Drastic attention catcher, I know, I could replace “useless” with “working” :rolleyes:
For about two weeks I try now to find out how the heck one makes Dynamic Hair that does not intersect with the emiter mesh, or for that matter collides with anything at all.
Am I missing something vital or is it not working at all or not yet?
My post in the support forums, as usual on in-depth questions, remained unanswered and I guess the hits it got was from people that landed there googling, hoping to find something about that topic, I most certainly haven´t found any useful information on it yet.
And when combing particle hair, is there any way to comb it like you comb a head and not always pushing and pulling them through the “head”?
And where have my good old friends the particle reactors gone?
When Sintel was in production there was much talk on the mailing lists and in the blog about how it really wasn’t working very well…
there was a lot of activity and attempts to fix it and patch it up… i’d guess that any fixes that happened won’t be in trunk, they * might* be in the render branch… but they may be very “sintel” specific…
Your best bet is pobably trying to PM someone from the sintel team about it… (or on IRC)
EDIT: though you should probably read the durian blog as a first point of call!
That is really disappointing to be perfectly honest.
Blender wants to be an animation package and lacks such a vital feature…
And the current documentation is beyond bad and I wonder who is going to write it if no one knows how stuff works in the first place besides the devs ^^
Well, who needs hair anyways, my next chars then will be Humpty Dumpty, Mr.Proper, Skunk Anansie and Grace Jones… lol. should be a neat group picture.
I think that’s how this sort of development works, especially with beta software. It’s much better for the developers to be out squashing bugs than writing a manual that will need to be rewritten with 2.6. My guess is that once we hit 2.6 there will be a stronger push for documentation, and I think the non-devs (like myself) will need to take a big share of that load. This community is dedicated and generous; I think we’ll pull through in the end, in spite of the growing pains we might be feeling right now.
To be blunt, I really don´t care to workaround everything anymore - and besides that it is not about doing hair because I need hair, it is about using the dynamic hair feature =)
However I deem hair pretty vital, i´d say the majority of characters got hair and looking at the bigger animation studios, they all use dynamic hair.
So there is a particle hair option in blender. Works nice for fur, works “somehow” for a hairstyle because combing and stuff is a PITA.
And there is dynamic hair in blender. Works nice, is just compeltely useless because it doesn´t want to collide with anything and ends in dead spaghetti.
Why is it even in the current official when it is not working?
Why do you need to be a P.I. to find information on such things - if it is not working currently how hard would it be to edit the wiki and leave a note there that it is not working?
And up until now I dont know if it is working somehow, not at all, broken, will be discontinued or fixed later and all that although its an already existing feature in the official build.
I think that the dynamics was assumed to work, and probably did under simple test cases… but found lacking during Sintel… given the time pressures on that team it’s understandable that they didn’t update the wiki…
I’d agree that finding up to date information does require super human skills sometimes and that when it gets to the hardcore nitty gritty there actually isn’t any documentation. I’ve found that a couple of times.
I think part of the trouble is that the really interesting things don’t get put on the wiki, they get used for self promotion as video or web tutorials on individual artists websites…
A very natural thing, but they rapidly get lost and outdated and become almost impossible to find because of the huge volume of superficial to basic to just plain inaccurate tutorials on the subject that stay seemingly forever in internet land.
…or worse, once pipelines and workflows are discovered in depth they just stay in the artists head…
I can’t really complain as I’ve not updated the wiki when i worked certain things out mysel that aren’t there… so will hold my hands up and say “i;m guilty”, but the point of a wiki is to centralise information and that we can all edit it…
I guess that irc chats might be your best bet (though again that’s not something I’ve tried). i usually only go as far as confirming if the issue is known or not…and report to the tracker
Maybe in this case you could confirm on IRC the status of hair, find out if anyone "own"s that wiki page and then update it?
Anyway, I feel your pain on this, but with a pang of guilt that the solution may actually be us finding out and documenting it ourselves in the wiki…
A lot of code that gets committed to blender has not been well tested in production scenarios - often it’s just been fun demo videos on youtube. It’s not really a good thing, especially if you need to use it in a production scenario, but it’s also the nature of the beast, given that most blender developers are volunteers and aren’t working in production. The best you can make of the situation as a user is to test things well and give good, thorough feedback, at least that way the developers can be made aware of the situation.
Hair is indeed very problematic in Blender. For Sintel we used a brute force approach; Campbell wrote a script that ran sims with a wide variety of settings, then an artist reviewed all the sims that were produced and chose the best one.
Even then, well, you saw the results.
It is my understanding that dynamic hair is, and has always been, a very difficult issue to solve in 3d. Just look at this videoabout the making of Tangled/Rapunzel. Even with a huge budget and massive development team they still found it challenging.
@arexma: i gotta say, you’re sounding a bit like a tourist. If you want to fund a developer (or a team) to focus on this feature, please feel free. If you want to contribute some documentation, also feel free.
(By the way, the “current official” is 2.49. 2.5 is a beta)
I think this is one of the major problems Blender faces. There are dozends of kickass coders hacking day by day, and dozends of kickass artists working with Blender day by day, but there is not really a platform where they can exchange, it would have to be a tight collaboration, you can’t write some feedback as artist somewhere and hope the developer implements it some time and correctly, and while the coder implements stuff very technical, smoke for instance (ask Michael how he feels about it =) ) - top feature, the artist has to use it like a physicist would use a finite element simulation, and not like an art tool.
I think some kind of interface artist<>coder needs to be developed, a platform but I got no practical idea. IRC isn´t it. I´ve been there numerous time and the channels are ruled by tumbleweed. And life chat is kind of hard being spread from the states over europe and australia back to the states.
I feel a bit insulted, not too bad though, and to be blunt, I find your attitude strange.
Usually I don´t have the big time budget to contribute. However I can donate.
I donated to the BF, to BA, donated to smaller Projects, BGE Projects, Projects based on Blender work, Ocean Simulation, bought stuff I don´t really need from the BF store and work with Blender to earn my money. So I take the liberty to say I am far beyond tourist.
And how would I contribute documentation when I can´t find any information in the first place, when the only one who knows how it works is the dev that wrote it? =)
Last resort is to check out the source and figure it out yourself, but not everyone got that time.
So now there is a feature in Blender that does not work.
If someone points his finger on it you get told:
“Please, you are free to contribute or fund a team to develop it yourself, after all Blender is free software”
If someone has a decent feature that should be in Blender first response is:
“We don´t need that and if you really want it, fund a team to develop it or do it yourself” which will be responded with:
“Thats why Blender will never be taken serious in the industry because it lacks vital features”
A “circus viciousus”.
To clearify, I don´t really want nor need dynamic hair, at least not at the moment, take all the time it takes for what I care to implement it.
What I decry is why there is a feature in Blender, in the “official” (after all 2.5 is heavily promoted) build, and with official i mean the one from blender.org, not some heavily under construction svn version, that isn´t usefull in any kind, while it works somehow you can´t use it and then you can´t find any information about it - there is a general lack of information.
And why is there no path established so dev´s and artists can work together better, with a free piece of software you have the chance to tailor a tool around the users, while currently tools are created and the users have to use workarounds to use it properly. I understand that 2.5 is still in beta stage but look around the forums.
More and more professional users disappear from this forum, others have problems with simple stuff and try to find workarounds or hacks to make stuff work that should work out of the box, and If not yet, why is there no place, easy to find where the artists can suggest changes, or get information if this is already working as intended. No one got any real information, just suggestions and assumptions in the average help with XYZ thread if its beyond basic stuff.
And again if someone raises the topic of a lack of documentation or broken features, you don´t get a discussion going on how to improve the situation, all you hear is:
“Contribute yourself, or raise team to do so…”
And don´t get me wrong, 2.5 is grand, I love it, the new features, context sensitivity and overhaul, I just feel that due to the heavy work under the hood other aspects got neglected.
Well… that’s a pretty pathological case. They were facing new challenges, for obvious reasons…
Hair simulation is still certainly a topic of research, of course. But even so, there are much better existing solutions than what Blender has. Blender’s hair simulation isn’t even based on hair simulation research, it’s based on shoe-horning the cloth simulation code.
One of my favorite DVD commentaries is Brad Bird during The Incredibles- he mentions a few times the hair/cloth simulation, how they had to set up crazy animated parameters to make it look right, and how amazing and wonderful it was when something actually worked! That was 7 years ago, and hair/cloth today is still a challenge- that says something about the complexity of it.
That is not really the worst approach.
In Maya for instance you can make planes which you make cloth-simulated and use them to “guide” the hair you had setup earlier, which doesn´t give too bad results.
I haven´t figured out yet, or was too annoyed not getting anywhere with the hair yet to try it, if it is possible in Blender - but maybe thats exactly how hair currently works.
And if the hair is heavily based on the cloth simulation I wonder why it can´t collide then, because cloth´s working pretty good.
If you create a head uvsphere and then a few curves for hair guides, duplicate them, make the duplicate a mesh and a cloth that collides with the head, then create a gazillion of hooks on the guides and create empty vertex parents along the cloth with a track to constraint to the mesh base to be able to copy the rotation and then copy location and rotation of those empties with the ones where the hooks are attached, you got a curve that moves like the cloth mesh and can guide the hair.
2 Problems:
the pinned part of the curve is the end, which can be changed somehow i presume, else the guide guides the hair inside the head.
I used dynamic hair in a fifteen second animation shot I did for a not-yet-published book, and it worked adequately for me with a little bit of tweaking and careful choice of parameters. Not sure what your requirements are though. I’m going to guess that really complex stuff isn’t going to work without an immense amount of time put into of. Of course, that’s the down side of simulation. What you gain in physical realism over hand-animation, you lose in controlability.
Also, as for the “but there is not really a platform where they can exchange, it would have to be a tight collaboration” comment, I kind of thought that that’s what the open movies did. If you’re a working production artists and using Blender in your studio, developers will be more than happy to listen to your feedback and collaborate.
Im working on a spider rig scene, where the spider has dynamic hair which sould animated too, but it doesnt work . i will provide scene in the bugtracker.
None really, it is broadening my horizont as I got air between projects, then I play with blender features till my eyes bleed. However:
I want an uv-sphere with hair on top. Beneath a cylinder with its height horizontal to act as shoulders.
Now I want the dynamic hair only not to intersect with the head and collide with the shoulders.
When I style particle hair and comb it, I want to comb it on my head (uv-sphere) and not push and pull it through it.
And I want to know if you can actually create a hairstyle and use dynamic hair on it without destroying the hairstyle.
I don´t know how to do accomplish it, 2 weeks of trail and error are enough, don´t know where to ask (support forum was a fail), nor if it is even possible with the current codebase =)
The open movies are great sure, but I think the development happens too fast especially as 2.5 is a milestone in blender history - 2.5 was born with durian and features implemented on the go. Now we got a brand new package but there are already talks about mango and it being a vfx project.
Why can´t there be another animation project with 2.5 where the features and workflows that have been worked out during durian are perfected?
I am, but so far I am happy with Blender, the features I need for work, work quite well, besides some special requirements, which I see as personal specialized requirements no one else would need, and then I work around it.
But if: Where do I find the ear?
It is rather to find out for instance about the dynamic hair. All I tried to find out was if it is possible to collide it with other objects.
It seems impossible to get a “no, because:”, “yes, simply do:” or “not yet, beacuse” somewhere. And here in this thread we got sintel team members, BCT, coders and professional users.
I was also asking where the particle reactors went to in 2.5 - if they are even there…
The wiki is empty. blender.org forum has the question from someone else since february - unanswered
Jahka´s blog said in january he´ll do a tutorial on reactor particles once he has implemented them… judging by that I guess they are not there yet?
I didn´t bother to search any further.
I also wanted to know why it isn´t possible to create a driver for the lens of a camera, it simply doesn´t work.
No answers.
IMO there is a general lack of information/documentation that is easy to access or reachable.
And if you ask for it, you get as reply feel free to contribute and fill the wiki, everyones busy… yeh… fill with what?
So basically we got an incredibly(surprisingly for all that was done) stable Blender 2.5, where it seems some features are there but not implemented working, some are missing - beyond that, its fully usable in a production.
The problem now is, we got no documentaion whatsever and are left to trail and error and eventually start to fill the wiki?
I was also asking where the particle reactors went to in 2.5 - if they are even there…
The wiki is empty. blender.org forum has the question from someone else since february - unanswered
Jahka´s blog said in january he´ll do a tutorial on reactor particles once he has implemented them… judging by that I guess they are not there yet?
I didn´t bother to search any further.
I also wanted to know why it isn´t possible to create a driver for the lens of a camera, it simply doesn´t work.
No answers.
A lot of info is available but you need to know where to look and have to keep on top of the revisions, something that most users can’t realistically do. For example reactor particles are still on the todo list as well as many other currently missing features.
Also things are constantly changing, drivers for camera focal length ?, just tried it with a recent build and it works.
Ah, tyvm. I wonder why I never stumbled upon the todo-list regularily randomly browsing blender.org. Reading along it makes one almost cry. 2.5 feels so complete and is far from it… lot´s of work ahead.
Most recent as in svn? I use the one from blender.org and either I am doing it all wrong, although other drivers work, or lens driving was just broken then ^^
Enough ranting for today… I need to get into a better mood… but I feel better already venting here lol And I don´t want to go back to 2.49 pouts
I wonder though, the to-do says “better handling of collisions & force fields for dynamic hair” yet I am completely unable to accomplish any collision at all…
And I don´t want to go back to 2.49 pouts