FumeFX

#5001

Thanks Allan for stopping in and helping out, it’s amazing what you can do!

Rich

#5002

Yes that is something I already do. But thank you for reminding me that this also can help a bit to resolve some issues.

#5003

Allan, thanks for stopping by! I bought your FumeFX-training right when I ordered FumeFX. Great tutorials, man!
Maybe it’s time to do some lessons on real high resolution effects :wink:

PS I’ve seen your new reel, looks awesome, the only glitch was the horizontally squeezed footage of Blade. Maybe if you could set that to the original aspect ratio, that would make the reel even better. Especially because the reel opens with this footage.

#5004

So lately I’ve been thinking about making some Fume tutorials. Here is a test render of what I think will be my first tutorial…

However, I can’t take all the credit, as most of it is inspired by Dan Chamberlin’s post here…

http://forums.cgsociety.org/showpost.php?p=7238518&postcount=4736

Anyway, I will add a few techniques of my own in there too, and plan to do it very in depth.

#5005

that looks awesome!
I never thought of making that type of clouds with Fume.

Reminds me of the days when I was a betatester for Ultrashock (also for 3ds max), a volumetric plugin distributed by Digimation. Great for explosions and clouds.

Fume simply takes it to the next level. And judging by that image, it looks very promising…

#5006

That would be a great tutorial. One which I’d gladly pay for. When you hoping to have it made.

#5007

Nice job on the shader :wink:

#5008

Explosion R&D V3. Still not perfect, but I’m working on getting there…

https://vimeo.com/44150355

Still Preview…

#5009

Guys, a quick update regarding solving voxel artifacting in high res sims:
It seems I may found a solution. After some serious testing I noticed that a lower step size% doesn’t result in a better quality render (which I assumed was the case). Slightly higher step sizes result in way better quality.
The current workflow I’m testing is setting the Jittering % to 0 and finding a Step size% (currently 10-30-range) at which the voxel artifacts at the silhouette are gone. Then upping the Jittering% to “clean up” some banding inside the flames.
Basically you have to find the step size sweet spot instead of choosing the lowest. Maybe you guys already knew this, but for me it was a revelation :wink:

To be honest I’m very excited about this solution. And I wasn’t expecting this. Does somebody know the reason why the lowest step size doesn result in the best quality?
fyi: I observed that when going to low (1-10 range) (step size%) you get a very pixelated outcome. Except when adding more jittering, which seems to benefit from a lower step size%.

#5010

is there any one who got fumefx3 window crashed (low ram) on low spacing even when u have a large capacity of ram available … it usually comes after I/O done according to Log file ?

Frame: 39
Vmax=9.457, dtReq=0.23, steps=1, CFL’=21.99
[larg fire area -src] Building voxels for front seat rock
[front seat rock] 4048 voxels, 0.1 Mb, 0.0s
[larg fire area -src] Building voxels for Sphere002
[Sphere002] 3762 voxels, 0.1 Mb, 0.0s
[larg fire area -src] Building voxels for Sphere006
[Sphere006] 4284 voxels, 0.1 Mb, 0.0s
[FFX Object Src001] Building voxels for Object003
[Object003] 36000 voxels, 0.2 Mb, 0.0s
[FFX Object Src001] Building voxels for Object
[Object] 11928 voxels, 0.1 Mb, 0.0s
[larg fire_back seat] Building voxels for Sphere003
[Sphere003] 1664 voxels, 0.0 Mb, 0.0s
[larg fire_back seat] Building voxels for Sphere004
[Sphere004] 1560 voxels, 0.0 Mb, 0.0s
[larg fire_back seat] Building voxels for Object003
[Object003] 36000 voxels, 0.2 Mb, 0.0s
Resizing: from (346,198,244) to (353,198,244)
Memory used: 748Mb
Resizing: from (353,198,244) to (348,198,244)
Memory used: 737Mb
ApplyForces done in 5.83 sec
Advect velocities done.
CG done in 31.35 sec with 200 iterations (104.698 MVOPS).
CG res e:-4.344 m:-4.344
Advect fields done.
I/O done in 1.67 sec
Saving fxd data FAILED (low RAM).
Frame time: 00:48.60
Simulation terminated.
Flushing larg fire area -src
Flushing FFX Object Src001
Flushing larg fire_back seat
Flushing bmw_main
Flushing Box002
Flushing Box001
Flushing RAM

#5011

No I haven’t experienced that. What I did notice is that especially 3ds Max 2013 takes a long time after a render to give the control back to the UI.
And also if you have in FFX Preferences set that you want to be able to resume after nth frame, you can get very big .fxc-files (depends on your spacing and other settings of course) and these can take a long time to save to disk after the sim is done (a few minutes).
It’s also strange you get this error, since when you have not enough RAM the system starts using the windows scratch disk.
Did you check that you have enough hard disk free to save the .fxd-file?
But the exact error you give here I have never received with FFX, sorry.

#5012

yeah actually its strange for me too, i have around 8gb of ram from which +4 gb was unused during sim… & as far as disk space is concern i have +50 gb of free space during sim… only one solution seems working & that is by increasing spacing value i can sim few more frames before crash… but obviously cant maintain quality… is it 2013 issue? do i need to try the same file on 2012 or 2011? also i m thinking to go back in fumefx 2.1c with no effector :frowning:

#5013

maybe a long shot, but are you running a 64-bit version of your OS (like WIndows 7 64-bit)?
If not (you’re running a 32-bit OS), not all 8GB can be used, so it would seem to FFX you only have 4GB in your system. The reason it wouldn’t occur in FFX2 may be that FFX 3 requires more memory due to more features or more channels you’ve selected.

But as I said this is a long shot since you really should have a 64-bit OS (and you probably have) to take full advantage of your RAM.

#5014

i got win7 64 bit 8 gb ram 500 gb hdd xeon quadcore processor with 440gt nvidia gcard… i dont knw if my pc specs doesnt support ffx3… however all of the other software are going fine even houdini & naiad which i am practicing these days…

#5015

A small private project,
inspired by this picture: thisiscolossal.com/2012/06/up-in-smoke/

   [https://vimeo.com/44549271](https://vimeo.com/44549271)

Uli

#5016

I like it very good effect.

Rich

#5017

Nice sim/render Uli !

I’m working too on a personal project and I wanted to know if is it possible with FFX to create a candle smoke simulation like this :

I’m trying since few days, with lower advection, Advection Fields, Turbulences, Buoyancy, systemscale 2, etc but I don’t find the best way, do you have an idea ?

Thank you in advance for your help

#5018

I would start by lowering vorticity first, and then tweaking from there. Say around 0.1 - 0.2.

#5019

My vorticity is set to : 0.
I think it’s hard with ffx to achieve this type of result, I’m lost :bowdown:

#5020

first create the effect without turbulence. So you’ll probably get a “straight plume of smoke”. Only add the turbulence as the last step. So you can concentrate on the turbulence setting to achieve the smoke movement.

Maybe you could post a screen of how your effect looks at the moment? So we can compare it to the screens you posted already.