PDA

View Full Version : True Skin & Muscle Simulation in LW...is it possible?



robertoortiz
02-28-2015, 10:26 PM
So guys i am curious to see if you think if true muscle & skim simulation are possible in LW for more accurate skin deformations.



I am looking forward to your comments,
-Roberto

jasonwestmas
02-28-2015, 10:40 PM
Sure it is, but that's not an easy thing to code. I think the makers of syflex could come up with something for Lightwave but we'd have to ask them. http://www.syflex.biz/tut_skin01.html

robertoortiz
03-01-2015, 03:07 AM
Sure it is, but that's not an easy thing to code. I think the makers of syflex could come up with something for Lightwave but we'd have to ask them. http://www.syflex.biz/tut_skin01.html
Maybe with a custom Genoma rig...if the lw group would play ball with them.

jasonwestmas
03-01-2015, 09:05 AM
Automation is absolutely necessary for this to be practical. In the article the Technical Artist setup was to model all the muscles; bind them to the skeleton and then deform them before any syflex skin simulation could take place. For bulging, SLIDING and deflating the TA used morph targets (blendshapes) and something called a sculpt deformer which looks like editable wrap deformer.

https://www.youtube.com/watch?v=8X4aXave-mg

To get the muscles to slide auto-magically I think we would first need a wrap deformer and a way to write some simple expressions for it. I have no idea how that would be done in Lightwave 2015 seeing as deformer tools are kinda limited to "sticky" deformers like linear endomorphs and bones. We need a sliding mechanism of some sort that can be automated according to the bone rotations.

For muscle Jiggle, the TA used dynamic splines that use a wrap deformer. But now there is syflesh available for maya and max.

jasonwestmas
03-01-2015, 09:29 AM
Aha, here is a nice nodal setup for auto-sliding. Pretty cool RH, it looks like you found a way to use the new spline IK stuff with joints? https://www.youtube.com/watch?v=6Xz9dhe-U78&list=PLTds3QePYrWGUd7fMuNAivIsORvpnKDqX&index=21

So the idea of course is to deform the muscle system using various techniques along with a SyFlesh sim. . . and then run the SySkin simulation on top of those muscle deformations as if it were a tight fitting shirt.

jasonwestmas
03-01-2015, 09:42 AM
In the sprit of nice "dynamic bone" deformation options. . .Here's a nice system for creating hold bones on the fly and automating that as each bone rotates. https://www.youtube.com/watch?v=0gGVVss_C1k&list=PLTds3QePYrWGUd7fMuNAivIsORvpnKDqX&index=20

RebelHill
03-01-2015, 09:43 AM
Cheers...

And yeah, using (active or inactive) stretchy bones as a base for effectors gives a great effect, which can be further extended and customised into joint rotations through nodal, cyclist, etc. No simulation overhead needed.

https://www.youtube.com/watch?v=6Xz9dhe-U78&t=05m27s

jasonwestmas
03-01-2015, 09:56 AM
No simulation overhead needed.



Which is the best part of your setup there. Doing a sim to drive a Syflex sim, ugg, that would be slow and or tedious.

jasonwestmas
03-01-2015, 10:56 AM
Here's a nice way to model a muscle, LWBrush would be able to do this nicely. https://www.youtube.com/watch?v=dWifw8yigR8

A good example where painting weights in layout would be extremely useful. . .http://youtu.be/dWifw8yigR8?t=15m40s

robertoortiz
03-01-2015, 02:30 PM
Let me post some videos to help the conversation:

Proper Muscle for Biceps
http://youtu.be/AD0qcmqn8wo
Back Muscles
http://youtu.be/puoOqYeeHoY

WETA Digital tissue System:
http://youtu.be/7VlthWa5pu8

jasonwestmas
03-01-2015, 03:21 PM
I didn't have sycloth a year ago and I just got a new computer so I am very curious what modern hardware can do computationally these days inside of lightwave using nodal motion, bones and splines with sycloth. I think I would have a better idea what we would need if I tried to build something realistic, otherwise it's just talk. :) Main reason would be to find out what details like sySkin and syFlesh would enhance the most as opposed to just using syCloth. Perhaps I can get syflex demo for maya and make a comparison.

RebelHill
03-01-2015, 03:33 PM
Let me post some videos to help the conversation:

I don't see how any of that helps at all.

Btw... whatever happened to that whole "12 principles centred tools" thing you were so hot in the biscuit and full of ideas for? Pro tip, finish one thing before starting something new.

robertoortiz
03-01-2015, 07:15 PM
Ok taking a break from this great conversation...

Thanks for the heads up and your concern. I am busy right now on paternity leave with my newborn son.

Thanks and you will be the first to know when I reopen that thread.
Anyway I have not given up on my ideas, I just have sharing my ideas outside the lw community offline.

Ok let's get back on topic.

robertoortiz
03-01-2015, 09:16 PM
Here is a cool video on muscle simulation inside native Maya.

It is quite amazing:
https://vimeo.com/46429322

03-02-2015, 04:44 AM
Ya know,
I do believe it is so amazing that it is patented, and thus, not to be replicated in this fashion by any other without permission.
As well, I believe AutoDesk's IP base prevents most other apps from picking up and including this kind of amazing stuff. Or, at least in a manner that can be shown to be like theirs in court.

Ya ever wonder why other apps don't have the bipedal walk system like they have. Yes, similar but not enough and not as easy as ad's.


I don't think a Lightwave artist would/could get paid for the time it would take to do all that rigging. That is to say, LW tends to be in a faster-paced production environment that would rather fake it and look good with a little time; the others tend to be team oriented over a long(er) period of time.

There are LW artists that have achieved the look. They don't get very warm receptions, it seems, in public places as some deride their efforts because, "there are easier ways to do that than lw ..." or similar sentiments. (Edit: These are the exceptions to the above, or so I hope they get paid for their Herculean efforts.)

Context is key; put this question into some context. I can't see the point of whether it can or can't if most of us aren't going to use it.

Ryan Roye
03-02-2015, 05:25 AM
Just a note here, there are some working examples of using dynamics for skin simulation in the 11.5 LW3DG content files. I don't feel they are super high detail examples, but they are examples nonetheless (pirahna alien, Mangalore soft face). The exact same concepts can also be applied to Syflex.

Of course, if you want even more options for rigging and editability in the mix, Cagedeformer beats everything out there.

jasonwestmas
03-02-2015, 07:23 AM
Ya know,
I do believe it is so amazing that it is patented, and thus, not to be replicated in this fashion by any other without permission.
As well, I believe AutoDesk's IP base prevents most other apps from picking up and including this kind of amazing stuff. Or, at least in a manner that can be shown to be like theirs in court.

Ya ever wonder why other apps don't have the bipedal walk system like they have. Yes, similar but not enough and not as easy as ad's.


I don't think a Lightwave artist would/could get paid for the time it would take to do all that rigging. That is to say, LW tends to be in a faster-paced production environment that would rather fake it and look good with a little time; the others tend to be team oriented over a long(er) period of time.

There are LW artists that have achieved the look. They don't get very warm receptions, it seems, in public places as some deride their efforts because, "there are easier ways to do that than lw ..." or similar sentiments. (Edit: These are the exceptions to the above, or so I hope they get paid for their Herculean efforts.)

Context is key; put this question into some context. I can't see the point of whether it can or can't if most of us aren't going to use it.

Yes, project context is key, which is why I don't worry about general statements about what is the easier application or plugin to use. Sometimes I like to load up lightwave because there is a more direct path in there. Sometimes I need more control/performance but not a huge nuclear reactor (which requires more maintenance) to power it. That's where affordable third party plugins for lightwave come into play. Also some of us like to use Lightwave for our personal projects, conceptual previz and other studies, not because we want to get paid with it.

ianr
03-02-2015, 07:59 AM
Anything that drives us towards Weta's Tissue System is a good thing.
Look what Marashiro's done with ' things lying about on the floor'
So to speak. Breathed new life into the way for LW 2Go, with his
3rdPowers appley named plug-ins.
Why ,I swear that the ghost of Chuck Jones ( a Hero of mine) did visit him
in his sleep one night!

jasonwestmas
03-02-2015, 08:35 AM
Just a note here, there are some working examples of using dynamics for skin simulation in the 11.5 LW3DG content files. I don't feel they are super high detail examples, but they are examples nonetheless (pirahna alien, Mangalore soft face). The exact same concepts can also be applied to Syflex.

Of course, if you want even more options for rigging and editability in the mix, Cagedeformer beats everything out there.

Thanks Ryan, I had forgotten about those. Examples are indeed scarce in general for this kind of stuff.

RebelHill
03-02-2015, 09:00 AM
Anything that drives us towards Weta's Tissue System is a good thing.

Well... that would probably have to include about 15 years of development and a couple hundred million dollars. Anyone wants to give me those things, I'll get your muscle system for you, no problem. Ofc, by the time it's done, and available "off the shelf", the big boys at the like of weta will have moved on to something even better still, and it'll just look like old hat, and we'll have the usual suspects citing the top level, all new "super realtime everything" system as "inspiration".

Honestly... it's like saying...

The hour long commute to work is a real problem eating into one's personal time... when you factor in the school run, it's nearly 2 hours each way, every week day. how can we improve this situation?
Here's a video to help the conversation along... https://www.youtube.com/watch?v=OC00qHonfj8

robertoortiz
03-02-2015, 10:16 AM
More on the WETA Tissue System:
http://www.fxguide.com/fxguidetv/fxguidetv-166-weta-digitals-tissue-system/

And here is are some cool papers on the subject:
Auto Muscle Tool for a Bipedal Character -
http://ecollections.scad.edu/iii/cpro/app?id=7331637339768542&itemId=1001845&lang=eng&service=blob&suite=def

Hand Tendon! simulation
http://www.cs.ucr.edu/~vbz/cs231papers/sueda2008.pdf

Muscle and soft tissue simulation within a character pipeline based on Autodesk Maya
http://jo3d.com/MasterThesis.pdf

ianr
03-02-2015, 11:31 AM
Yes, I could run a team to do that as Well !

I might buy an island & do Thunderbirds Again?

But your Usual 'Zen' accuracy has Slipped R.H.

A jaded slight exaggeration in Time & Money

you have put down here, I was only trying to be positive.


I hope that you can see that & don't Flame Up.

RebelHill
03-02-2015, 11:49 AM
Sadly... its not an exaggeration in either respect. Weta (and other big boys) have had proprietary muscle sim systems going back 15 odd years... Their tissue system (and the systems of others) didnt just spring into existence in isolation, they're just the evolution of those which they had before, and stem from know how and techniques that have been built up over many years. There's a reason that you only see the big companies like weta, ilm, et al producing this level of toolset, and its because of the long periods of time that they have to grow them, their focus on high end, very specialised tools, and their massive multi million dollar budgets per film.

Its really a very simple equation... How do you get tools that good?.. With resources that large.

robertoortiz
03-02-2015, 02:37 PM
This is a muscle Simulation video done by a student in SCAD
Maya + MEL
https://www.youtube.com/watch?v=uWwDMCAthOc

And this other video was done by an independent rigger.
Also in Maya

https://www.youtube.com/watch?v=pKWpZSfFwrA

jasonwestmas
03-02-2015, 02:53 PM
I've always liked these spiderman 3 maya muscle examples.

https://www.youtube.com/watch?v=fNCB9wZP8k8

silviotoledo
03-04-2015, 06:06 PM
Muskeelar has a free code. We need someone to port this for lightwave :) https://www.youtube.com/watch?v=x0akFtKiayo

Blender have muscles

now unreal also does https://www.youtube.com/watch?v=qOgj47gck_0

Davewriter
03-04-2015, 08:51 PM
Well... that would probably have to include about 15 years of development and a couple hundred million dollars. Anyone wants to give me those things, I'll get your muscle system for you, no problem. Ofc, by the time it's done, and available "off the shelf", the big boys at the like of weta will have moved on to something even better still, and it'll just look like old hat, and we'll have the usual suspects citing the top level, all new "super realtime everything" system as "inspiration".

Honestly... it's like saying...

The hour long commute to work is a real problem eating into one's personal time... when you factor in the school run, it's nearly 2 hours each way, every week day. how can we improve this situation?
Here's a video to help the conversation along... https://www.youtube.com/watch?v=OC00qHonfj8

Yea... but what kind of mileage does it get?
Kool! But I'd be afraid to touch the damn thing. I'd have it broken before shifting out of 2nd

jwiede
03-04-2015, 09:58 PM
Muskeelar has a free code. We need someone to port this for lightwave :) https://www.youtube.com/watch?v=x0akFtKiayo

Where did you see that Muskeelar (and its algorithm(s)) is/are free/OSS? I saw no mention of that, nor any links to src code. The paper is published, but that's not the same thing obviously.

robertoortiz
03-05-2015, 09:46 AM
Great point about the context on how this could be implemented.
Ok imagine a next gen 'genoma' rig, that allows for some basic muscle dynamics over a basic biped skeleton. The advantage of this sort of simulation would be more realism, thanks to amore precise musculature reaction.

m.d.
03-05-2015, 11:45 AM
not sure what genoma has to do with this...

it's a scripted rigging system....what you are talking about is more in the realm of modifications to deformations in Lightwave.


Realistically the best hope we have for (semi)realistic skin in the near future within Lightwave will likely be provided by syflex

The developer told me they plan to implement syMimic within LW at some point

syFlex has 2 approaches in maya...
sySkin is a 'sandwich' approach creating 2 or more layers of cloth with springs in between...the bottom layer attached to the deformed object
syMimic creates a single cloth layer with springs to the deformed object

there are definitely more advanced muscle/skin simulations out there...but this has a good chance of actually being implemented within lightwave

jasonwestmas
03-05-2015, 01:16 PM
The thing with setting up dynamic muscle groups is that they need to fit the volume and anatomy of the model in a precise manner. So really anyone doing this would have to model the volume and placement of the muscles at some point. I guess a genoma rig could come with a quick and easy way to attach muscles to the rig, but again, each muscle will need the ability to model and weight map it to the rig and then define the associated effectors anyway.