide all you want and these crystals still form.

do you have M4? you can try out my t-shirt for M4. you will get some crystal ridges.

http://www.renderosity.com/mod/freestuff/details.php?item_id=60835

Yes! Behind the left shoulder - perfect example.

These ridges, which render with a dimpled look (what I was calling the concave shading...didn't know what else to call it..lol) I've always gotten with quads. With X tris, it was slightly different, but still about the same amount. With the straight diagonals I got fewer yet, and with the random tris I get almost none.

Laurie

Looking back to the linked tutorial on cloth parameters, it says:

Quote - Shear resistance has a range of 1 to 1000; it defines how much the cloth tends to retain its shape as it folds. This differs from fold resistance; the shear resistance of a fabric determines how much or little it deforms when actually folding.

This means absolutely nothing to me as far as making predictions and understanding what it does.

I think or thought I know what shear could mean. Consider a 4-point square. Grab two opposite corners and squeeze them closer together, while holding the four side lengths constant. You get a diamond. It is still coplanar, so this is movement within the original plane. That's what I thought shear was. And I thought it did not apply to triangles, since there is no coplanar motion possible with the side lengths held constant in a triangle.

Stretch, I assume (wrongly?), is a variation in the length of edges.

Folding is a changing angle between two adjacent polygons. Maybe.

The collapse of a quad - is that fold or shear or a combination or neither? The crystal ridge suggests it isn't either one, since I was unable to affect crystal ridge with any parameter, other than creating a cloth that stretches however much it needs to in order to avoid a fold. Any quads with a diagonal fold produces crystal ridge.

Now I have another thing happening I'm curious about - the mysterious crossing of a fold through itself. I see it all the time, even with hex tris. Without making any conclusions, I just want to share that I made it stop by altering shear resistance to a low value!!! What?!?! Impossible!

So - this tells me we need to understand shear or we're not going to get anywhere. I reject anything I thought I knew about shearing motion. Starting from first principles, how do we figure this out?

I wish stewer would just come in here and give some direction. I don't expect he knows everything about the cloth simulator, but a few clues as to what to look into would be so helpful.

hi all,

great question you just asked BB, as was about to tell what I stumbled into today. I did not Poser anything, but jumped a bit around on the net.

As a result, I'm under the impression that the Cloth Room is a direct derivative of the ClothReyes plugin for 3DSMAX, which was in use in the 1998 - 2002 period. And I do have some more, similar impressions as well.

No formal proof in my hands, but there seems to be something in the fine print on the box of Poser 5 as well, so if someone has that one still around ...

This means that when we can dig up that plugin material / documentation from a 3DSMAX user, or take is for directions when looking further on the net, we might get somewhere.

Full story and references in the next post, but now you know.

we need a cloth room rewritte or a new free plug in .

and since thats unlikely, how about we concentrate on what we DO have ice-boy.

you constantly remind me of the saying "if we had eggs, we could have eggs and bacon. if we had bacon".

MagnusGreel i understand why you are angry at me. if you have a problem with me you can ignore me

thanks. **
**

since I can't ignore you, there's no ignore function, why don't just stop posting things like "if we had moonbeams we could..." or "this program does it this way, why doesn't poser" or "we need function" over and over and over and over.. and post things like your M4 shirt that are constructive?

I don't think we need to go as far as rewriting the cloth room. All we need is a little understanding of the cloth room. For a $250 dollar program, that we have dynamic cloth at all is hard to believe. If it's not what you want, there are other programs out there that do dynamic cloth, but we aren't talking about them here because we're still fleshing out Poser's cloth room :o).

I'm sorry I haven't posted any pictures of my results yet - I've gotten sidetracked by different things. It's true tho, that anything I post of the dress will not really help with real understanding of the cloth room because of all the variables that exists with a whole clothing simulation. bagginsbill and artbee and the others working on the flat planes are doing well in that area :o).

Laurie

Aha!

Maybe it's a mistake to assume data on other cloth simulators are the same as Poser's, but this is pretty informative!!!!

http://softimage.wiki.softimage.com/xsidocs/cloth_ControllingtheStretchwithSpringNets.htm

The shear spring does what I expected for a quad mesh in SI.

But how does a triangle perform shear? The shear springs connect non-adjacent vertices of the same polygon. In a triangle, there are no such things. And yet altering the Poser Shear Resistance parameter changes the outcome. According to this definition it should have no part in the configuration of a triangular mesh.

That page also says

Quote - For polygon mesh objects, the polygons are not always quad, so the shear net is not defined. What you can do to solve this is to triangulate the polygon by selecting Triangulate Polygons on the Cloth property page. The edges coming from the triangulation are then used as the shear net.

It didn't say so, but it seems to be referring to polygons with more than 4 vertices - not triangles.

What is the shear net for a triangular mesh!?

Full story (no pictures).

After my first keywords into Google, I stumbled into http://www.animares.com/book/Secondary-Motion/Cloth/Hintergrund.html

Its in German, on issues with cloth simulation. His first note is that although regular meshes do fine in most areas of 3D, cloth simulation is not one of them. It benefits from a "Delaunay-Method" which gives irregular triangulation of meshes. It's implemented in Sim Cloth / Clothmesh, as Cloth Reyes / Hexamesh, as ClothFX / Cloth Panel and Maya / Cloth.

His second note, on which I wont elaborate much further, is to watch out for parameter settings and conversion. In short, if the parameters are determined for seconds, meters, kilograms etc and the mesh is defined in inches or internal (poser) units, then you're off. I'll let this alone, the Poser parameters are in metrics (m,s,kg) as I recall to have read somewhere in the documentation, and I already posted on the gravity matching the Earth surface conditions.

Cloth Reyes got my attention, since "Reyes" was referred to related with Firefly. So I found two interesting timelines. So; watch the dates.

Timeline One, on 3DS MAX

mid '90's, the Spanish REM Infografica launched for working on fluid dynamics. In 1998 the company split, the fluid people can be found later with RealFlow and the like. (interview http://www.cgarena.com/archives/interviews/3daliens/interview.html ). About that time, the remaining and renewed Reyes Infographica started to do plugins for 3DS MAX 2 (wikipedia + myself; I was into MAX 2 those days, my avatar was done then). Also see http://forums.cgsociety.org/archive/index.php/t-725233.html.

For my story, the relevant ones are; Cloth Reyes with the HexaMesh routines, and NPR:Reyes, the renderer. NPR stands for: Non Photoreal Rendering (as in Toon and Sketch), and Reyes is a routine to make photoreal renderers with a minimum of time consuming raytracing (wikipedia).
From BB we've learned that the Reyes routine is in Firefly.
I guess the company took the algorithms name as part of there own: they incorporated the routines into their already existing NPR, then made products with Reyes in the name and rebranded their company. So: Reyes Inforgrafica did not create the Reyes algorithm but they incorporated the routine in their rendering product. Just to be clear.

mid 1999, release of 3DS MAX 3 by Kinetix. The merger into Discreet started then, mid 2000 we get 3DS MAX 4 by Discreet. (wikipedia). The Reyes' plugins were available till 4.2, I could find a date but the forumthread on installing patch 4.2.5/4.2.6 was on jan 19, 2002. (both: CGSociety forum,  http://forums.cgsociety.org/archive/index.php/t-725233.html resp http://forums.cgsociety.org/archive/index.php/t-1006.html)

It's interesting to note that 3DS MAX 5 had a completely renewed set of... rendering features.  So something happened between Discreet and the Spanish boys.

Timeline Two, on Poser

mid 1999, Poser 4 is launched by MetaCreations, but within a few months the 4.0.3 patch is issues by Curious Labs, the Poser people from MC which is splitting up. P4 introduces conformed clothing and magnets, by the way.
It then takes a few years, till mid 2002, before Poser 5 sees the light. A disastrous release full of errors, well described on JCH Digital Designs  in http://www.cocs.com/poser/poser5mess.htm, but that aside. 

Timelines meet

The interesting part is that P5 is completely renewed, including new functionalities like: Dynamics for Cloth and Hair including Wind effects, and also: a Reyes based Firefly, supported by new feautures for camera, light and materials, plus typical NPR features like Toon and Sketch rendering.

So... where did all that come from?

A few years later, Reyes Infografica pops up again, with a product called Poser Fashion (is has a Pro too, and made it to version 1.5). You can import Michael, Vicky and the whole lot, and make advanced dynamic clothing with it. Nowadays, the poser-fashion site has gone. (I closed the links on this, lets try to find them back)

So... why are those Spanish 3DS MAX guys into Poser?

You may draw your own conclusions, but for me there is a link between those timelines. So, ClothReyes might be a direction to look into.
Also interesting then is the association between ClothReyes and HexaMesh. I did not into the latter yet (it did nring up links too CAD engineering), but given our quest into tris and quads, hexagons shine a new light and will definitely fit into the opening remark, that too regular meshes don't do well in cloth sims. 

So far my share of today. Have a good weekend.

Cool.

I looked into cloth sim of 3DS Max. All examples seem to be Delauney triangulation. Cool.

3DS documentation explains

Quote - **Shear—**Resistance to shearing. Higher values result in stiffer cloth fabrics. Shear defines how much the individual triangles can deform. If you were to lay the edges of the triangle out in a strait line this value would represent how long this line can stretch out to. With a high value this length will only be the sum of the length of all of the sides at rest. A low value will allow this length to be greater then that off all of its sides at rest. This length of stretched sides is not on a one to one basis. One side of the polygon may stretch more then another

So - this is what I wanted to know. It is the resistance to changes of the PERIMETER of a triangle. This is not at all an analog of the quad shearing effect - in quad shearing, the perimeter is unchanged BY DEFINITION. Here the change in perimeter is the definition. Very odd that they would use the same word to mean exactly opposite concepts.

Just for those that tend to get lost at the vertex level.

Just lay a piece of cloth flat on the table. Pick up one corner, and the cloth raises "out of its plane". That's bend, or fold. Thick leather has a high fold resistance.

Lay it back, fixate it on the table with one hand, put the other hand in front and push or pull while the hands stay in one line. That's stretch, deformation in the direction of a straight force. More precise: along or perpendicular to the threads in a woven cloth. Some cloth sim software recognizes U and V stretch, with different values.  

Back again, fixate on the table, hands in front of each other, now you move one hand aside. More precise: along a diagonal of the threads in a woven cloth. That's shear. Leather is non-woven, it's grown, shear and stretch are similar. Rubber is non-woven too but the very lengthy molecules allow for some freedom, apparently.

Chain mail for instance is interesting. The iron rings it's made of allow almost no stretching, but a lot of shear.

For woven cloth, the issue is: do I pull on the threads it's made of (stretch), or do i require the threads to move and bend aside (vertical thread moves horizontaly or v.v) That's shear.

to BB: the poserfashion values were not for Poser but for Poser, ClothReyes (!) and Stitcher.  Maya Cloth  has different values but recognizes Thickness. 

= = = ALERT

found a mishap in my previous posts.

From the Poser manual:

The Cloth Density parameter specifies the mass-per-unit area density of the cloth in grams per square centimeter. If you take a 1cm x 1cm piece of cloth and weigh it, the weight in grams is what this parameter should be set to.

So, the default 0.005 means 50 grams per m2, and density 1 means 10 kg/m2. For lead (over 10.000 kg/m3) this implies a 1mm sheet, and not a 0,1mm sheet as I wrote before. I'm learning. And that German guy said: watch out for units. We have to sort all parameters on this.

So, Poser default density reads 0,005 = 50 gr/m2. It's understood to be table cloth. Maya default is about similar to their denim / cotton, and reads 0.01 density at 0.2 thickness. Thats 0.05 density at 1 thickness. If this is the same type of cloth as Poser, than Maya does kg/m2 for density. Metrics, at last. Maya thickness then, is in mm I guess, makes sense too. kg - gram, m - mm. Steps of 1000. Feels good.

= = = MORE UNITS and thoughts

According to the P8 manual (and the dials), Collision offset is in cm (from 0.1 to 10), and I guess so is Collision depth (from 0.0 to 10).

Collision depth works inwards (P8 manual), and determines the distance between the zero-thickness cloth mesh and the figure (to be tested). Everything below this value is considered a collision (to me). Everything above it gives room to move towards the figure, in order to fit. Default value is 1 (cm). This means that dynamic clothes might have to be a bit wider than conforming ones !?

Collision offset works outwards. For cloth which is outside the body, at depth distance or more, this is a quite useless concept (to me). So - I guess - actually it's some sort of calculation helper, to avoid that the sim has to search the entire scene to find its collision targets. Its meant to be a massive calculation speedup. This might imply that when, while moving from frame 15 to 16, the figure moves so much that the cloth - before calculating frame 16 - ends up more into the figure than this offset, the sim might get hard times to move that specific vertex outside the figure. Then we've got pokethrough due to too fast figure moves. Wild moves might require large offset values. To be investigated upon. 

Unless someone can explain to me why a cloth sim routine needs an inoutward thickness next to an inward one, while nature needs one thickness value only. Is there something wrong in my thinking (again)?

got it (and fixed another bug in my post):

= = =

Scotts Valley, June 21, 2006 - e frontier announced today the release of Virtual Fashion Basic for Poser for Windows®. 

Virtual Fashion is a division of Reyes Infografica, the 3D software company set in Madrid, creator of the set of Reyes plug-ins: MetaReyes, CartoonReyes, DirtyReyes, JetaReyes, SurfReyes, NPR1 and specially ClothReyes, the first commercial cloth simulation software.

= = =

In http://www.creative-3d.net/newsviewer.cfm?NewsID=1162 

So Reyes Info did not continue on poser fashion but on Virtual Fashion. It got a Basic and a Pro version. VF-Pro 1.5 dates from mid 2009.

Well, nowadays www.virtual-fashion.com as well as www.poserfashion.net have become dead ends anyway. I mixed up the tomb-stones. Sorry for that. 

some time ago I ran couple of simulations trying to find out what 'collision offset' did.

Well... no luck - but anyway, you may check it in following thread:

http://www.renderosity.com/mod/forumpro/showthread.php?message_id=3441266&ebot_calc_page#message_3441266

I can show you what collision offset is for. It's very important. In fact I have to show many new things I've learned.

Still pictures don't cover this too well. Any suggestions on how best to make animations of wireframes (don't need renders) from Poser? I tried a 400x300 movie using the preview renderer, 90 frames, and it's almost 7 MB. Without buying software - what's the quickest (preferably scriptable) way of making an animated gif that is small? (Really only need like 4 colors).

Just tried animated SWF - 4.4 MB. Ugh.

I was doing some testing of quad shear, and stumbled upon the need for collision offset. Watch this video. Observe how the cloth hangs on the block by just a few vertices. This was with the default settings.

http://tinypic.com/r/mkkf0l/7

Changing collision offset to .2 fixed it - the cloth drops to the floor.

You should see me dropping my dish towel against the edge of my table for the last hour.

Quote - You should see me dropping my dish towel against the edge of my table for the last hour.

No wonder our families and friends feel we've rather gone off the deep end ;o).

LOL...funny.

Laurie

{Gobsmacked}

I've been doing my own - albeit incredibly lame - experiments with Blender-generated quad-based materials, all experiments failing spectacularly, which I attribute to 1) a very inconsistent testing approach and 2) a complete lack of understanding of even where to put settings in Poser to start my testing. I guess what I want to do is start draping with plywood, and work my way to silk. At this point, my plywood sheet falls through my "smoke" square past the finish point in frame 30 to the ground.

Which makes me the compleat dummie - needless to say, none of these experiments will ever see the light of forums.

I'll keep trying, though, and eventually catch up to *this point, at which point you will all have moved on to re-writing the cloth room from scratch. :biggrin:

Quote - ... Any suggestions on how best to make animations of wireframes (don't need renders) from Poser? I tried a 400x300 movie using the preview renderer, 90 frames, and it's almost 7 MB. Without buying software - what's the quickest (preferably scriptable) way of making an animated gif that is small? (Really only need like 4 colors).

you can use mencoder (no istalation required) - not directly to gifanim but the file size of similar animation to yours (two props - one movable - and ground, with res 1080:560 and 100 frames took 800kB) - of course everything depends on the output quality you want.

You can use it inside python script (as I do) because it is a command line. Poser must render series of images and mencoder puts it together with any (almost) codec you want (and you have FULL acsess to all codec's command so you can get really good results).

In doc file they mentioned it could make anim gif directly but you need to download additional library which I just cannot find anywhere.

hi all,

I continued my strolling around the Net and in the end I found myself buried deep down in a worldwide scientific community on cloth simulation, US patents included, which evolved from about the mid '80s. That was about the moment I stopped in the scientific arena after getting my MSc on applied math and experimental physics. I never looked back. This means to me that I will stop looking further into this area on the vertex level. Back to the front! But I'll give you what I found on the ways forth and back.

From the saved pages from poserfashion.net I learned that the author (Serge Marck) was into cloth sim long before Poser started doing it. He stated that all cloth sim routines (those days) were based on the modeling of physical fibre behaviour in cloth. His main source was a tutorial by P.S. Karthikeyan. More on that later.

He also flagged and demonstrated that because of this, the simulation results depend on non-virtual things as real world cloth size and structure. In effect, he showed differences (from Cloth Room) based on cloth size at constant vertex density, on vertex density at constant cloth size, and on mesh type (quad vs tris) at constant density and size.
This implies to me that the mesh structure (and not the vertices only) is mapped onto an internal (fine grained) cloth representation. From that, layered ball/spring nets are deployed to run the sim, and then all this is translated back to vertex positions (and speeds?) in the cloth mesh. Things might get lost in translation.

Serge Marck also noted that since the Dynamics experiences in the Poser community were very thin those days, it could be an advantage to look into the forums of other applications, like ClothReyes, Stitch and SimCloth.
This last one is quite interesting, as it's an Open Source plugin for 3DS MAX, in its final state since 3DSMAX 3 (that's before the Poser Cloth Room) and only recompiled after that, and still here for 3DSMAX-2010. I also found references to SimCloth for its use in large scale movie and TV-series productions, in 3500-figure crowd shots. So, its code and docs can tell us something, I guess. 

Back to the tutorial. It's nowhere nowadays, I only could find a Russian translation of one chapter out of it, on the modeling fundamentals. This said that all started in the '80s from the well understood behaviour of cables (hanging bridges, telegraph cables) and the effects of fibre parameters, the effects of fibre thickness on those parameters, and the effects of gravity and wind (those two again!). Cloth then is interpreted as a two-way weave of cloth fibres each described with cable behaviour.

It also said that - at those days - P.S. Karthikeyan was a 3rd year student on the Aerospace Institute in Madras. I did not find any other refernces to him related to cloth sim, after that. But I did find that the tutorial must have been very good, because it's quoted in about every paper on this matter from those days on. Google rocks.

One of the papers I ran into looked worth downloading and sharing, it looked readable, an overview on fundamentals and implementations in the popular cloth sim software mentioned above and in previous posts. Recommended.

It's http://www.risc.jku.at/publications/download/risc_2354/manasilib.pdf 

Then we are into RISC, deep down the science dungeons. Get me out!

So, one way or another, the mesh structure is mapped onto a simulation of cloth physics based on fibre material characteristics, fibre thickness (which effects stretch and bend etc all in different ways) and weave/knit tightness. Poser does not grant us access nor insight to any of this.

With respect to all of this, Cloth Room definitely is a different beast than Materials Room or alike. It's not a Room at all, it's the Magic Kingdom Castle, or the complete Disneyland Park around it. I'm back to the front gate. See you there.

okay, hold your breath, we're going to do air resistance.

Athough there are various ways airflows can effect the motion of cloth, the simpest form one can expect in a Cloth Room is: 

• Air is flowing through a piece of cloth, at some speed v.
  There is an air pressure difference P between both sides of the cloth.
  The ration between is the air resistance,  a.
  Then  P = v * a.
  Like Ohm's law in electricity, current i, Voltage V, resistance R: V = i * R. 
  And yes I know electricity through a wire happens in a completely different way as air through cloth.
• The air pressure (difference) creates a force perpendicular to the cloth, proportional to amount of cloth (surface S) in m2: F = aSv  

Now, consider our drop-drop experiment. There is no atmospheric movement but the cloth is flat falling down at speed v, same effect, same force. But initially, the cloth is dragged down by gravity, at a force F=dSg . Same surface S, mass density d and Earth gravity acceleration g (9.8 m/s2).

While the gravity pulls the cloth down, it accelerates, speed increases, upward air damping force increases and finally, both forces equal out. From then on, the cloth falls down unaccelerated, at constant speed. 

From the equations we can see that from that moment aSv = dSg. The surface amount cancels out as the reult is the same for any size of cloth (and any shape, as long as it falls flat). And the final speed reads v = d*g / a.

When v is in m/s and g in m/s^2, then the ratio d/a is in sec. All other units in d and a should be similar, so if d is in g/cm2 as the Poser manual tells us, then a is in g/cm^2 too - per second.

We know the basic formulas for non-damped motion:

• acceleration g, a constant
• speed v = g * t, increasing at constant rate over time t
• displacement h = 1/2 * g * t^2

Now for air-damped motion:

• acceleration g - v * a/d, gravity minus air damping, it varies with speed v itself. So
• speed v = d*g/a * [ 1 - exp( -a/d * t) ] and therefore
• displacement h = dg/a * { t - d/a * [ 1- exp( -a/dt) ] }

Now back to Poser. I took my 32x32 cloth, raised it to 19,6 mtr, and folding etc to max to make it stiff as a marble plate, but kept desity and air resistance at default values. That's 0.005 gram /cm2 for density and 0.02 gram/cm2 per second for air resistance. The ratio a/d reads 4 /sec, and that implies that at 1 sec, the exp(...) thing is reduced to less than 2% and can be ignored, to simplify calculations.

That means: 19,6 mtr = 0,25 * 9,8 * ( t - 0,25) or: impact at 8,25 sec, that's just beyond frame 247 (poser time in cloth room is always 30fps, whatever your animation settings). That's theory.

Sim run. Drop down. Impact without edge distortions! (It was the 32x32 cloth that trouble me, you remember?). At frame 247 it's just above the ground, at frame 248 it has had the full hit. That's Poser reality meets theory.

So now we understand air resistance, in its simplest form. Now we can check out the Wind Generator. Later on, we can investigate on more complex forms of air resistance, like air blowing against a windproof / airtight massive plate. 

So, hold your breath again, I'm on Wind Force this time.

Actually, the experimental setup is simple. Piece of cloth, setup straight at a few meters above the ground. Clothified, and the upper row of vertices put in the choreographed group. We don't need a pole or so, the sim routine will leave choreographed verts untouched (so you can drive them by keyframes or what). And since we're not going to touch them either, they will stay put, and so make a hanging flag.

Give the cloth maximum stiffness (fold etc), zero friction, and default density and air resistance.

Then we put a Wind Force generator aside, and make it blow straight into the cloth. Take some distance (twice the cloth size will do), leave the angle at 45, and give it a serious range (4 to 5 times the cloth size will do). A Wind generator is quite a rude thing actually. It will proceduce the full force in an area determined by angle and range, and nothing outside it. No fall offs like spotlichts. Leave the Amplitude at 1, our queste will be what it means.

Run the sim, and find the flag hanging at an angle of 45. 

Now we know what's meant by Amplitude = 1. That's the wind that - when blowing horizontally - will push a default piece of cloth up with the same force that is exercised by gravity to pull it down. 

To look at it more closely, let's do the math first.

Take angle z as the ZRot value, the angel between the flag and the horizon. Z = 0 means a horizontal flag, extreme winds, and z=90 means hanging down, no winds.

The (vertical) gravity force on the skewed flag can be decomposed in a part along with the flag, stretching it and being countered by the pole (if we had one), and a part perpendicular to the cloth, making it rotate downwards. This latter force is

F = dSg*cos(z), d for mass densty, S for cloth surface, g for gravity acceleration constant.

The (horizontal) wind force has a similar effect, but we have to adjust for the fact that a skewed flag will present a smaller surface to the wind. Again, the force can be decomposed into one along the flag stretching it as well (you know cloth is pulling when the wind blows in), and a force perpendicular, rotating it upwards:

F = aSw*sin(z)^2 with airdamping a, windspeed w (in m/s), surface S and the sine squared thanks to the mentioned adjustment.

The flag hangs at equilibrium when both rotational forces cancel out, and since sin^2 equals 1-cos^2,  we can make it to

0 = awcos(z)^2 + dgcos(z) - a*w

From this we learn that when a or w equals zero (no atmosphere, or no wind) then the equation reduces to cos(z) = 0, z = 90, flag hangs down.
And we learn that when a and/or w grows really big (under water, hurricane) then the equation reduces to cos(z) = 1, z = 0, flag fully stretched horizontally.

From this we can determine the meaning of Amplitude 1: we just solve the equation for z=45 and find a windspeed of 3,46 m/s. Default density at 0.005, default airdamping at 0.02.

So, what will happen to the windspeed when we adjust Amplitude?

For this, I had to repeat the sim at various amplitude settings. 
Amplitude 2 created an angle of 23,5, that's a windspeed of 13,6 to 14 m/s, that's about 4 times as fast.
Amplitude 4 created an angle of 10-12, that's a windspeed of  80 - 55, or: another 4 times as fast again.

Amplitude 0.5 gave an angle of 56, that's 1.993 (say 2) m/s, or about half as much compared to amplitude 1. 
And Amplitude 0.25 gave an angle of 68, windspeed 1.067 (say 1) m/s, or another half as much again.

It looks like the amplitude and windspeed relate in a linear way below amplutude 1 or a 45 flag angle (half the amplitude gives half the windspeed), and relate in a squaring way (double the amplitude quadruples the windspeed) above amplitude 1 or a 45 flag angle. This gives artists more control in both regions, as if the dial changes sensitivity.

Now you know windspeed, air-damping, mass density and the gravity constant values, you can easily predict the angle by solving

cos(z) = x for x^2 - (dg/aw) * x - 1 = 0.

That is: cos(z) = { -dg + sqrt[ (dg)^2 + (2aw)^2 ] } / (2aw)

Note that the cases aw= about 0 or aw= really large already were discussed above.

I ran my sims for 100 frames, flags were still waving a bit at the edges (they make a full wave despite the extreme stiffness settings) but still enough to make estimates of the angles. My standard 32x32 quads cloth, as in the previous fall-down test.

happy sailing 

hi all,

it's quiet this weekend. Everyone is measuring up her/his own wardrobe, I guess LOL.

For now something completely different, closer to the original question, which was about: why does the Cloth Room interface look like an alien spaceship cockpit dashboard?

Well, I found out it doesn't once I understood the structure of the cloth room functionality. It's quite organized, even. And while looking at it, I found a few nifty things I didn't know before. Of course you did, most of it is in the manual after all. They just forgot to give me the overall picture.

Okay, this is what I've got, from a 'database perspective'.

**0. Poser handles one scene at a time. **

You don't have questions on this, do you? It's in for starting up the next one.

1. Each scene can contain multiple simulations.

Once created, all sims can be recalculated in one go (one after another), from the menu Animation > Recalculate Dynamics. You don't need Cloth Room for that. The recalculations (clear sim and run sim) are performed in the order the sims have been created. Renaming the sims is for your reference only.

Example:
A skirt is simulated in simA, then a simB is created to simulate a blouse (over the skirt). The skirt collides with Vicky, the blouse collides with Vicky and the skirt.
At recal, simA is redone but simB has not be cleared yet. For the skirt will not collide with the blouse, the blouse moves will be there, but are ignored by the simA calculation. Then simV runs, and the new results for the skirt are taken into account for derived the blouse behaviour. First all simA frames are done for the skirt, then all frames in simB for the blouse.

The catch: if the sims themselves were created in the reverse order, this all-sims-recalculation will turn out wrong. Then first the new blouse frames are derived from the old skirt positions, and then the skirt is renewed. In that situation, one might have to do all recalculations in the correct order manually, sim by sim, separately.

Now, the Cloth Room interface is not one big cockpit display, is presents several separate panels. The panel upper left (called 1. ...) deals with the sims. Presented in a list (in order or creation, same as the recalculation menu option, pick the one you want to work on). You can add, delete and edit sims settings. I've not found a way to copy/paste or save/load sims. Sims settings relate to frames (link to animation, draping) and the way the selected sim calculation is done.

2. Each simulation holds at least one sim-item (aka 'piece of cloth').

At least one. So you can put the blouse and the skirt in one sim together? Yes indeed. The pane below the sims, so mid-left of the interface (called 2. ...), is about sim-items. Or cloth objects, they're called. Presented in a list, you can add (clothify), delete (unclothify) and edit sim-items and their settings.

rule 1: each piece of cloth can be in one sim only, you cannot run two different sims on the same skirt.

rule 2:two pieces of cloth which are in the same sim, cannot be set to collide to each other.

And... you can copy/paste and save/edit. Sort of. I will come back to the item-properties later, like the groups and the parameters. At this point it's relevant to know that those properties are tied to the piece-of-cloth object as such. So you can clothify it, assign groups and parameters. the unclothify it, create a new sim, (re)clothify that previous object, and all the properties are there again. Or, the other way around, add an object from a new sim to an existing sim, next to another object in that sim already.

This works, as long as the above two rules are adhered to. When you pick an object which is in another sim already, you get a message.
But when the blouse is already set to collide with the skirt, then you take the blouse from its original sim, and add it to the skirt sim, it won't work. No message. But the blouse cannot collide to the sim when being in the same sim, so you have to unselect the skirt as a blouse collision target.
And... when blouse and skirt are in the same sim, and you want to select collision targets for the blouse, the skirt won't be there.

Then, how do we get the blouse to collide against the skirt? That's simple. Go back to the sim settings, and tick the cloth-self-collision box.  

This is good news for all these people struggling with moving multi-part conforming clothes into dynamics. You can put all parts in one sim, and run it. It will take time, but it will be more convenient than running 20 individual sims yourself in the right order.
Hence,  there might be no need at all for any conversion of conforming clothes to dynamics. Just use the things. Dynamic sim results take precedence over conforming (try it, animate one piece of cloth both ways).

And for testing purposes, you can take a single piece out and put it in a separate sim for the time being (and add collision targets), and then put it back again (take the collision targets out first).

When you select a piece of clothing, and you save it to the library, all properties will go with it. Groups, parameters, the whole thing. Then you can bring it in later, or in a new scene, or as a second clothing item. 

So yes, you can cut/paste and save/load (and duplicate) sim items, in a way.

3. Each sim item (cloth object) contains of one or more dynamic groups.

Next to that, each sim item has one group Choreographed (vertices remain untouched by the sim routine. They will stay put, or you can use keyframes, or they are moved as they are part of conforming cloth driven by a figures animation).
And one group Constrained (vertices follow the figures/props the cloth is set to collide, but after that the sim dynamics should not move them).
And two groups  Decorated. Vertices in there will not be dealt with from a cloth dynamics point of view, but will follow the cloth movement that results from the dynamics. The rigid group will not deform (buttons), the soft group will (loop belts).

Note that the manual issues a health warning for intersecting vertices: when you assign a vertex from the cloth itself to a decoration group, some hole in the cloth will remain, effecting the dynamics. So mesh-parts which are separate from the main cloth mesh but grouped into the object, are favorite. Otherwise, take care.

All groups are called: Cloth groups. they're handles in the left lower pane (3 ...).

That leaves us with the dynamics groups. There is always one: default. But you can add, delete and edit to the list.

Each group consists of vertices. Each vertex can be in one group only - dynamic and non-dynamic together (), and vertices which are unassigned end up in default.
() not true/sure, there are checks for vertices in multiple groups. But then what are their properties and behaviour? To be investiated upon. 

4. Each dynamic group has its own parameter settings

That is what the grouping is for. You can have a lace-group, and a leather-group, and create a mixed piece of cloth with patches or stripes. This is why the group editor offers the ability to use material-groups for vertex selection. This way, the cloth-part with lace material gets lace cloth behaviour, the leather looking part behaves like leather, etcetera. 

The parameter settings are in pane 4, top right. for you convenience, or confusion, it starts with two buttons Calculate Simulation / Play Simulation that might better be in the first simulation pane, as they concern simulation, not parameters. On the other hand, you need to take steps 1,2,3 first before Calculate Simulation can take place. That's why it's in pane 4. But it sure raised the alien cockpit impression of the cloth room layout.

In the same pane, below the parameters, you've got a button Reset to restore the parametrs to their defaul (linnen), and a Clear Simulation you need to do before a proper recalc can take place, after changing any or all of the parameter values.

And in addition to all this, bottom right, a collection of buttons to miscellaneous routines that can be of help, especially when dealing with cloth and clothes. Has nothing to do with simulation or dynamics directly. Handy to have them around, confusing at first to have them in the same room interface.

So, essentially, (4) you've got paramaters per dynamic group. You've got (3) multiple groups per cloth object / sim item. You've got (2) multiple (!) sim items per simulation. You've got (1) multiple simulations per Poser scene. And you can (0) recalculate them from one menu option.
Each step in this has it's own pane in the Cloth Room user interface. The panes are numbered 1234, and you can tear them off, and resize them. You can make your own interface, sort of. It's not that bad after all. I just lacked the oversight.

I hope this helped a bit. It did for me. And wherever I went wrong (I do have a track record for that), please contribute.

happy clothing.

(edited for the most obvious typing errors)

This post is - obviously - the page missing from the manual !

BB, nice hat in your avatar. Full Dynamics, I presume?

I needed to do some conforming cloth, no dynamics, just for a change. Result is in the gallery, best wishes and a creative year to all of you.

I'll continue the Dynamics Queste soon. We'll have to get on top of it. Where is everyone, BTW? Am I missing a party, or what?

No parties, I'm afraid... real life takes over during the week, I think, aRTBee.

According to the credits for Poser, the cloth simulation was written by Size8 Software. All I can find for them is a plug-in for Max which is sold on Turbosquid. On the miniscule one-page website they claim to be the developers of the Stitch cloth simulation software, but elsewhere I see that this product was developed by Digimation.

It would be so great if we could have insight into the logic of the code. But trade secrets, I suppose...

hi,

Thanx for sorting out.

I see that in my first post on the code background, I mentioned SimCloth, ClothReyes, ClothFX and Maya/Cloth. All referred to in poserfashion.net. The pages of that site do not mention a creation date, but say: MS Frontpage 4. 
According to wikipedia, Frontpage was issued in 1997 as frontpage 98, in 1999 is was called Frontpage 2000/version 9 in order to follow to numbering as for Word etc. In 2001 the FP2002/version 10 was issued. Wikipedia notes that the 4-8 versions were never mentioned formally but html tags did refer to them. 

This places the creation of the poserfashion page (and the knowledge itself) around year 2000. That's just about the release of Poser4 and two/three years before the first release of Dynamic Clothes in Poser5.

The Size8 site (page) links to TurboSquid, and there is says; ClothFX. As a plugin up till Max 7, then build into Max 8. So that's what's happened to ClothFX. And ClothReyes for Max is clearly history.

On www.maxplugins.de : Clothfx (formerly known as 'Stitch'). Digimation issued Stitch Lite per feb 2003, compatible to Max 4, Max 5 and Stitch.
http://animationartist.digitalmedianet.com/articles/viewarticle.jsp?id=15461 

Since Digimation was buying about a lot, including some parts of MetaCreations as well, they might have acquired Size8 on the fly too. So, Stitch etc is from Digimation as well as Size8, both statements might be correct. I don't know.

ClothReyes seems to be ended up in Poser Cloth Room, see earlier posts.

SimCloth is still around as a Max plugin. It's opensource, so just go http://www.spot3d.com/simcloth/ and get the code, if you like. That is; if you want to find out about the internals of Poser Cloth Room / ClothReyes, SimCloth could be an equally good choice as ClothFX/Stitch, but far more accessible.

Hope this helps. Enjoy, I'm curious about your findings.

aRtBee been following along, even on the weekend. I have nothing to contribute so rarther than say "oh, ah", I stay quite. But that doesn't mean I haven't found the work you guys have done totally fantastic, and a good historical read too. I found lots of great tips, and I'm not so scared or confused by the morph brush now oddly enough.