[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/3/ - 3DCG


View post   

File: 2.27 MB, 1920x1080, 00CFWPvcAW6JoZSo.webm [View same] [iqdb] [saucenao] [google]
648925 No.648925 [Reply] [Original]

what's a good tutorial, either as a book or video series, that covers making polymodels/rigs like this where they bend really cleanly even without blendshapes in maya? i can't find anything that actually shows how to make this work. it usually covers basic rigging techniques, which is fine, and that constraints exist, but implementing them for a wrist or muscle group seems impossible without morph targeting to the result. also there's no normal map for this either and i can't grasp how the muscle crests were modeled in. the artist said they used booleans but i still can't get it. is there anything in English that goes over similar methods? he used basic plains for the hands and feet from what i could read.

>> No.648940

Oh boy, yet another cheaply made rig made by yet another worthless horny loser who thinks he made a breakthrough.

Look into muscle systems and using influence objects.
More joints can also go a long way and they're game-friendly

By the way, the thing you posted fucking sucks ass.
Source: I do exactly this for a living.

>> No.648944

>>648940
You sound upset.

>> No.648946

>>648940
>Oh boy, yet another cheaply made rig
nope

>> No.648958
File: 523 KB, 367x219, 1540267506032.gif [View same] [iqdb] [saucenao] [google]
648958

>>648940
>discouraging newcomers to eliminate possible competition

>> No.648967
File: 225 KB, 373x327, edgelord detected.png [View same] [iqdb] [saucenao] [google]
648967

>>648940
posting to this thread only to increase the visibility of your retardation.

>> No.648978

>>648940
does houdini support influence objects?

>> No.649013
File: 80 KB, 637x488, DunningKrugerGraph.png [View same] [iqdb] [saucenao] [google]
649013

>>648940
Let me reality slap you, see that Dunning Kruger chart? Ive marked the spot where you are now based on that comment.

OP: There is many different ways to get there. I use a system called "volume joints". Its basically a set of up to 6 joints (one in each direction) positioned around two main bones, for example upper and lower arm. These volume bones move outwards based on the angle of their driving bones. The system can be made quite complex using ramp graphs to map the angle values different for each bone. Using different means of measuring the angles for increased stability and so on.

An other way for example would be using an RBF solver like Hans Godard demonstrated a while back.

Or use the maya Muscle system to drive bones on the surface of the muscles that then drive a skin cluster (so it stays game-ready).

Thats 3 different methods, which should be enough food for thought for now.

>> No.649076
File: 782 KB, 500x534, dgF7vebm5FsKOPfL.webm [View same] [iqdb] [saucenao] [google]
649076

>>648940
You're a fucking idiot.
But OP is a fucking idiot too for posting a shitty webm.
https://twitter.com/sakuramochiJP/media

>> No.649079

>>649013
can they be used simultaneously to obtain high fidelity?

>> No.649097

>>649079
I suppose you could do that. Apply different bones for the different regions and tasks.
My volume joint setup is mainly used to preserve volume on knees, elbows, shoulders, hip, belly, wrists etc. When done right it saves the hassle of doing corrective shapes simply to fix volume loss and bad skinning.
I also have a thing called stretchy bones which can be used to stretch over a larger muscle region like latissimus. Its just an aiming joint that scales to meet its end target, optionally with volume preserving scaling of secondary axes. Its constrained between key points and given certain limits to create the desired effect.
Anyway, the next step would be to have bones that deal with different muscular details or skin sliding and such. Similar to what I mentioned above. That however becomes quite a lot more complex with increasing demand for detail and you should think that through in depth where your goals are, what your target platform can support, what your time constraints are, and so on, before pushing forward.

In my projects im working with a game engine so I cant use too many bones. Thtas why I resort to volume joints and stretchy joints for the general body shape preservation and then make blendshapes for muscle detail. Those shapes then are automatically triggered in real-time by a bit of code in the engine.


Here is what my rigs look like in Maya, its an older demo but shows whats going on: https://youtu.be/7xfWAZ4Nz5w?t=65

And this is the implementation inside the engine, take a look at her abs and neck region to see how the muscle flex shapes trigger based on the motion:
https://www.youtube.com/watch?v=7p_HNOvsoUE

>> No.649125

>>649097
Don't blendshapes cancel each other out in game? Or do they blend automatically?

>> No.649138

>>649097
what are some good guides on volume/stretch joints for these tasks?

>> No.649160

>>649097
awww dang, been following you guys on youtube for a long time, you guys do awesome work, i hope your big project does well.

>> No.649170

Wouldn't a muscle system conflict with a mesh that has muscle contours sculpted into it? Or is that negligible?

>> No.649175

>>649125
How would they cancel each other out? Can you elaborate the thought process behind that question?

>>649138
Those are hard to come by. After a quick search I found this: https://www.rig-it.net/blog/maya-advanced-rigging-deformation-tutorial/
Its done by one of my former coworkers who knows his stuff quite well. I havent looked into it much but judging from the few descriptions on that page it should show more or less the sort of stuff im talking about here.

>>649160
Thanks my dude

>>649170
Depends on the muscle system.
If the muscle details are etched into the topology by using thigns like edge sharpness or so then the system wont work well when sliding skin over muscles as the skin with the muscle detail will travel over the actual muscle and create a wired look. Thats why you want to model things in clean evenly spaced out quads if they need to simulate muscles in such detail.
If the muscle system is limited to simple flex action then it should be fine as the underlying muscles only control volume of the body while the details can be smoothed in and out using an other deformer.

>> No.649177

>>649175
>How would they cancel each other out? Can you elaborate the thought process behind that question?
When you enable two (or more) at the same time, does the second one overwrite the changes of the first one our do they all add up?

>> No.649179
File: 181 KB, 868x458, V98-KneeRigDetails.jpg [View same] [iqdb] [saucenao] [google]
649179

Since this is a rigging thread, I thought I may ask here instead of the question thread:

Which system is better for performance in video games?
more bones, no blend shapes
or
less bones, blend shapes

I want to make a video game that is gonna have low poly models and hand painted textures, but many models in one scene with fancy animations.

pic not related to my project, but related to the question. Is it better for performance to have multiple corrective bones, or blend shapes/morths?

>> No.649182

>>649177
By default they add up. All a blendshape is is a list of vectors for each affected vertex to move the amount specified by the delta of the base mesh and the shape target.
So turning on multiple shapes will just add up the vectors for each target.

>>649179
Depends on a couple things, whats the target platform, does it even support blendshapes.
How lowpoly are your characters going to be? At a certain level you might be better off using topology tricks to fix some of the prolems.
What sort of game is it going to be, do characters have different clothings, different body shapes, etc, that you would need to reflect in the corrective shapes again and again?
What does
>fancy animations
mean?
Is the picture related? Is that the sort of polycount you are referring to?

>> No.649185
File: 433 KB, 756x749, HIGH QUALITY CG 09.png [View same] [iqdb] [saucenao] [google]
649185

>>648925
This artist is in another level, anon... let´s just say that he raised the bar, even for industry standards. And i´m talking hollywood/game industry here.
No book will give you the knowledge to do what he just did. He 'created' his own knowledge, so to speak.
His tweets are the greatest source of reference and technique you will ever find for Maya, no matter your pipeline - and it´s a 3ds Max user who´s saying it.
/3/ is infested with losers and trolls of all kinds, but this guy is the real deal - notice that he DOESN´T waste his time posting here.
If he did, he would receive a rain of shit instead of compliments of any useful advice - a guy like him would GIVE advice, not receive it, for starters.
/3/ is not useful even for technical advice anymore, /b/ raided here a long time ago.

>> No.649186

>>649185
i just wanna know what rigging method to use coming from both houdini and maya. i feel like it's easier in houdini but i can't set up helper joints as clearly, since it completely restricts you with using bones, but then its tissue system is leagues above maya's base configuration.

>> No.649187

>>649185
Ironically you are the only one not being useful here...

>> No.649188

>>649185
How can you even get that good? Just years of practice?

>> No.649190

>>649188
yeah and not being afraid of experimenting.

>> No.649193
File: 412 KB, 638x654, HIGH QUALITY CG 01.png [View same] [iqdb] [saucenao] [google]
649193

>>649186
Every software has its rough edges - i myself will try and model that Cammy anatomy and rig result with Zbrush and Max, and the techniques used will be quite different. I will be focusing on some aspects i found appealing - the muscled thin anatomy, some deformation bulges, etc. Pic related, his hair solution is awesome.
>>649188
Yup. And always, ALWAYS get only good influence on your artist skills. Bad art is like /3/ posts, or a filthy drunk bum in the streets: you ignore completely.

>> No.649194

>>649193
I'm thinking of using the houdini tissue system with a zbrush sculpt.

>> No.649196
File: 403 KB, 1171x725, large[1].png [View same] [iqdb] [saucenao] [google]
649196

>>649193
The guy is seriously great.

>> No.649199
File: 458 KB, 1024x768, Where's Wally scene.jpg [View same] [iqdb] [saucenao] [google]
649199

>>649182
>>649182
>whats the target platform
PC
using Unity
>How lowpoly are your characters going to be?
up to 2500 tries

>do characters have different clothings
just textures

It's less about corrective shapes more about having bones for fingers and mouth, or using blendshapes.

I want to have long idle animations for many characters on the screen

I want to create a game in style of "Where is Wally?" but from 3rd/1st person perspective, so given how scenes are gonna be cluttered with many characters doing their thing (although not as many as pic posted), I don't want to make a mistake of adding too many bones to low poly characters and defeating one of the reasons why I am going low poly in the first place.

>> No.649204

>>649199
Does unity have LOD system for skinned meshes? If so I would definitely think about using that. So you can add bones for fingers and mouth but only for the handful of characters closest to the player.

>> No.649208

>>649204
It has LOD for mesh rendering, I have no idea if it can be used for bones.
So should I just go with more bones and then look into optimization later?

I honestly have no idea how much impact on performance they would have in the first place.
I know games like Warriors series from Omega Force make obvious optimization choices when it comes to graphics, but I wish I new if more bones could be a big step back.

>> No.649293

>>649188
>>649185


His earlier posts from 6 years ago were not as impressive, but you can see his progress over the years.

He went from generic anime shit to pseudo anatomy realism.

His twitter history is a fucking goldmine,

https://twitter.com/sakuramochiJP/status/1023954680743772160

The insanity of it all is how he starts his meshes!

And its all poly modeling, not zbrush , amazing

>> No.649294
File: 865 KB, 2047x881, DjiPG20U8AAbK6C.png:large.png [View same] [iqdb] [saucenao] [google]
649294

>>649293

>> No.649295
File: 377 KB, 1376x480, DkAj-H7VAAETXIW.png:large.png [View same] [iqdb] [saucenao] [google]
649295

>>649294

>> No.649326

>>649295
His topology looks pretty horrible tho

>> No.649330

>>649326
I have to add: seeing the output views of his renders makes me think: fuck the topology, shit looks goooood

>> No.649331

>>649185
>>649293
Is this higher level shitposting?

>> No.649347

>>649294
>>649295
That looks disgusting

>> No.649357

>>649294
>>649295
>anorexic torso
>thicc meaty ass and breast
>garbage topology
but hey at least he nailed the fake bolted on tits, so..

>> No.649389

>>649326
I'm pretty sure those are work in progress shots.

>> No.649390

>>649389
Thing is everything on his twitter looks like that from the early days to now. So I dont think its because of wip. I think its just his "style".

>> No.649486

>>649326
>>649330
>>649357

This kind of topology isn't problematic, it's a kind of technique that is used exclusively for subdivision modeling. The triangular faces are actually folds and humps. It looks like they could be combined back into quads, but they're no longer planar. Basically, this technique lets you create different sorts of sharp folds and detailing that would otherwise require a lot more subdivisions. It does give you a pretty neat look without using any normal or displacement maps at all.

It does have some drawbacks though, suckiest one imo you is can't select edge loops anymore. It's a bit of a pain to paint skin weights with this kind of topo, and you also lose the ability to share body mesh topology between characters if you are trying to set up a pipeline.

I couldn't fap to those results but they are pretty stylish.