Character Cloth Simulation Explained
http://vimeo.com/40797021
By Jeannot Landry.
Jeannot writes:A lot of ppls complains about blender cloth simulation and say it doesn’t work properly. But it does work, they just don’t understand cloth enough to make it work, or simply use it the wrong way. Here is a breakdown of how I achieved it and it works great ;)
Nice demonstration ! I like the tip of parenting to the root bone, i usually use my character’s armature, and it’s a bit complicated with the pinning setup :S Thx !
I’m sorry, but that doesn’t look good. I wouldn’t flaunt this video. Perhaps you don’t have enough divisions to deform right. The jump is rather fast, so if you can do some sub-frame sampling, it might work better, but, for right now, I wouldn’t flaunt that as though it worked. It’s too jiggly and the folds it makes aren’t realistic looking. Also, the forty frame pause at the beginning shouldn’t be necessary in most animation packages. Most animation packages with decent cloth will allow some way of running the simulation and caching the settled position for that frame for a start. If blender doesn’t have that, it should. I hate raining on parades, but this result worries me more about the program, not assure me that it’s viable.
It is not perfect. But it does look pretty decent IMO.
He did a good job on that. With some tweakings here and there it will look even better.
easy, move to maya, good luck ;)
If you’re going to be like that, then at least say move to Houdini – which at least has a free learning version available for Linux. Remember that this is part of the Open Source community.
On another note, he never said it was perfect, just that it worked. The animation is rather rough, and I’m sure there are better cloth animations out there done in Blender. Also, higher quality cloth simulation would have taken a LOT longer – as it is, it took 6 hours for this guy (you probably didn’t watch that far into the video though, judging by your response). Not that ‘ZOMG MOAR ACCURATE!!!’ is the answer to the problems necessarily, but you never know – maybe this guy’s computer was just kinda on the slow end, and couldn’t handle much.
I was being sarcastic, if you dont like blender then go and suffer with Autodesk
I think that in real life air drag also has to be taken into account?
to the whiner…….he tried to find a way what have u
done…except…well….whine…..this is a good start I never saw him
for moment suggest that this was THE way or that there was no room for
improvement ……..why dont u make yourself useful and try to improve
upon it…or come up with a better way………….I am not opposed to
criticism but nonconstructive criticism is well as useful as a fart in
the wind
But this was posted on Blender’s main site. When I go to Blender’s main site, this is put forward like an example of what Blender can do, and it’s not impressive. It makes the program look bad. I could understand your response if this was a forum post, but this was presented as, “Eh! This feature works fine. Look at what I did and what Blender can do.” Not somewhere off the main site as, “Eh Everybody, I’m messing around with the feature. Here are my results. What do you guys think?” If this was the latter, I’d have been quiet. As a header on the left column of Blender main site, no. No apology.
Well in case you didn’t notice, this is Blendernation.com and it is comprised of both Blender users, amateurs and professionals, and the links on Blender.org are just updates on what’s being posted here. So no… this was NOT i repeat NOT on the blender.org official website as an official example of what blender can do. That is what the open movie projects do. WATCH THEM.
Yeah. Well you get what you pay for. (And with Blender you get a hell of a lot more.)
You are the same guy that already bitched about the ‘No more Heroes’ model right?
Don’t you have better to do with your valuable time than that? I mean, if you know how all that stuff works, go and show us your skills. I’m sure we all would appreciate to learn from a master like you :D
But just bitching around under an anonymous name is pretty lame.
I didn’t realize these were just auto links off Blender Nation. That’s my bad. I had thought it was a designer trying to justify a feature as functional, when some were complaining about its implementation. That kind of stuff drives me crazy. If something is counter intuitive or hard to use, designers should fix it, not defend it. Anways..
Sorry to the artist. I didn’t realize that work posted to Blender’s front page was, indeed actually enthusiast work, not designer work.
As for the site designers..What the heck are you doing? You realize what’s shown on your main page reflects the quality of your program, right? Reserve news for actual development and even then, push that further down in favor of showing off your project’s best accomplishments. It’s great to encourage enthusiast work, but unless it’s making it into magazines or comes from a larger production, it doesn’t have a place front and center. By all means, have a link to Blender Nation, so that enthusiasts and would be enthusiasts can find a place to get involved, but the main and only thing you should be concerned with in your front page is showing what your program can do.
Sorry, I accept your apologies…
But I want you to know that I am not just an enthusiast, I have been doing 3D professionally for over 11 years now. I am a former maya user, know it very well, still use it from time to time for contracts. I am not an animator, but a character artist and a rigger (Although I am learning animation). I am a contractor and I use Blender so I don’t have to pay for Maya anymore since its pretty expensive and its hard times and economy at the moment. This was my first ever real try at cloth SIM in blender on a full character. I didn’t want it to be posted here, but it did. As far as the mentality you have about company page, blender is full open source, not like maya or any other Autodesk products where everything is shared in secret, so they do want to share what ENTHUSIAST can learn on their own and achieve. That’s the entire point of Open Source, Share knowledge.
With the cloth sim, you had some good frame by frame deformation. I wish Vimeo and Youtube allowed a frame forward, frame back functionality that quicktime allows, so that I could really see how it looked every frame, but oh well. I think your issue was that the cloth was too active, needed some form of dampening. More divisions may have helped as well. Cloth deformation does a vert by vert calculation, so the more verts the better the end result, though the longer the calculation. Also, when working with any simulation, you generally need a sub-frame sampling, otherwise the difference between one frame and the next can sometimes cause large kicks of material. Lastly, you may have to key some of the attributes, like dampening. Your character’s jump isn’t bad, but it’s stylized, meaning he’s moving a bit further and faster than a natural object would. This is problematic for simulators, which assume that the influencing objects resemble some form of reality. Because your character is making some pretty big per frame movements, you’ll likely need to key up and down dampening or some other factors to accommodate that movement.
As for the site, I understand that reasoning, but you have to see it from a developers standpoint. If I take this program to an art director and say, “Hey. This program has made a ton of improvements, really can do some good work on it, and it’s free. I think we could save a lot of money and produce better work,” that art director comes to this site and looks for examples of work. If the first thing that art director sees is enthusiast work, you just lost the benefit of first impressions. If Blender really wants to get a foothold in development, it needs to control its image. It needs to put forth that it is a quality program worthy of a production environment. If it can manage to land a few large projects, some real credits to its name, more money can flow to development, and the whole industry benefits from some real competition (likely forcing autodesk to actually listen to its customers).
Blender SHOULD link to enthusiast sites, but its front page has to lend it’s program credibility to its potential for projects.
I understand your point, and I agree until a certain extent. The reason why blender is popular is because it is free, most certainly, but also because it is user friendly. It can obviously do anything and more then any other 3D apps on the market, and blender have a few movies to prove it such as Elephant Dream, Big Buck Bunny and Sintel. If people would actually read instead of just clicking blindly, its obviously written News Headlines from Blender Nation and not News Headline from Blender Foundation news.
As far as simulation goes, thank you for the pointers :) I have actually went ahead and did more research on blender cloth sim, and I am currently baking a new one, with the same animation, and hopefully, my understanding has increased and the results will be better.
From my experience in studios, which I hated for the simple fact that they didn’t even want me to try out animations, it was more like, you are good at modeling and rigging, so stick with that even though my dream is to be an animator. We are the one busting our ass making the work possible, but the person doing almost nothing is putting big $$$ in their pockets. Which is also the entire reason why I became a contractor instead and forgot the idea of working in studio.
To me, blender is more like, here, try, learn, have fun while doing it, its not all about money, but about passion.
My experience with larger studios were that they were bureaucratic nightmares with a frustrating amount of wasted work and time. Smaller teams are much more organic, and you gain the freedom to try and do more. It’s nice to actually put your name on something, rather than being a speck on a three hundred long list, too. Doesn’t pay as well, but definitely feels better.
I’m trying to get one of the projects I’m working on to support Blender. The game wants an active modding community, and if they supported a free software like Blender, I think it would be huge for innovation.
Well, we seem to be a on similar page after all :P
Good idea of wanting to opt for blender community, it is extremely active, full of great potential and a lot of informative contents for anybody to learn. Some areas of blender are still not very much understood, like cloth for example, but its up to us to test, try and share our knowledge.
Perhaps showing a bit more appreciation and support while still sharing your opinion or knowledge to help would be a better way to gain the trust of the community such as myself.
I felt offended by you in some ways since this wasn’t a tutorial and all, but at the same time, I tank you, because now It sorta motivated me to dig further into cloth and go try to prove you wrong, and if I am successful, I will have a nice tutorial for everyone to be able to learn from, add to and what not ;)
Sorry about that. I came at you thinking you were a designer, and I’m afraid I brought some of my prejudices along with that. Didn’t realize you were a bystander roped in from a separate site. I hope your work pays off.
Hi, Please show us what you do and stop whining. Please put a link.
lol @ sorry.
Blender is always a work in progress. It challenges users to think of unique methods and workflows to make things work. Don’t bash on someone trying to show their method for doing something. Theres more than 1 way to skin a cat. And thousands of ways to model one. If you’re worried about the program then you must be new because the rate that this software expands and improves is awe inspiring in itself. Just read all the release logs. The only program advancing so well between itterations is the zbrush
I second that, ZBrush is pretty amazing.
I had the privilege of using it in the past and I am hopping to still use it in my future as well. I use sculptris in the meantime, since my computer doesn’t handle a lot of poly’s in blender, but in sculptris it does, so its my new choice which sort of takes care of the Zbrush side of things ;)
While I feel that this reply was a little harsh from the user, I agree to a certain extent. The cloth does look too jiggly, but I think that is probably a cloth setting problem. Setting it to leather (even though the shirt is not, of course, leather) would probably keep it more in shape, or just tweaking the stiffness and bending settings of Denim would probably work too. It looked like a cotton animation, and to me, the cotton and the silk don’t work so far for character animation. But this poster obviously has not spent much time researching Blender, and to him/her I say this: the development team is amazing. Hands down, no questions asked. They release a new version at least twice a year, and every time they improve. Look at Sintel and tell me the cloth doesn’t work. It DIDN”T work, but it does now. It’s tedious, yes. But Blender will always keep improving, and I think by version 2.7, they will have a good cloth simulator that can do as high quality stuff as everything else in blender does. If the poster of ‘sorry’ reads this, then please please look into blender and their development pipeline a little closer before worrying too much about the program. If you look at the Open Movie projects, you should have absolutely no doubts in your mind that Blender works. If you can’t be patient enough to do it, then maybe 3D animation isn’t your thing.
thank you :)
yeah i agree, the video you posted was alot better.
I think the chief point you’re missing is that fundamentally, Blender – both as a software and as a foundation – is run differently from a commercial corporation in lots of subtle ways.
The first key way is that since the community is open, there’s a wealth of information and it comes from sources with different levels of experience. So whereas the Maya front page will only list the AAA films and games it’s been used on, Blender lists indiscriminately. Since it is not a commercial program, it has no reputation to uphold, and since it’s not in a competitive market and doesn’t serve a competitive user base, users are far more open with sharing ideas and giving feedback.
So saying that this video makes you worried about the program isn’t entirely valid. The goal of this video isn’t at all to get you to download Blender – the Blender community gets nothing if they win you over. Instead, you have to see a site like BlenderNation as a window into the community of people already using and developing Blender – they share ideas and sometimes the ideas are half-baked, but that’s okay. Since many of us are hobbyists, you have to understand that we’re not trying to sell you on something; we’re doing this for our own edification and for the benefit of the program itself.
That’s what Blender has been built on. Different people with different experience bring different benefits to it, which is why it’s a far wider-reaching program than most commercial programs, capable of motion tracking, compositing, modeling, animation, rigging, rendering and editing all within the same window. All those functions were initially made by different developers who decided to take it upon themselves to turn the program into something suited to their individual needs, and over time those developments were integrated into the program as a whole, creating the program that exists right now.
Given all that, I’d say that your criticism of the video was valid and probably very helpful. On the other hand, I’d say that your comment saying “I wouldn’t flaunt this video” misses the point of the program, the community, and the entire mentality behind open source software.
The video is not bad at all, i have toyed with cloth simulation for capes and flags (so very simply and basic shapes) and while i always had some more waiting time to do with the caching than i thought considering the the low poly and few frames i was using, the following really worried me :
“It took 6 hours to cache the 160 frames”
Even subdivided, the model and the clothes still look rather low poly, and it’s “only” 160 frames, and still it takes 6 hours to precache the simulation for them ?
Why does it take that long for something low poly and for a short animation ?
Collisions are the most expansive parameter, and he has some on the body and the short…
Have to agree with wildduck, it’s the collisions and self-collisions. When you have capes and flags, it only occasionally has to self-collide, and rarely (if ever) has to collide with a fully animated and moving object. For this animation, the artist has both set on maximum – so it will take a lot longer than a flag or cape.
Keep in mind that fully animated means deformation as well as changing position. This isn’t a ‘sphere moving against a cloth’ animation, this is a ‘humanoid figure jumping and moving every part of his/her body’ animation.
Great demo ! Congratulations.As it requires many tests and a long baking process, I think that most of us give up at some moment by a lack of patience or use too low quality parameters to speed up the process and so we are not happy of the result.
Heh! Looks like his tshirt is full of jelly! ;)
But yeah with some mild tweaking this would work well. :)
Waiting 6 hours for 160 frames.
Oh my Buddha!
People people people… he has attempted something new and shared his knowledge… Even if you don’t appreciate please stop discouraging…!! BTW… Kudos brother, that’s a nice demo… I will make use of this in my upcoming animations… cheers… :)
Amen
+1
That’s great! Thank you so much for sharing your setup with us Jeannot!
Wildluck: thanks :)
Sorry: Well, I know its not perfect, but it doesn’t go through the geometry. I am a former Maya user and I know Syflex very well, blender cloth Sim isn’t as good, but it does the trick. Maya is overly complicated for nothing, and its over priced as well. Maybe this cloth simulation isn’t realistic, nor the animation perfect, but it still proves that it does work with a bit of patience. If I would have more time to tweak this until it looks good enough for you mister, I would , but I don’t…
Jay:Thanks :)
Lapinou: yes, I should’ve brought up the air drag a lot more, it would look better.
Gobuis: There is a lot to take in consideration for fast movement so it doesn’t go through, and there is 2 collision, the short one being the most expansive since it has to repel the shirt as well. Taking the shorts out, you bring down cache time to about 1/8th
Philippe: Thank you :)
Dusty: Yes I know, I find that it sorta fits the cartoony character :P
Jack Qoop: I didn’t wait in front of the computer lol, I let it bake over night :P
Amarneethi: Thank you :) and I am happy to share any discoveries I make here and there :)
Chicortiz: Thank you and you are welcome ;)
To all, well, this isn’t what I would normally do for this sort of cloth, I would probably make the shorts part of the rig instead of calculating it, which adds a lot of calculation. I would normally only use it for like capes, dresses, and very loose clothes. But in this case, I really wanted to test the cloth and I did, its not perfect, but I am happy that it does work, now I know where to add tweaks to make it nicer.
Happy Blending ;)
Jeannot, I forgot to congratulate you for your great tutorial! my comment was not aim at you (your results are IMOH great!) but rather to user “Sorry”. It is now up to us to experiment with the parameters. Kind regards
You did a great demo. I have been messing with the cloth a lot, and the T shirt proved to be a B@#$& and like one of the posters said, I gave up. But never once did I think it didn’t work. I have been using Blender for only a little over a month, but for a free software, it is extremely top of the line and I proudly joined the Blender community. Maybe ‘Sorry’ is willing to spend six hundred to several thousands of dollars for a top of the line, expensive production suite that chews up his ram, but for me, I just wanted to animate, and people like you are helping me do it, one tutorial at a time. So thank you. Thank you thank you thank you.
I just wanted to ask: I noticed that you have depth to your cloth, and it sort of acts as a fully closed mesh… But all the other cloth simulations I’ve seen/done myself (albeit simple tests – I’m no animator or artist, sadly) have been with flat meshes without depth. Does this affect the realism and/or bake times?
I was thinking that, since it (seems) to work just fine with flat meshes, meshes with depth would have lots of self-collision calculations and would take a LOT longer to simulate.. But I may be wrong about this. I’d just really like to know more information on this :)
The depth as you refer to is a simple solidify modifier added after the sim, it just gives a thickness to the shirt, perhaps much in this case, but since the character itself is cartoony, I went with thicker since it sorta fits better with the style.
Nice demo. Everything was perfect until you said: 6hrs for 160 frames ! Also I have done cloth simulation without using softbody, it makes render reaaaaaly long
I did not want to comment on this because although its a good video and it explains the technique well its is not a good demonstration of usable cloth. This is not a reflection on the OP I think that the cloth simulator is just not up to the task of doing full body clothing in a reliable and repeatable way.
I have eneded up modeling shape keys and rigging the cloth deformations on my characters as it was the only way I found to get reliable and repeatable results. You can use a softbody to controle a rig so that the shapekey moves like a real flexable cloth.
Could you make us a tutorial on your methods please,,,,,,,,,,,,,,,,,,,,thanks
he won’t, this fatty guy is always complaining of something. He thinks he is too good : he hasn’t even updated a useless ANT rig ! What the heck ! He is always supporting Houdini in other Forums ! Bye Bye Mister Fatty !
EEee Gads! This needs to be rated R for nudity. I had to close my eyes.
So, It seems to be kinda like with Blender hair. It may need some serious work, but it can be done.
I’ve seen better with the blender cloth simulation, but its still good.
Nice to see your Demo. There is not many good tutorials on clothing on an animated character on the net, mostly just the basics like a cloth falling onto a sphere, and just to many of those, so it was nice to see your method. I have had barely acceptable results with the cloth simulator in blender. I will be trying your methods. Thanks
Congrats on making it to BlenderNation,
I liked the video. The bakes however are killer in their times but eh. All the bakes are eternal.I liked the idea about letting it prebake. I forgot all about that one. The settings you also have work well for your animation. I want to try and apply that but reduce the amount the cloth simulator. Like maybe only affect the sleeves and bottom of shirt? Then use the armature to deform the rest. But yes, excellent demonstration!
thank you very much for this. i’ve tinkered with cloth sim with poor results. i’m going to be giving this a try real soon.
Thanks for posting this. I’ll see if this is a viable option for my work – I haven’t seen a lot of cloth tuts out there so I always welcome different and/or any approach that works! :)
As far as bake times go, what would be the biggest hurdle in getting baking for cloth (and others) to be multi-threaded? Doesn’t make sense to have 3+ cores sitting idle during such a process.
Very strong point Matt. For those of us with plain old Macs, it would be nice to have a way to fully utilize multiple cores.
The biggest problem is the physics itself. Whether or not a problem is easy to distribute over multiple cores/cpus/nodes, depends on how strongly all the variables in the simulation influence each other. Rendering is easy to parallelize. BI does it by rendering in tiles. Each tile is handle by a core. Particles is somewhat harder, because you have to calculate the forces between all the particles for every step in the simulation. But still, one can relatively easily code a particle system to run on a GPU for instance.
I have no experience with cloth and I don’t actually know how it is simulated, but I would imagine that one would simulate the position of every vertex in the cloth. The forces that would be exerted on these vertices would 1) little extending springs to the other vertices to simulate stretching. 2) bending springs on all edges. 3) Forces fields (gravity, wind) on all vertices. 4) collision detection on every face. The last problem is probably the hardest, because the collision detection should not only work on the faces of the cloth mesh, but also on all the meshes that it can potentially collide with.
So in short, there is definitely something to win in parallelizing a cloth simulation, but it is very nontrivial. In the end, if at all possible, one would want to run this on the GPU, because the more cores the merrier. But this is not something someone just does for fun on a sunday afternoon.
A further explanation of clothing is always appreciated. THANX
Jeannot
Welcome :)
You can greatly reduce baking times by having the Collision calculations only use the low-level mesh. (by ordering the mod-stack on the character object as Armature –> Collision –> Subsurf). If the cloth object’s low-level mesh is sufficiently dense you should not have any serious clipping problems (nothing that the subsurfing, a VG-targeted Mask modifier, or a small bake-edit couldn’t fix). This does tend to cause an “air cushion” between the cloth and the skin after both objects are subsurfed. As long as it’s not too wide, you can fill in that space with the virtual thickness that the Solidify mod provides by turning it inwards (opposite the direction of the normals). As an aside, I find that putting Solidify before Subsurf in the mod-stack is less-CPU intensive and you can also take advantage of Solidify’s auto-crease weight options (to get sharper edges).
That’s an unusually large amount of Repel force, by the way. I usually only set it to be identical to the Collision Distance, and then set the Repel Distance to double that.
Thanks for the additonal helpful info!
I hope it’s helpful! Here’s a few extra things I didn’t mention that can save more time.
Use only as much Quality as needed. — 40 is more than likely overkill and each quality step takes more baking time. I rarely use more than 5-10. Same thing goes for the Cloth Collision settings. These don’t really need to be taken above 5 Quality unless you need to address serious clipping problems. Like I said above, minor clipping can be fixed without rebaking.
Use a proxy mesh for collisions. — Duplicate the character object and delete any geometry from the duplicate that you’re certain won’t be touching the cloth objects during the animation. You can give this cut-down proxy the collision modifer instead of the actual character (and set it to be nonrendering). As I understand it, collision takes every face on the mesh into account, even if they never actually collide with something, so this can reduce ‘wasted’ calculations.
Use object groups to filter colliders. — This is more for when you have multiple clothsims in a scene that don’t share colliders. Without object group filtering, the clothsim will take every colliding object in the scene into account, even those that never touch the cloth object, meaning extra baking time from wasted calcs.
Thank you for those pointers :) I am testing those right now :) Do you perhaps have any examples of these that we could all see?
Very informative. I always wondered how someone would make pants cloth without them falling.
You’re officially a tailor. :-P
I guess this just shows how far Blender has come. The naysayers are actually visiting.
That’s right; it seems like all the naysayers are just spectators and don’t actually use Blender.
Haha indeed! I was wondering where the sh*tstorms were coming from recently and that would certainly explain it.
A walkthrough like this was overdue. To my knowledge there’s not much help on this topic so far. Thanks!
There are some good hints for improvements in the comments, too. But they’re hard to find within the noise ;)
I agree, this is one of the reason why I gave it a shot, at least tried :) I am trying to improve it :)
thank you for sharing.
You are welcome :)
Hi, Jeannot!
Your other vids on Vimeo costs an attention for sure!
I’m surprised why it’s not spotted here on BlenderNation. BTW the language sounds almost without accent… Really a bit of French :)
This serie http://vimeo.com/27920689 of vids looks complete and really good. Guys, I found there a lot of cool advices and hints! :)
Thanks for this, Jeannot!
Thank you :) I am flattered :)
As is, That looks bad. It does show that with additional work it could probably be used effectivly.
Still, the additional work should have been done before releasing it.
First of all, I can post what I want on my vimeo, secondly, I didn’t choose to post it here, I didn’t even ask, Bart saw it, liked it and I guess he posted it, so I don’t have to wait to release any of MY OWN STUFF… I am however, after some of that hatered, trying to make it better.
If you see it as hatred you probably shouldn’t release anything publicly… ever. lol
I just said, your work is promising but not finished.
If I shaved half of my beard off and went into public I would expect people to say it looked bad… because it isn’t finished.
(I also wouldn’t use it as an example of how to do something.)
Do you see anything in the video that says, Here is a tutorial on Cloth, or Here is how to do it? No, this video was only to show that is indeed possible after at least trying and how I achieved it. Never did I say this is the right way to do it or not. I just wanted to share with some of my friend through my vimeo and get feedback. I didn’t expect it to be posted here since it was not complete! I didn’t even want it to get to here in the first place since I knew it wasn’t finish and it was only meant to be a test, otherwise I would’ve requested it.
Now I am getting judged for something I didn’t intended in the first place. How nice!
Maybe Bart should take this into consideration in the future even though it requires extra effort on his behalf.
I can understand the frustration of getting WIP stuff bashed and critiqued as it was a polished masterpiece even though it never was intended for a larger audience. I guess you already were aware of most of the harsh critiques in the first place since it only was a test. I think it’s a shame you had to read all this crap and it could have been prevented with a simple question of “Do you want this featured?”.
You are not suggesting anything but just generically berate a work-in-progress blender diary.
Are his audience paying customers? Are his claims outrageous? Hell No!
The community is known for asking productive feedbacks not just some generic statement and awful analogy that you just made. With your mindset, blender.org will just release major version numbers instead, just because people like you think versions like 2.61 and 2.62 are ‘half-ass’ and ‘needs additional work before releasing’ due to bugs and incomplete features.
Good tut,I download the file and reset the lighting setup and render it in 1Hr.58 min. A lot better then 6Hr.
6 hours was how long it took to bake the physics, not how long it took to render.
Richard, it took me only 30 min to render, it took 6hrs to bake the simulation ;P
One hour and half bake and 30 minute render time. I have a different blender build then the Blender foundation…
Two questions:
1) Why did you parent the cloth to the root bone? Shouldn’t the collisions hold the cloth in place?
2) Why did you set the spring setting on the cloth so high? Taste or necessity?
Actually, the only reason I parented it to the root bone is because I didn’t use an Armature modifier, and if I didn’t parented it, the pants and shirt were left behind :S Still testing and learning cloth. These are simply my findings so far …
Thanks!
I appreciate it.
I’m also really wondering about question 2). Did you need to set the spring constant that high (at 40 or so) to keep the clothes behaving predictably?
Actually, I thought it would help the cloth not going through the geometry, but I was wrong, that is essentially what caused the jelly type like final feel it had. I am rendering a newer version, a lot better and now I understand cloth a lot better and I will try to explain as best as I can. I am not sure this new video will get posted here after all this, but I will come over and post the link.
Oh, very cool!
Do let us know!
Holly crap dude I did not even watched the movie yet , now I am bit scared , I may make some popcorn, and coke will do fine .
that was cool .. with a nice soft cap ,hat would have been nicer .
Well done for working it out :) I’d still be inclined to say that it jiggles around way too much though, it was really distracting my eye. That said, I haven’t gotten it working myself so you’re miles ahead of me.
What’s the big deal with the 40 frame settle? Any simulation I’ve done in Max has needed it, it it really such a problem?
the 40 frames that sorry is refering to is, in maya, you can set the timeline in the negative for that sort of things, well, in blender too, you just have to enable it in the preferences, but most ppls just advance the animation a few frames like I did, even in maya using syflex. The user sorry think he knows maya, but I have used maya for 8 years in studio, and its good, but its a pain in the ass. Therefore I switched to blender and i love it ;)
even while not being a tutorial, for the people open minded enough to look past the flaws it was a good try that provided good helpful hints….as I said above I saw nowhere any claim that this was THE way or meant to be THE way, just A way that still required improvement ….but still provided tonnes of useful info……….I personally would love to be kept up to date with where u go with this and hope that the naysayers don’t turn u off of sharing what u learn………..
Ah, sorry, I see what you mean. I’ve always done it that way (negative timeline) in max too. I’m just learning blender so I’m not sure about these things.
oh dude I had a matchstick in between my teeth and I drop it. Emotional punch line was ”
Therefore I switched to blender ” .
The following sentences are written by the machine translation.
Evolution doesn’t exist in the perfect one.
Therefore, there is the future in Blender.
His animation is not all of Blender. The one having worked within the range of the time having been permitted by the thing and man whom he understood was expressed. I think that it is good work.
Thank you and Amen to that :)
Thank you for this helping and nicely presented tutorial.
I’m in no position to evaluate it from a technical standpoint, being no animator myself. Anyway, I have a question and an observation.
What do you mean by saying that you made UV textures besides textures? It seems obvious to me at first glance.
The final part of the animation, when shirt conforms itself to the standing character, is way less realistic than the other parts. It looks like shirt’s time is more dragging than that of the other parts. I cannot say if this comes to me from the fact that I’m accustomed to Maya standard, and if some tweaking could resolve this issue in Blender. Anyway, it’s a fairly good animation work, altogether.
This isn’t a tutorial!!! It was meant to be a breakdown of how I achieved it and for my friends, I have no idea how it got here in a video tutorial section!
Thanks for the efforts to test this and put it into such a nice well presented tutorial!
I think that your video shows a nice workaround for frustrated users, and above all it shows one thing:
Blender needs a better cloth sim !
You are welcome :) And thank you :)
I disagree about the fact that blender needs a better cloth sim, it is very powerful, what it needs is simply to allow the entire processor (CPU) power to bake the simulation. I have an AMD Phenome x6 3.2GHz processor (6 core CPU) and baking seems to take only 1 of core out of all 6 as opposed to rendering which takes it all if I want it to, since I can specify how much core to use. If the baking process would have that, I would be able to make way faster test since my baking would be roughly 6 times faster, and that would allow me to find proper settings faster since there isn’t much info on the topic and I relying on my intuition and logic to test it blindly.
Wow this thread was a bit more confrontational than I was expecting after watching that vid. Good stuff in my opinion, I’ve done a couple of things with Blender’s cloth sim, no character anim yet, and I’m quite happy with it. Anyway kudos for following your dream of animation and thanks for the tips.
Taking time to reply to everyone is too time consuming so, thank you all for your awesome support :) Blender is most definitely the best 3D community on the web, also very very supportive and motivational ;) Blender just pure Rocks!
Awesome work Jeannot! I have a feeling I’ll come back to this someday. Thanks much for sharing the results of your efforts!
penduin, thank you :) here is part 2, basically just improvement on the previous simulation :)
http://vimeo.com/40976671
Hopefully, this one is better :P
Thank you for sharing. The tips will be very useful. I’m glad I have Blender as I would not want to even attempt to try to do something like this in Maya. At least Blender has simplifies the process of cloth simulation.
Super :)
Yes, that looks a lot better :-D
Lowering the spring constant helps a lot – it doesn’t behave like rubber anymore. Did you lower it to the cotton preset value?
I’m wondering if it would look even better if you increased the damping constants. When he lands, the cloth sort of bounces around a bit, but it might look better if it instead would simply plop into place … flumpf.
Great material there.
I think the simulation is doing everything right except for the time speed, wobbly effect sense may be due to the fact that real fabric goes back to normal faster, from what I can tell, this based on slow motion camera videos I have seen.
e.g.:
Youtube: Action Figure Slow Motion Punches
Looking at that reference there when the people get punched you can see their shirts don’t move that much, when they deform is quick a second on that video is miliseconds in realtime, if the faces of people get horrible deformed when a force is exerted but they go back to normal very quickly.
Maybe the simulation speed should be double the amount of the scene speed.
Just to complement, the simulation has to much detail, it looks like it is running in slow motion, remove some of the details and it should look great and realistic hopefully.
Love it!
hey there, the video seems to be taken down.