Hair & Fur simulation in Ornatrix

Discussion started by veer

Well,

It's been a few days now that i'm trying to play with ornatrix simulation (ox dynamics) and sincerely I didn't manage to nothing.

I'm using Ornatrix 3.1.2.0 with 3dsmax 2014.

My workstation is a powerfull i7-3770, 3.90GHz with N500GTX Twin Frozr II.

I watched the 2 related tutorial videos many times and even a basic sphere as showed in example don't give me the expected results or simulate diferently from one attempt to another.

Some parameters works once and trying the same later gives me another result: i'm really getting insane ! It works well sometimes but for my case it's driving me crazy.

I even started with combing my hair and spent a bit time to stylist them. I loose the main shape of my hair as soon as i start the simulation (even with very low global limit) and sometimes hairs become crazy without any logical explaination.

First question about all this is how MassFX engine handle all this:

I know you can choose to use Multithreading in MassFX options; in my case it don't change nothing...

I also tried cloth simulation for my case. but no differences !!

Please Light me up with this thread.

Thanks,

Answers

Posted over 8 years ago
0

Maybe try the demo of this one? (http://www.hair-farm.com/)
They claim very fast rendering and gravity compensation for preserving modeled hair shapes. In this demo video the hair seems to return perfectly to its original shape after movements (https://www.youtube.com/watch?t=1&v=TDh_gSPmysE).

Hair farm also presents some novel features (most notable)to model hair shape by so called hair meshes (http://www.cemyuksel.com/research/hairmeshes/), you should see the video. It's clear this technique also presents some efficiency opportunity's for the dynamics simulations.

Maybe the techniques and tools presented in hair farm work better for what you are trying to achieve?

Also don't forget to checkout the gallery, they have some amazing renders on display. Most notable is the stunning work of Luc Bégin (http://www.hair-farm.com/gallery.php?f=Luc_Begin-Beautiful_Woman04). The full project overview of this master can be viewed here (http://cgvilla.com/2015/04/07/realistic-wonder-women-work-process-by-luc-begin/). Also nice forum post here (http://www.hair-farm.com/forum/viewtopic.php?f=13&t=1439.

He claims Hair Farms is the best 3ds max plug-in for hair.

It is relatively expensive hair simulation software but maybe for good reason?
In this comparison chart you get a good overview of the available features (http://www.hair-farm.com/hair_and_fur.php).

Posted over 8 years ago
0

I agree with hair farm. but due to lack of resources i just want to finish this task with ornatrix only, i have seen a few videos regarding hair simulation, and what i got till now is ox Dynamics method , even i could simulate it with cloth modifier. and i tried it with a sphere and worked but once i try it with my model got nothing. But I have asked if you could upload examples for ornatrix. so i could backward engineer how it could be done.

Posted over 8 years ago
0

I have some past experience in dynamics simulations and hair rendering/animation (in joe alter shave and a haircut). It's not expert experience but enough to know it's a very difficult thing to do.

I'm not directly familiar with Ornatrix toolset and its proprietary ox modifier and connection to massFX engine, but as far as I can tell, this combination is not going to deliver you a toolset for retaining complex hair models and performing the complex dynamics you are trying to get.

The developers of the toolset are warning to not set to high expectations on the connection with the massFX engine, which is originally a game oriented real-time dynamics simulator with some serious limitations when it comes to producing film like effects or delivering production quality results (more focused on speed not so on precision). The simulations performed by the engine are more granular.
You could compare it with trying to perform complex lighting with 1 GI bounce and few samples.
Maybe on a very simple object you could get some nice results but in a complex scene the lighting you get would be very low resolution. The same is true for dynamics simulations, you have different sampling factors involved for getting a good final result, things like time scale (amount of samples taken in-between frames) and collision radius and so on.

The more complex/dens and closer points are to each other the more precision you need.
The craziness of the hairs you are talking about is a clear indication of limitations (to low granularity of the calculations) of mass FX collision detections and precision on vast amounts of overlapping and close to each other laying points and lines.

Also as far as I can tell, the massFX engine will affect things globally and so in this case introduce problems for retaining a hair model, the only way to prevent losing the shape (in how you currently do things) is to stiffen things up globally in the dynamics parameters, but that will not deliver realistic results in general.

You need to be able to selectively control the effect (assign regions of effect) and get groups/clumps of hairs to interact, slide and have various amounts of friction and spring values and so on. The hair farm toolset provides isolated volumes via the hair meshes and thus delivers a different approach to simulating complex hair models, you need to get to a similar kind of approach in order to retain that hair model.

Maybe try it with bones/IK chains and perform the dynamics simulation on them (simplify the dynamic system to be salved, reduce it to maybe 20 IK chains)and then try to get the hairs attached to the bones via skin modifier or something like that (that is when you stick to the toolset you are using now). In this way you would at least be able to control/retain the base shapes of the hairs.

Also I think You need to use a dynamics simulation system that can deliver more precision and provide more flexibility then current mass FX engine can deliver.

In general I'm a bit afraid maybe you set to high expectations on a toolset not specifically designed to perform the tasks or deliver the results you are trying to get from it, and also maybe underestimate the complexity involved for getting the result you are trying to get.

You will definitely not solve it in 2 days.

Your answer

In order to post an answer, you need to sign in.

Help
Chat