SimCloth test and question

[apu]

Hi. Attached are two simcloth test of an arm like object moving and a cloth that reacts to it. I,ve been playing with the collision tolerance and solver precision settings to get the look that the cloth is in fact reacting to the movement of the arm. The problem is that in order to avoid the cloth penetrating the arm, I have to set the collision tolerance real high, and the look I get is like the cloth is floating on top of the arm, that they don't get to touch. If I set it too low, the cloth penetrates the arm.

Any ideas, suggestions?

Simcloth seems like such a great plug in, but this kind of problems ∞¢#@#¢∞ off!!!!!!

Thanks!!!!

Test 1.mov

Test 2.mov

[BuzzLightyear]

Hi Apu - I agree, SimCloth is really amazing, but a bit quirky. If you could post the EI file, I could play around with it to see if I could improve the intersection problem. In general, I would suggest experimenting with the subdivisions in two ways. First I’d try modeling the cloth in different resolutions, without using the subdivisions option in SimCloth (just run the sim with your modeled resolution). Then I’d take the cloth with the best results, and then try subdividing it inside SimCloth. I can’t say if it will work, but I’ve found that SimCloth seems to like certain mesh resolutions over others. Must be something in the math…. If you’ll post your example file, I’ll take a look at it if you want. Chip

[apu]

Thank you Buzz.

Attached is the simcloth project. Lets see if you can figure out the solution , becouse I surely cannot!!!!.

Thanks a lot, I appreciate it very much your help.

Apu

SIMCLOTH.zip

[RichardJoly]

Apu,

Part of the problem seems to come from the excessive calculation SimCloth had to make to position the cloth on the arm. You arm was scaled 20.3 times. So every frame SimCloth had to expand the model 20.3 times, calculate the bend from the bone, check the new position of the polygon and position the cloth where it should be...

I exported the arm with transformation and the cloth behaved a lot better.

It is not perfect but it's a place to start... The cloth is still floating a little over the arm but it could be the cloth resistance. Try different settings to check of it hug the arm better.

Cloth_on_arm.mov

Project2.zip

[Tomas Egger]

Ola Rick,

You are a “Jedi”...

Thanks

Tom

[apu]

Thanks for your answers! I m travelling and i wont be able to check it out until next week, but ill post my results!!

[BuzzLightyear]

Hi Apu,

I just checked your posting a day after you uploaded it. I haven't had a chance to look at it yet, but just helped another user with his SimCloth problem. As it turns out, part of his solution was also a scale problem, as Richard discovered in your file. In the other users case, scaling objects too small causes SimCloth to create unpredictable output. Richard's Quicktime looks a lot better, and clearly gave improved results. Tweaking the tolerances in SimCloth will no doubt get the motion you're looking for. I hope you post the final results; it would be great to see them! Chip

[apu]

Hi Richard, I just saw your project. Amazing. Thank you very much. I just had no idea about scaling the objects could affect simcloth so much. Whats a good scale for simcloth? I mean EI units, couse I saw that the arm in your project was about 100 units long.

If its too small (like 1 unit) simcloth will not make the simulation properly, right?

Again, thank you very much for your answer.

Apu

[RichardJoly]

Apu,

I don't know that small units would inevitably cause problem but if I remember correctly, VERY small polygons from a group (like your 1 unit arm) could be problematic. The idea is why chance it? I usually work with models 100 units and over.

In this case, I think the problem was that you scaled the arm 20 times and SimCloth had to "guess" the position of every polygons (after deformation) of your arm. Exporting the arm at 20X the original size removed this calculation for SimCloth and the results where more accurate.

This was an educated guess and it worked... :)