Hi everyone,
here is my first test of FumeFx & RayFire
http://youtu.be/6DCP4rYpAbY
In this test,
I pack all debris object ONE BY ONE in Obj/Src tab, this process is boring and time consuming.
Is there any method to add all object which have to influence smoke as once in FumeFx?
I use 2.1a
Sorry for this dumb question:)
When you save with the .jpg, the 2.2 gamma is probably being baked in, however if you’re opening the png/exr in a compositing package, are you viewing the image in the proper color space or applying an sRGB gamut node to the file? If not, you’re essentially just looking at the image in 1.0 color space. More than likely, your data is there, you’re just not loading/viewing it properly.
What are you using to composit with?
Hi Cryptite, thank you for reply, I use After Effects and I have tried to change color space to sRGB but anything was changing. In the PDplayer when I use colorspace : sRGB and I increase gamma to 2.2 my smoke is too bright not like the .jpg. So i need to use an gamma node ? (I don’t know which). Did you have problem with fumefx and vray ?
With .jpg :
With .exr :
When I import my .exr/.png in After effects (I lost a lot os smoke details)
I opened both a jpg and exr both in Fusion and AE and they looked the same. How are you saving the jpg’s and the EXR’s exactly? Can you show the settings you’re using?
Seeing as your gamut settings appear to have the output gamma set as 2.2, you probably don’t need to adjust your color space, in hindsight, so I don’t think viewing color space is the issue anymore.
I appreciate a lot your help, I hope my pictures can help you to see my problem.
My .exr settings
My .png settings
My .jpg settings
Today I have change my smoke shader and for the same render with vray or scanline with a gamma of 1, I have this :
In my 3DSmax buffer :
This is my import .png in after effects :
And this is my import .exr in after effects :
I can use for compositing my .png over my .exr to have approximately the same result as 3DSmax buffer, but I think it isn’t the good method.
I believe the fault is with After Effects trying to handle 32 bit full float EXR images.
Try this, Save your EXR as Half Float (16 bits/Channel), then in After Effects (with the regular color space, not sRGB), interpret the EXR footage as usual with the Straight - Unmatted alpha, but the key is tick on Preserve RGB in the Color Management tab. It’s trying to apply linear light to the EXR sequence which appears to be muddying up the color for no apparent reason.
Hope that works.
Edit: I thought that worked, something went wrong though, still looking into it!
Edit2: I’ve tried almost all possible combinations, and unfortunately all I can say is After Effects just won’t reproduce the same picture you’ll see in your max render buffer. You can get close, but you’ll probably have to adjust curves/levels to get the same-ish image. Perhaps now’s a good time to learn Nuke or Fusion :surprised
Edit3: If you ignore the alpha, however, you get the proper colors. Your final alternative could be rendering the alpha out separate and importing it; could work.
You’re right Cryptite ! After Effects don’t display the same as 3DS max buffer, and in Nuke my .exr is really the same 
I’m learning Nuke since 1 week so I would like to composite my project with AE that I use since 2 years. But I think I will work for this project on Nuke. 
About .exr you prefer : full float, half or integer ?
Cheers
Hi Kieran,
I experienced today what you mentioned above. I was able to squeeze a very detailed grid from my 16Gb-machine and the fire behaved completely different from the slightly coarser draft-grid. It was if all the fuel was completely gone(only very small flames remained and very large bubbles of unburnt fuel). And the scale seemed larger (very large voxels). it seemed that the sim freaked out and resorted to large voxels (very coarse grid)… very strange…
I hope I can rectify and get decent high resolution results, otherwise a new machine won’t do any good if I can’t solve it…
Does anybody has any experience with FumeFX changing behavior drastically when simming very detailed grids, and what do you do to solve it?
@AdrienSliver: I usually always work in 32 bit, since I know I’m losing the least amount of data. When it comes to packing in render elements through vray, though, you only really need full float data for passes like ZDepth, SamplerInfo, Velocity and things where you will have values outside the standard range of color (0-255). The other things are usually fine to leave at half float and you’ll save yourself a ton of space by doing so.
Everyone whom uses fume does. 
There are two general methods (well there are more really but in this case)
Test sim at or very close to your final spacing. These results are predictable and well… you get what you pay for.
Sim a low res grid and wavelet. In the long run you can iterate faster but your final may take just as long since, in the end, the process takes two simulation steps to complete, your default pass and your wavelet pass. Also Wavelet has its own “look” that a standard sim does not have, which may or may not be desirable.
John,
I prefer the first method, because the wavelet has (as you mentioned) a certain look that I want to avoid (in this project). The only drawback is the long sim time 
(60 hours for 30 frames).
It’s a pity really because the look I had was spot on, but a few steps of extra detail completely changed it and made me have to start over again, spending long sim hours…a bit frustrating… but rewarding in the end (I hope).
So no matter what, I’m ordering a new machine to speed things up 
Added: Is it a correct assumption that you can use a network-slave to sim but the sim can not be divided between several network slaves? So running a sim is always restricted to 1 machine. Just wanted to be sure that 3.0 hasn’t changed anything.
Hi everyone,
here is my first test of FumeFx & RayFire
http://youtu.be/6DCP4rYpAbY
As I want debris influence the movement of smoke,
I pick all debris one by one, it is really time consuming and painful
Is there any other method to add all objects as once ?
I use 2.1a
Sorry for this dumb question:)
@Cryptite , I’m going to save my pass like you in 32 bits, in half float.
Thank you again for your help
At last, the new custom build “Fume” workstation is ordered. Now I have to wait 2 weeks and than FumeFX scaling will go “very” low
Select all of the objects you want to add to Fume, then paste and run this snippet into the maxscript listener:
for i in selection do <FumeFX>.addSource i
Where <FumeFX> is exchange, with the name of your FumeFX container.
I ran across this last night which I thought was pretty cool. One of the best fume explosions I’ve ever seen.
Hi
can i somehow render the smoke with repetitive texture pattern?
So that when i would apply a evenly tiled grid (tiles) map, the smoke would look like made of a grid (which would be deformed with smoke motion)
When i would simulate smoke from wide rectangle with gravity vector in for example Y axis, without any turbulence, after some time i would see a clear grid pattern in the width of my rectangle emitter.
Is it possible with fume?
You can apply a pattern/map in the Rendering-tab.
For each effect (Smoke and Fire) you can assign a map and how you want to map it (“Map source”), this can be set to Fluid mapping (texture moves along with you smoke/fire) or World/Object coordinates which keeps it static.
I don’t know if the above helps you, but if so please read up on it in the Help of FumeFX it’s all described in there.