Cloth - Preserve Shape for Ornatrix

 
 
 
Posted by:TAT Studio
Data created:28 April 2016

Hello !

 

I'm on R&D for doing hair dynamics with Lucid & Ornatrix for a feature film.

 

First simulation is pretty fast & easy to get, but tweaking simulation for final resul is pretty hard.

 

For exemple, simulation stretch a lot my based cloth, even with a calm animation.

It has for effect a motion blur style but not exepected :)  It also kill my clusters when re-importing in Ox.

 

First real test : no collisions

preview : http://pingus.tatprod.com/BUG/Lucid/LucidOx_Dyna/P_SIM_branche_preview.avi

render : http://pingus.tatprod.com/BUG/Lucid/LucidOx_Dyna/P_SIM_branche.avi

 

First real pain : slow anim, ground collision, arm/leg collisions

Here, i tryed to convert the guides as pipes (instead of flat bilboard) and get some 'betters' results but is hugely heavier (simulation time, space)

preview : http://pingus.tatprod.com/BUG/Lucid/LucidOx_Dyna/P_SIM_marcheLente2_tubes.avi

render : http://pingus.tatprod.com/BUG/Lucid/LucidOx_Dyna/P_SIM_marcheLente2.avi

 

I found Lucid good for a fast try, but there's a real lack of documentations, impact of each parameter.

 

How the hell dynamics are manage in production (Blur-like) ?

Do you consider Lucid valid for end-quality production ?

 

Keep Going !

Thanks !

-p-

 

 

 

 

 

http://tatprod.com/

Hello,

The tests so far are looking great! We are looking into this. Pavel should get back to you about your questions tomorrow.

Marsel Khadiyev (Software Developer, EPHERE Inc.)

I think cloth stretching could be fixed increasing simulation substep/iterations counts (Lucid/Show Flex Settings/Simulation). I think 10/10 should be fine. But cloth can't enforce original hair shape.
For arm/leg collision dynamic SDF can be used. For ground static or convex collision object.
We can look at the scene file if there are other issues (private FTP - ftp://dropoff:G00dStuff0nly@ephere.com/)
We are working on direct hair simulation (hair as polyline). This method should be more robust.

Hey Guyz ! 

 

Just saw Lucid Update Orientated Hair dynamic.

Trust me i'm so happy about it :D

 

I tested on teapot / sphere workflow, and indeed the simulation is munch better to handle.

But i wanted to try it on my character and it's looks like the Lucid Modifier doesn't work on Ox_Baked_Guides yet. 

I've try EditGuide or even GuideFromSurface, but same result.

 

Lucid_Tony.JPG

 

Do you reproduce it ?

 

Thanks !

-p-


Attached Files:

>Lucid_Tony.JPG (356203 bytes),

http://tatprod.com/

We're looking into this, thank you.

Marsel Khadiyev (Software Developer, EPHERE Inc.)

I can't reproduce the issue. Lucid handles baked guides fine for me. Please look at the attached scene and the screenshot. If this is still the issue please upload sample scene.


Attached Files:
Image 1
>BakedGuidesTest.max (319488 bytes),

Hi back !

Ok, indeed it's working on baked guide. I've lately realized that i've download & installed demo version (still available above licence).  My bad !

 

So, here's my 2 mains problem with the new dynamics for fur (haw can we call it ? :) )

- Dynamics is globally too springy. Even with no animation, the hair never stop boucing & moving. It's gives unrealistics result for me. I would love a parameter to fix and attenuate the movements & dynamics.

 

- I've got some (random ?) guides that goes bug (huge length) when turning simulation on. Sometimes they can stick into air too. 

 

Links :

Tony Without Animation

Tony 3dsmax file

Miguel animation

 

Thanks !

http://tatprod.com/

"I found Lucid good for a fast try, but there's a real lack of documentations, impact of each parameter."

We would love to have each global lucid setting have a  very detailed description and have pictures and videos for each setting:

For example Drag 0.0  this happens,  drag 0.5 this happens,   drag 1.0  etc etc

Even after reading manual it is still not clear or we do not see visiable difference in simulation  what is  drag, lift, dissipation, damping, surface, restitution, inertia bias etc etc  when those must be use, why.  are they used for cloths or liquids,  should values be between  0 -1   or they can go upto 100 etc etc

Thank you

 

 

Did you guys got time to watch my files ?

I'll need a pipe for hair dynamics soon.. what can i expect from this method ?

 

Thanks

-p-

http://tatprod.com/

Yes, we're investing the scene, stay tuned.

Marsel Khadiyev (Software Developer, EPHERE Inc.)

Thank you for the report. We have found and fixed the issue with very short hair strands. The fix will be available in the next Lucid release.

I also tuned a little the attached scene for better hair simulation:

Important changes:
Damping: 1.0 (for moving hair attenuation)
Inerta Bias: 0.001 (any small non-zero value should be fine, in new Lucid versions default value for new scenes will be 0.001)

Minor changes:
Sub-steps: 10
Iterations: 10
Gravity: -100 (no need for strong gravity)
Particle Radius: 0.5 (set particle size manually and do not use resolution parameter)

 


Attached Files:

>P_Tony_SIMU_02_fixed.mp4 (1231948 bytes), >P_Tony_SIMU_02_fixed.zip (25717566 bytes),

Looks a bit low gravity but it's far better than what i had.

Waiting for next release then !

 

thanks

http://tatprod.com/

I think the gravity can be increased, or you can increase the density. The preview also looks to be real-time capture, not pre-recorded.

Marsel Khadiyev (Software Developer, EPHERE Inc.)

Any news of next release ?

http://tatprod.com/

Should be soon, but meanwhile you can always download the latest nightly build which should contain this fix.

Marsel Khadiyev (Software Developer, EPHERE Inc.)