[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vr / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/3/ - 3DCG


Thread archived.
You cannot reply anymore.



File: channel_box.jpg (66 KB, 800x355)
66 KB
66 KB JPG
Should I learn to rig in houdini or maya?
>>
Maya has always been the king or model rigging. Houdini started as the gold standard for VFX particle sim, and has over time rounded out it's toolset to include decent rigging and hard surface modeling features.

If you're doing pure character animation use maya. If your character is going to have something specialized like muscle deformation or belly fat sim, then houdini might be better for the rig. It depends on what your going to be doing.
>>
>>696607
Houdini is not normally used for rigging or character animation. As >>696608 explained, it's for procedural generation and effects like physics/smoke/etc.

You certainly CAN rig in Houdini, but it would be a strange choice. Maybe in 5-10 years, Houdini will become the new standard for everything... but right now, I'd stick with Maya for regular rigging and animation.
>>
>>696608
I want my character to have flesh tearing and shit. I figured houdini fem solvers would be best, but is Maya Lacking there?
>>
File: no.png (637 KB, 693x693)
637 KB
637 KB PNG
>>696607
AHAHAHA rigging in Houdini
>>
>>696633
lol why not?
>>
>>696659
Can't compete with the one and only gold standard for rigging and character animation.
>>
>>696663
Yeah but standard sucks at everything else
>>
>>696664
animation is like 90% of 3d. The other 10% is zbrush and painter
>>
>>696667
The fuck u talking about? There's also cloth, soft body, fluids, fem, hair, and so on. Tons other shit if you arent making a basic deformation too. Like CD mfm for dynamic fracturing. Neo hookean skinning.
>>
>>696668
>There's also cloth, soft body, fluids, fem, hair, and so on.
>basic deformation
>fracturing.
that all falls under the animation category, assclown
>>
>>696669
do you actually think an animator knows how to set up an FEM solver? i know you aren't shitposting that hard.
>>
>>696669
ALSO, maya has no decent set of those tools at all built in, so its a moot point. saying it's part of the category means nothing if you can't implement it effectively.
>>
>>696672
you can implement whatever you want. there is a reason why tissue is built into maya
>>
>>696675
that...you're not being real right now are you? that shit is so primitive that it's essentially just deformation cages in the form of capsules. it has no consistency in any other regard like dampening, and trying to configure that is unrealistic in terms of parameter types. it's just not in the same realm as houdini, where not only you could configure what parameters you use, but you can just flat out implement your own coordinate space that they influence.

i get that maya is a standard, but rigging has gotten so good in a lot of packages that it's more or less homogenized. the only exception is that houdini uses bones which means you can't have rig components in isolation for doing certain techniques to aid in the workflow, but i think the necessity for it is debatable if you can utilize a superior workflow for every other step of the way, because it helps you focus on how to minimize the need for those shortcuts.
>>
>>696677
>the most advanced muscle deformation system ever seen in movies
>used in the still #1 grossing movie to this day, Avatar
>so primitive

kys
>>
>>696679
it's just FEM solvers applied to every single part of the model. that's it. if anything, gathering the data and modelling/scanning the geometry is what's impressive.
>>
>>696680
stop shitposting kiddo. It looks good, its proprietary, you havent used it
>>
>>696681
Why are you so defensive? Do you have weta stocks?
>>
>>696685
weta is one of my favorite studios, get over it
>>
>>696686
Lol so you're just dicksucking. K then.
>>
>>696687
whatever kid
>>
>>696688
*slurp*
>>
>>696689
kys
>>
>>696690
*gulp*
>>
>>696663
Sure, Maya is the standard in those areas. But is it technically superior? If so, in which areas?

I'm not asking for the sake of it. I've seen a few statements regarding Maya vs Houdini for rigging and animation, but wasn't able to find a real, somewhat detailed comparison. Not just "joints vs bones lol", but an explanation on why (if at all) joints are superior to bones, and in which cases they show their advantage best.
>>
Parallel rig evaluation is the only maya rigging technology I wish Houdini had.
>>
>>696712
Yeah, it's a game changer.
Having seen the rig performance of C4d and Blender, it's quite nice to see my character rig running with 120fps.
I don't even have a monitor that runs at this framerate (so i cap the max frames to 60) but knowing that there is 100% more performance than what i actually use is astonishing.
>>
>>696679
Weta's tissue system is Maya-based, LARP more kid.

>>696711
First, see >>696712
And look, it's as clear as day to anyone who does rigging for a living that bones are bullshit. Both as logic and for an ever-changing production. You're free to propose why you think bones may be superior, and I'll be glad to tell you why that's bullshit. If you're just trying to find an excuse to say Maya bad lol, then at least I won't have wasted my time with a LARPer.

And at the end of the day, even if every Maya functionality was cloned into Houdini down to the button labels, computed faster, in parallel and even if every existing Maya-based rig, pipeline, tool, plugin magically became Houdini-compatible and retro-compatible thanks to pixie magic overnight, there is still one insurmountable obstacle.
ANIMATORS.
Even in this make-believe case, animators would still stick to Maya, because they are THAT daft. You literally need a new generation of Houdini animators to come around before anyone even thinks of switching to Houdini for rigging.
And since there are no rigs for Houdini but plenty for Maya, nobody will learn Houdini as an animator.
So you'd need either an animator able to create his own Houdini rigs or a Maya rigger able to animate his own Houdini rigs to pave the way.
Fucking good luck finding anyone who fits either description.
>>
>>696771
>You're free to propose why you think bones may be superior
I can't, because I'm just beginning to learn rigging and know of no reasons why bone rigging could be better than joint-based rigging (or the reverse). All I can point to is this, from Houdini's docs:

"Joint based systems provide the flexibility and control of being able to move joints freely; however, improper joint orientations often produce unwanted artifacts which require post-processing. Bones, on the other hand, have an orientation. They always point down the -z axis, while up is the y dimension and x is the flat dimension. Bone hierarchies can also understand the whole bone chain, which makes it easier to squash and stretch rigs with bones."

I wish there were an extended article comparing the pros and negatives of both approaches, but I couldn't find any. In any case, I plan to also learn rigging on Maya, so I guess I may end finding the answers by myself. It would still be interesting to read/watch a good comparison, though.
>>
>>696771
>You literally need a new generation of Houdini animators to come around before anyone even thinks of switching to Houdini for rigging.
Maybe you already know of this by the way, but Outside's trailer has been animated in Houdini: https://www.youtube.com/watch?v=6YpWBIoR5mk It was done in collaboration with SideFX and they had to come up with their own solutions, of course. But it may signify a future for character animation in Houdini, if the egg-and-chicken problem you mention somehow gets solved.
>>
>>696609
>As >>696608 explained, it's for procedural generation and effects like physics/smoke/etc.
That's what I thought Houdini was. Procedural Generation of Physics and things like smoke/fire. It doesnt seem like a program that you would normall model/rig/animate in.
>>
>>696818
I just can't see it being easy to import a rig or export simulations. So I figured houdini would be worth mastering overall just due to flexibility and performance for dynamics and also rigging characters with those dynamics.
>>
Cool a rigging thread.

Learn Maya because it's the standard for rigging and animation. There's plenty of resources for it. Also do yourself a favour and double check your linear algebra knowledge.

Regarding joints vs. bones you end up needing joints if you want to make something production quality and animator/tools friendly. Being able to control orientation and rotation planes is essential. It can take a while to really appreciate the flexibility they provide. The Houdini devs don't really seem to understand rigging.

Maya's performance still sucks balls though. I increasingly hate the transform stack and eagerly await the day I can totally bypass it. Their constraints implementation sucks too and desperately needs a rethink. Not that Houdini is any better. Rigging is still a dark art requiring a ton of custom tech to do it properly.
>>
>>696834
Right now I'm in a tough spot trying to figure out how to place helper joints for things like the glutes and pelvis, but there aren't a lot of tutorials for that. I'm going off a creature rig tutorial that has a spline ik and space switching system, and possibly something similar to pic related with the thigh double bone to help with twists, but I'm lost on muscles because I can't decide if I want to blend shape them, or use bones. I don't mind sculpting but I can't tell if it will conflict for extreme poses, since each axis might need it's own and I don't know if morph targets can overlap like that well enough for a production rig. I was also gonna try out deformation cages but Id like to stick with joints for now.
>>
>>696834
>Being able to control orientation and rotation planes is essential. It can take a while to really appreciate the flexibility they provide.

I've never used bones before, could you provide a clear example of how the joints give better rotation plane control? I don't have the experience to understand the precise nature of the advantage joints give and would like to know.
>>
Fucking 4kike eating my posts like baby foreskins yesterday.

>>696790
Yes, I'm familiar with the Houdini docs.
Here, let's break down the bullshit.

(1/2)

>however, improper joint orientations often produce unwanted artifacts which require post-processing
Which is true, but given more weight than it ought (or rather, less), since what makes a rig a good rig (and by consequence, a good rigger) basically boils down to proper joint orientations. It's as useful as arguing that if your're bad at your job you will obtain bad results. DUH.

>Bones, on the other hand, have an orientation.
And a length, and two indissoluble ends.

>They always point down the -z axis
There goes animation retrocompatibility out of the window. 99.99% of Maya rigs are built with joints pointing down the +X axis. The geniuses could've at least kept that convention.
And:
>always
You don't ''always'' need a joint to point down another. And when they do, the don't ''always'' need to point down the -Z axis. Or the +X axis. They only need to be properly oriented for the geometry they're influencing to move correctly. Proper joint orientations, hmmm... sounds familiar.

And the most ironic and moronic thing is that bones also have an orientation. Or rather, an up vector.
>up is the y dimension
Yeah, but ''where'' is up?
It often isn't "world up". And not always +Y.
A bone still needs an explicit up vector, or it will be free to roll along its Z axis.
Which basically means that you still need to care about proper joint orientation with bones as well.

>which makes it easier to squash and stretch rigs with bones.
True, but that's because each bone is aware of its own length and two ends by its own design.
A feature that's only really useful for cartoony stuff, and only made "easier". It doesn't mean it's impossible in Maya, just slightly longer to setup, which means literally nothing unless you're not scripting your rigs, in which case you're either learning or out of a job.

(cont.)
>>
>>697033
>>696790

(2/2)

And if I need a single joint to drive anything, I can't, because I'm stuck with two.
>but a single joint alone is just a locator with some extra attributes
Technically true.
But I can't skin my geometry to a locator/null/empty/whatever and my game engine (read=''every'' game engine) only reads joints and skin weights.
And I'm only talking about driving a simple geometry in FK. If I need a single joint in the inner workings of my rig (and you do, every single time), I still can't have it.

You're basically giving up several degrees of freedom in order to eat shit right out of SideFX's anus and whatever their devs thought "would be best".
The manual also says you can emulate a joint-based system in Houdini doing some awkward shit setup.
I'll one up them and tell you you can also emulate a bone-based system in Maya.
Spread two joints apart, parent them together, and either have the child aim-constrained or create a single-chain IK. Now build a rig only using a bunch of these, even when you only need single joints. Good luck holding on to your sanity.

And, most importantly, you give up
>the flexibility and control of being able to move joints freely
which is what happens on an hourly basis in any production.

>>696791
Which is as useful as Blender-funded shorts are useful to convert professionals and future-professionals to Blender. i.e. very little or not at all.
>>
>>696991
Joints let you change the rotation planes independently of the translation planes, i.e. your up vector doesn't have to point down a translation vector.

For example if you have an ik chain that isn't in a world plane, like the whole thing is rotated. You'd want the rotation plane of the end effector aligned to the ik plane otherwise you'll end up countering rotations. But it might be the opposite case for translations if you want to translate the end effector in world space, where you'd end up having to counter translate in at least two axis if it was aligned to the rotation planes. Joints let you dick with the orientation to achieve things like that.

Examples for this are things like legs (particularly quadruped legs) which are sometimes posed in an A, depending on how the (anatomical) bones sit against the rib cage.



Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.