FumeFX

[VladimirStefanovic]

I have seen this with xmesh. Do you have any modifiers on top of the xmesh loader?

None, seems like Xmesh has some weird issue with fumefx compability, I searched a lot and never found an explanation or something

[JohnnyRandom]

Hmm, it is possible there are some strange vector lengths coming from somewhere, that seemed to be what happened to me, I had added a relax and I had approx 5 faces blow off some extreme velocities. Can you recache the mesh to see if it was just a bad frame?

You may be able to effectively control this with an effector if you are using Fume 3.+ by testing the per voxel velocity and replacing it with a more normalized value. So test velocity from min-max and replace or multiply with a less value. Not perfect solution but it may get you past that bunk frame.

[Ash007]

Hi, at the moment I do a little RnD for a dragon flame. Everything works fine so far, except the point when the RayFire simulation starts to break apart the handcart.

Here is a little test video (~1 MB): VIDEO (http://dl.dropbox.com/u/20751640/test03.mp4) The last couple of frames are the important ones.
Here is a viewport preview of the whole scene (~2MB): VIDEO (http://dl.dropbox.com/u/20751640/_scene.mp4)

When all the little bits and pieces start flying around, the FumeFX simulation starts to fill out the whole simulation container and that's look weird.
Is there any chance to simulate the fire without spreading so much?

thx

[VladimirStefanovic]

Hmm, it is possible there are some strange vector lengths coming from somewhere, that seemed to be what happened to me, I had added a relax and I had approx 5 faces blow off some extreme velocities. Can you recache the mesh to see if it was just a bad frame?

You may be able to effectively control this with an effector if you are using Fume 3.+ by testing the per voxel velocity and replacing it with a more normalized value. So test velocity from min-max and replace or multiply with a less value. Not perfect solution but it may get you past that bunk frame.

I found the solution to my problem: I was Xmeshing the whole thinking particles instead of just the groups, now that i saved just the two groups of TP it works fine with no artifacts :)
Thanks for helping :thumbsup:

[turas29]

Hi, I was trying to do some clouds, what do you think?
http://imageshack.us/photo/my-images/713/rendclouds.jpg/
https://vimeo.com/51777803

[AMParticles]

Hi guys,


I'm wondering if there is a way to render out cross-sections of a Fume sim. I would like to render the internal volume, not just a outline of the cross-section. The goal is to do very large renders for print of the various cross-sections.

I have tried clipping planes on a camera setting the near and far this does not work for fume. Also I know I can do it with a magma flow delete node and geometry to essentially slice the resulting particle clouds. But I was wondering if there are any other ways you guys might now or can think of...


Thanks !

[benio33]

@AMParticles: I think you could just place a box over unwanted-to-render section of fume grid and put a matte/shadow shader on it.

edit: Supid idea - doesn't work :( - checked ... sorry.

[3ak]

Hi guys,


I'm wondering if there is a way to render out cross-sections of a Fume sim. I would like to render the internal volume, not just a outline of the cross-section. The goal is to do very large renders for print of the various cross-sections.

I have tried clipping planes on a camera setting the near and far this does not work for fume. Also I know I can do it with a magma flow delete node and geometry to essentially slice the resulting particle clouds. But I was wondering if there are any other ways you guys might now or can think of...


Thanks !

Fumefx is not flexible enough so you have to do some extra work:
1) Sim your main grid.
2) Create it's clone (don't forget to change cache files!)
3) Add fumfx source and 2 void sources to the clone (change their type to box and place them so you leave only slice of your 2nd grid).
4) You can use effectors instead of void sources if you need something more specific then just delete voxels.

[JohnnyRandom]

Fumefx is not flexible enough so you have to do some extra work:
1) Sim your main grid.
2) Create it's clone (don't forget to change cache files!)
3) Add fumfx source and 2 void sources to the clone (change their type to box and place them so you leave only slice of your 2nd grid).
4) You can use effectors instead of void sources if you need something more specific then just delete voxels.

Nice someone actually using an FFX source ;)


You can also use Krakatoa you won't have to run another sim. Pretty easiy setup, same basic premise as 3ak described.

[max3d2008]

Hello guys. Does Anyone know what is the temperature value set by default in the fumefx grid without emitters?

[JohnnyRandom]

It is 0.0, there is no ambient temperature (yet)

[3ak]

Nice someone actually using an FFX source ;)


You can also use Krakatoa you won't have to run another sim. Pretty easiy setup, same basic premise as 3ak described.

Btw, didn't you try some approximation for multiple scattering in krakatoa? I didn't mention (thanks) cause 90% uses MS and i just don't know (yet)) how to approach it in magma mod (or at least something plausible)

[JohnnyRandom]

Btw, didn't you try some approximation for multiple scattering in krakatoa? I didn't mention (thanks) cause 90% uses MS and i just don't know (yet)) how to approach it in magma mod (or at least something plausible)

You have all the necessary channel data, it is just a matter of applying it the same way the Multiple Scatter function does. I have tried with limited success, someone else I am sure could hammer right through it, me takes me a while to wrap my head around it. Man that was a long time ago.

[max3d2008]

It is 0.0, there is no ambient temperature (yet)
Thank you.

[Daniel-B]

Is it just me or has it been anyone else's experience that the Temp Buoyancy default in FumeFx is way too high? When I make a flame with default settings it's way too stringy vertically. If I put the Temp Buoy around 0.2-0.25 and the vorticity between 0.6 and 0.75, my flames look instantly better.

[JohnnyRandom]

Yeah you are probably right, I know there are a few settings that are similar. I personally think that the timescale default of 1.0 is too slow, or the ambient smoke color being dark grey instead of black, or any one of a dozen others :)

In fact I always create "my" default test grid with a script that sets my general params. Which is something that I think is another of Fumes greatest strengths, the depth of its MXS is pretty deep. There are but a few functions/params that aren't accessible.

[3ak]

Hey guys, how would you approach a subject of emitting smoke from a sphere like simple source one, but with falloff near the edge of it. What I mean is it emits much more inside the sphere and at the egde almost none. What I already tried:

1. Emitting with particles - gives great control, but gives to "dotty" look - very hard to smoothen out.
2. Emitting from simple source with noise map - too hard to get "sphere" inside using noise parameters, and the shape is not sphere enough.
3. Emitting inside a fume container from a simple source with high diffusion then use it as a fume source in another container - works best but I can only "read" values from fume source, and no "add" possibility. That way when I use few containers they "eat-out" their results :((

Thanks in advance.

I'd use effector - create spherical one with the same size as source, align it, and set it to multiply channels with constant (==1) and enable radial falloff.
Example:
|https://dl.dropbox.com/u/46208032/3ak_source_with_radial_falloff.max

[benio33]

I'd use effector - create spherical one with the same size as source, align it, and set it to multiply channels with constant (==1) and enable radial falloff.
Example:
|https://dl.dropbox.com/u/46208032/3ak_source_with_radial_falloff.max

Thank you 3ak I really appreciate that! It is what I'm looking for, but the thing is, it works only for one emitter. When 2 are mixed they eat-out each other, and I need it to work when 2 emitters intersect. But the idea is really good, and maybe I can figure something out using effectors.

[3ak]

Thank you 3ak I really appreciate that! It is what I'm looking for, but the thing is, it works only for one emitter. When 2 are mixed they eat-out each other, and I need it to work when 2 emitters intersect. But the idea is really good, and maybe I can figure something out using effectors.

I don't think you can use effectors for more than 1 source in this example. they are too restrictive.
Use particles - create object with shape you need for emitter, emit particles inside, calculate distance to nearest point on surface of the object (fake SDF), normalize it and use as multiplier in FumeFX. I use tp so i can use custom particles channel.
You need quite big amount of particles, decent radius and low smoke value - this way you can "blur" channel and avoid popping while your multiplier gives you desired density from inside to outside of the shape.

[benio33]

I don't think you can use effectors for more than 1 source in this example. they are too restrictive.
Use particles - create object with shape you need for emitter, emit particles inside, calculate distance to nearest point on surface of the object (fake SDF), normalize it and use as multiplier in FumeFX. I use tp so i can use custom particles channel.
You need quite big amount of particles, decent radius and low smoke value - this way you can "blur" channel and avoid popping while your multiplier gives you desired density from inside to outside of the shape.

Yeah, you're right in both cases - using effectiors didnt work for me, and using particles did. What I did is have a spheres filled with particles, calculate the distance from the middle, the largest are on the edges for most blur and smaller are inside, the interior ones emit the most based on a track curve. Funny is I started with particles, but didn't think of fading their emission near the edge - works like charm. Thank you 3ak! You saved my day.

[Debneyink]

So fumefx is great obviously but I'm looking at Phoenix Fd also as an option...

I'm working on a nuke and this is what I've got so far without too much time, there's frames missing in the sim which I'm working on but you get the idea.

https://www.youtube.com/watch?v=4D_uEVI11SA

https://www.youtube.com/watch?v=swHP8WCJj8E

http://www.youtube.com/watch?v=bz5Th1fy8qg&feature=youtu.be

what i like most about these and Phoenix is its ability to work with Vray, both these shots are using one light with a hdri map!

Does anyone know if you can use hdri with fumfx?

oh and I'm going to continue my phoenix tests on the Aura sticky thread...

cheers

[finalight]

Hi everyone,

Wondering if anyone knows if this has been resolved or possibly a workaround. I'm getting the same error as below using fumefx3.x. What I've found is that if writing to a local drive, it doesn't give the error. Have tested on multiple 64bit machines with 16GB ram each. As soon as the sim pushes to 4GB ram usage (total usage - max and win), it throws the error. Haven't seen this prior to v3.x

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 ?

.......Saving fxd data FAILED (low RAM).
Frame time: 00:48.60
Simulation terminated. ....

[Daniel-B]

Brian, I don't have a solution to your problem, but I just wanted to say that the Fume explosions on your site are the best I've ever seen. Really excellent. Did you use particle or object emission?

Any info on how you made them would be appreciated.

[finalight]

Heya Daniel,

Thanks very much - your RnD on Vimeo is really cool - nice work, good progression - eh, all those blasts look cool actually. It's obvious you have patience to get some good ffx movement/behavior, mess with shader tweaks, and pull away from default settings.

For the stock blasts, some use object sources, and others, particle. Some tips would be the following (though a lot of these are probably commonly known):

--I rarely use TimeScale 1 nor real world scale to drive parameters chosen.
--I usually have to dial explosions at full res since they come out very different from higher spacing (low fi) tests. It's tedious, but pays off. To save time, I use the 'slice' method to test the cross section of movement first, then expand out to the box depth I need.
--I don't use a lot of noise - just enough to similate ambient air currents that help break stuff up as in real life. I try to use either heat or velocity to get curling motion.
--Low burn rates of course
--Low vorticity
--Contrary to suggestions, sometimes raising the Quality to 6 or 7 has yielded mushrooming I was after, whereas 5 would not. Just depends.
--Note taking regarding RnD (i.e. v030 - changed vorticity from AA to BB; changed burn rate from CC to DD). Sim, review, fall back to previous version? push on?
--Patience. I ended up RnD'ing fire/smoke/explosions for 2 years before getting what I was after. Most of that was sim time. Morning routine - check sim, tweak for 3 min, jet to studio. It helped enable dialing of what works.
--For Multi-scattering within super large scale blasts, MS settings should be very low - falloff versus fire strength takes some dialing. Best method is to pick a frame, save shadow to disk, then render/tweak/repeat until the right amount of falloff versus fire strength is obtained. Fire shader also affects it, so it's a parametric tweak session.
--Sleep sometimes.

Cool looking fire shader notes on your Vimeo. Fire is definately a strange beast, kind of occludes - kind of doesn't. After messing with a bunch of fire stock all year, shader tweaks definately can go a long way.

Cheers,

[garryclarke]

Hi,

Is it possible to suck smoke out of an enclosed space?

I have a room where a noxious liquid on the floor is giving off a nasty vapour - this bit shouldn't be too hard.
Then a hatch opens in the ceiling and the vapour starts to head toward the hatch (acting a bit like a chimney.

How would I go about setting this up?

[3ak]

Hi,

Is it possible to suck smoke out of an enclosed space?

I have a room where a noxious liquid on the floor is giving off a nasty vapour - this bit shouldn't be too hard.
Then a hatch opens in the ceiling and the vapour starts to head toward the hatch (acting a bit like a chimney.

How would I go about setting this up?
Try to use spherical wind with negative strength to suck in smoke through hatch and void source to kill smoke behind it. It will work but don't forget that wind knows nothing about walls so it will create velocity field that points right through wall . so you may add another wind poining to hole in wall or even more complex - stream of particles, used as velocity field (with big radius and small vel multiplier to make it subtle force).

[JohnnyRandom]

Or use a Gravity Vector and set it to point, since 3.x gravity param now accepts negative values.

You could also play with an effector too, test smoke density values to give greater effect for more dense smoke. Or half a dozen other things to add some uniqueness to it :)

[garryclarke]

Thanks for the tips, I'll give those a go.

- Garry

[Daniel-B]

Is it possible to suck smoke out of an enclosed space?


Here is what you do. Put a simple source over your hatch that emits temperature/heat only (upward), and nothing else. The cold air in the Fume container will rush in to fill the void left by the low pressure created by the hot air, and pull the smoke along with it.

I had to do this to suck smoke through letters, and it worked beautifully.

Heya Daniel,

Thanks very much - your RnD on Vimeo is really cool - nice work, good progression - eh, all those blasts look cool actually. It's obvious you have patience to get some good ffx movement/behavior, mess with shader tweaks, and pull away from default settings.

Cool looking fire shader notes on your Vimeo.

Interesting. Your notes are similar to my methods. One of the differences that stood out to me is that I can never get an explosion to look good with low vorticity. How low are you talking? My explosions are usually between 0.6 and 0.75 vorticity. Anything lower just ends up looking like a "bubbly" smoke. I also usually have my Advection Stride between 0.125-0.3. This helps keep some of the curling motion.

Do you emit fuel at all before you ignite it, or do you simply emit fuel and temperature at the same time? I'd love to compare notes via email sometime if you are up for it.

Also, I always work at real world scale.

[turingm1]

Hi Everyone,

I just finished my 2012 website and special effects demo reel. It's located at http://www.chaddelliott.com

All feedback is welcome

Thanks

[Daniel-B]

Hey guys. I was wondering if those of you who make Fume explosions could give me some detailed pointers. I am really liking the way my explosion sims are looking, except for emission. All of my explosions are a blobby fire when they first ignite. This is my latest attempt...

https://vimeo.com/47798297

However, I would love to get the first few frames to look something more like this...

https://vimeo.com/47717040

Any thoughts?

[benio33]

Guys, how do you change for example "Smoke Opacity" in rendering tab with maxscript? I can change most of the other params listed in "show $fumeGrid" e.g. gridSpacing, gravity etc., but there is none changing Fire, Smoke and Fuel params in the render tab :( ... or is there?

[3ak]

Guys, how do you change for example "Smoke Opacity" in rendering tab with maxscript? I can change most of the other params listed in "show $fumeGrid" e.g. gridSpacing, gravity etc., but there is none changing Fire, Smoke and Fuel params in the render tab :( ... or is there?

s = $.getshader()
show s

[benio33]

s = $.getshader()
show s

Of course, these are shading parameters! Damn, you're good!! Thanks 3ak!!

[finalight]

Do you emit fuel at all before you ignite it, or do you simply emit fuel and temperature at the same time? I'd love to compare notes via email sometime if you are up for it.



Heya,

Sometimes I do. It depends on what the client is after. Sure - I'm up for emailing. Free free to shoot me a msg thru briandemetz.com

Cheers,

[Solbester]

I created a tornado using pflow particles, and then tried to use them to create a fume fx tornado. Unfortunately once it hits 200 frames the tornado explodes outward even though the pflow particles don’t. How can i fix that? I want the fume particles to follow the pflow particles perfectly. thank you in advance.

[Menesis]

Is FumeFX something that is really used in blockbusters? I can't find any references really (except a little list on their own website).

[JonathanFreisler]

What's the point of a question like that? Yes? It completely depends on the studio.

[bhavya2]

What's the point of a question like that? Yes? It completely depends on the studio.
Very true sir...

[adom86]

currently the best question of 2013.. wonder if it will hold on :D

[JohnnyRandom]

Is FumeFX something that is really used in blockbusters? I can't find any references really (except a little list on their own website).

It is used in more places than you might think ;)

[Glacierise]

It is used in more places than you might think ;)


Ooh yes indeed :)

[fireknght2]

Most of today's movies use FumeFX and other Fire/Explosion/Smoke CG tools. The Avengers Movie used a very large amount of FumeFX shots. Its a safer method and doesn't require all the permits that old school pyrotechnics require.

Rich

[Daniel-B]

Most of today's movies use FumeFX and other Fire/Explosion/Smoke CG tools. The Avengers Movie used a very large amount of FumeFX shots. Its a safer method and doesn't require all the permits that old school pyrotechnics require.

Rich

You sure about Fume? I know ILM has their own inhouse tool called Plume for such things. They could have used Fume, I just would think ILM would use their own tool. :shrug:

[SoLiTuDe]

You sure about Fume? I know ILM has their own inhouse tool called Plume for such things. They could have used Fume, I just would think ILM would use their own tool. :shrug:

ILM isn't the only company that worked on Avengers. Hydraulx also did a few shots, and they have been using TP and Fume for a while now.

[fireknght2]

ILM does not do all the FX for Marvel/Disney. There are many companies even small time single employee companies that work on the projects as well. Grant it most work under ILM, but they use there own favorite software...FumeFX is a biggie for most. Since ILM won't share their software for use the smaller companies have to turn to alternate software.

[Aboubakr]

hi every one!

i have some R&D to share with u, and i need any idea or suggestion :

http://www.youtube.com/watch?v=hfsIX69rtgg

thanks in advance

new update of the helicopter

http://www.youtube.com/watch?v=uZ4BRs37dog

and this is the reference that i'm following

http://www.youtube.com/watch?v=jwr8a-5MruE

[draxysd]

Hi everyone. First post in this topic, yay :D

Anyhow, I wanted to access objects that are in ffx simulation, so I used the $Fumefx001.getObject index
This is all fine and dandy, but I need to find out the name of the object I am acessing, and it seems there isn't a way to do it. All I can get when i use show on the object is 3 parameters (use, speedmultpilier and objecttype), but nothing else. I can see that the object type is ReferenceTarget, can anyone explain in short whate ReferenceTargets are, and how you use them. I have run into them quite a lot lately, and can't seem to get a grasp on them. They seem to have a lot of hidden functions and properties (or so it seems to me). Is there a way to find out all the properties and methods without documentation (maxscript listener alone).

Thanks in advance,
Dragan Najdenov

[kogden]

Hi everyone. First post in this topic, yay :D

Anyhow, I wanted to access objects that are in ffx simulation, so I used the $Fumefx001.getObject index
This is all fine and dandy, but I need to find out the name of the object I am acessing, and it seems there isn't a way to do it. All I can get when i use show on the object is 3 parameters (use, speedmultpilier and objecttype), but nothing else. I can see that the object type is ReferenceTarget, can anyone explain in short whate ReferenceTargets are, and how you use them. I have run into them quite a lot lately, and can't seem to get a grasp on them. They seem to have a lot of hidden functions and properties (or so it seems to me). Is there a way to find out all the properties and methods without documentation (maxscript listener alone).

Thanks in advance,
Dragan Najdenov

Its been a while since I threw somemaxscript at FumeFX....i have a feeling you have to work backwards from the object back to fumefx to find what you are looking for... recursive function of somesort.... its been a good while though since ive done any maxscript/ fumefx work so im not sure what might have changed with updates from Kresmir...

But i think if you dig around fume/sources etc useing this code, it might help you a little :)....

showproperties <node>

ie showproperties $fumefx01

and

getpropnames <node>

ie getpropnames $fumefx01



hope that helps a little.

Cheers Kieran

[3ak]

Hi everyone. First post in this topic, yay :D

Anyhow, I wanted to access objects that are in ffx simulation, so I used the $Fumefx001.getObject index
This is all fine and dandy, but I need to find out the name of the object I am acessing, and it seems there isn't a way to do it. All I can get when i use show on the object is 3 parameters (use, speedmultpilier and objecttype), but nothing else. I can see that the object type is ReferenceTarget, can anyone explain in short whate ReferenceTargets are, and how you use them. I have run into them quite a lot lately, and can't seem to get a grasp on them. They seem to have a lot of hidden functions and properties (or so it seems to me). Is there a way to find out all the properties and methods without documentation (maxscript listener alone).

Thanks in advance,
Dragan Najdenov

This should help (select ffx grid first):
for i = 0 to ($.numobjects() -1) do print ((refs.dependson ($.getobject i))[3].name)

[draxysd]

Right, forgot about the dependents. Thanks a million.

Dragan Najdenov

[jigu]

Hello Everyone,

I am doing simulation on new SSD and it got full. So next day I deleted some last 100 frames of sim data. Now when I try to resume simulation in fumefx, it starts out from 200 frame instead of 120th frame. Is there a way to tell fumefx to read last frame 120 and start continuing sim from there?

Thanks,
Regards,
Jigu

[Mills]

What is the workflow for running many simulation of the same fumefx grid to see changes in settings. Something like "FFX MultiRun" I cant find a link for the script.

I would like to run a number of simulation in one max file, similar to these movies:

https://vimeo.com/20700366

https://vimeo.com/20701396

Cheers!

[JohnnyRandom]

Hmm, I think I know what your talking about, LOL Fabian Buckreus, Jeff Lim, and I have each written one, in fact I can't remeber for sure but I think someone else has written one as well.

Fabians (FumeFX - R&D Maxscript Beta) alter values to a single grid
http://www.afterworks.com/BBoard2012/viewtopic.php?f=12&t=1239

Mine (ComparaSim in the first post) builds an array of grids with different increment values
http://www.afterworks.com/BBoard2012/viewtopic.php?f=14&t=1125
I have not updated the param list for FFX 3.0+, if you want to use this one let me know I will dig in some time to update it. It should not take too long.


Can't find Jeff's :(

[Nickel0531]

hi everyone ,

i have question about zdepth,
as we know Fume has its own camera zdepth ,
I was wondering what if we need world position zdepth .
In maya fluid , there's a way like projection mapping (lassign gradient ramp),
does Fume (or 3ds max) have similar way to reach my need ?

BR

Niko.

[Glacierise]

You can do lots of custom shading stuff if you render with Krakatoa.

[Nickel0531]

You can do lots of custom shading stuff if you render with Krakatoa.

thanks for reply,

Im not really familiar with krakatoa ,
so does that mean ,
we cant do it on Fume but krakatoa ?

[Glacierise]

You can create particles from the voxels of FumeFX and use Krakatoa's magmaflow to assign colors and other channels to them in a smart way, so you can get more shading options than with FFX alone.

[Nickel0531]

You can create particles from the voxels of FumeFX and use Krakatoa's magmaflow to assign colors and other channels to them in a smart way, so you can get more shading options than with FFX alone.

thanks a lot !
that sounds great !
really appreciate that. :)

[Mills]

Hmm, I think I know what your talking about, LOL Fabian Buckreus, Jeff Lim, and I have each written one, in fact I can't remeber for sure but I think someone else has written one as well.

Fabians (FumeFX - R&D Maxscript Beta) alter values to a single grid
http://www.afterworks.com/BBoard2012/viewtopic.php?f=12&t=1239

Mine (ComparaSim in the first post) builds an array of grids with different increment values
http://www.afterworks.com/BBoard2012/viewtopic.php?f=14&t=1125
I have not updated the param list for FFX 3.0+, if you want to use this one let me know I will dig in some time to update it. It should not take too long.


Can't find Jeff's :(

Thanks for these links! I realized that if i send it to back burner with network simulation "on" on all the grids, it does what I want it to. Dont know why I never tried it before.
E

[Daniel-B]

Man, must be nothing interesting happening in the Fume world. This thread has been pretty scarce of comments lately.

[jdrouse]

Did Luma and Sitni Sati release FFX for Maya? Maybe everyone migrated over there and we're just sitting here.

Maybe the 3dsMax gang is paying them not to release it to keep a few Max licenses in the pipelines of Maya studios.

In other news, if somebody wants to donate a sim of fluffy red smoke for an emergency pickup I'll be needing that finished ASAP and I haven't gotten to it yet! heh heh heh.

http://media.naplesnews.com/media/img/photos/2009/10/11/091007NS-GK-lesslethal293_t607.jpg

[Nickel0531]

hi again!

I have question about initial state ,
when im done default simulation with initial state ;
the wavelet sim doesnt seem to right ,
does wavelet sim not allow to use when default cache start from initial state?
or am i missing something ? :sad:

[Glacierise]

@Rouse - I'd swear if you render this out and show to sups, they'll say it needs to be more cauliflower and less low-res CG smoke :D

[JohnnyRandom]

hi again!

I have question about initial state ,
when im done default simulation with initial state ;
the wavelet sim doesnt seem to right ,
does wavelet sim not allow to use when default cache start from initial state?
or am i missing something ? :sad:

No you are not doing anything wrong, it has never worked. I reported this back the end of March last year, it got pushed down the list, I just bumped it maybe we will get a response sometime soon on what is possible/not possible. :)

[JohnnyRandom]

@Rouse - I'd swear if you render this out and show to sups, they'll say it needs to be more cauliflower and less low-res CG smoke :D

LOL funny how everyone thinks ALL smoke is digi now :D I augmented some practicals with some ancillary smoke a little while back, I read some comments on it after it was aired and was funny to read Joe Blow bagging on the all CG smoke and fire. :wip:

[Nickel0531]

No you are not doing anything wrong, it has never worked. I reported this back the end of March last year, it got pushed down the list, I just bumped it maybe we will get a response sometime soon on what is possible/not possible. :)

thank you johnny ;)
hope it could possibly on this part soon.

[galagast]

Hmm, I think I know what your talking about, LOL Fabian Buckreus, Jeff Lim, and I have each written one, in fact I can't remeber for sure but I think someone else has written one as well.

Fabians (FumeFX - R&D Maxscript Beta) alter values to a single grid
http://www.afterworks.com/BBoard2012/viewtopic.php?f=12&t=1239

Mine (ComparaSim in the first post) builds an array of grids with different increment values
http://www.afterworks.com/BBoard2012/viewtopic.php?f=14&t=1125
I have not updated the param list for FFX 3.0+, if you want to use this one let me know I will dig in some time to update it. It should not take too long.


Can't find Jeff's :(

Sorry Johnny, I might have to dig my old script up and slap in a UI while I'm at it >__<

[JohnnyRandom]

Go for it! I'll mess around with it for sure :)

[Daniel-B]

Hey guys. Looks like video copilot might be releasing some FumeFX training, or at least a Fume FX comping tutorial.

http://www.videocopilot.net/blog/2013/02/explosive-training/

I'm not sure if he will go over creating the elements, because those elements look pretty similar to the ones found at...

http://www.finalight.com/-/maindir/webpages/order/_EVT1_HD.php

But I can't be sure. Anyway, looks like a fun tut either way.

[AdrienSliver]

Thanks Daniel !

[AnoPrkl]

I'm trying to achieve something like this (http://www.youtube.com/watch?v=OIYns2tWnzY&t=6m5s). Not red color but the look of the smoke. Haven't really been able to get it look like that. Any tips what should i do to achieve that? How much particles, do i need really low spacing to get that much detail or is wavelet enough?

Thanks.

[finalight]

Hey guys. Looks like video copilot might be releasing some FumeFX training, or at least a Fume FX comping tutorial.

http://www.videocopilot.net/blog/2013/02/explosive-training/

I'm not sure if he will go over creating the elements, because those elements look pretty similar to the ones found at...

http://www.finalight.com/-/maindir/webpages/order/_EVT1_HD.php

But I can't be sure. Anyway, looks like a fun tut either way.
Howdy Daniel,

Interesting - We shall see! Yep, looks like a cool tutorial.

[cojoMan]

I stumbled onto this the other day : https://vimeo.com/22346903
And I get that it was made in Houdini
How would you go about doing something like this in max with fume ?

I have 2 starting points :
1.custom shaped fumefx containers, that after the sim is cached would be linked to a RBD simulation using geometry. But to my knowledge, no such thing as custom shaped containers exist.
2.Simmulate fume. Put a box over the whole sim. Chop it with rayfire, volumebreaker, etc. Sim the pieces. PFlow/TP particles that folow the fume sim, but linked to the RBD as long as it is in the chunk that is flying. This I think it could be pulled off. But how do you render it back USING Fume, not krakatoa, afterwards... ?
3. please sugest if you have any clue...

thanks

[JohnnyRandom]

Right off the bat I would say just prt cache the final sim frame then mesh it with frost, then fracture it. As for the shader well, maybe projection map and bake it. The shader I think would be the most difficult part.

[prieMAX]

hi there...
last few days ago..my office ugraded our fumefx to ver.3.5.1 ...but i got problem..it seems my sim stop at certain frame..without error msg or crash. let say my sim range 0-250 frames. at frame 240..always stop..its like fumefx finished the sim. i checked that my sim range at general tab was correct, at range (0-250). same with time config anim bar 3dsmax. and RAM ok too.
and same thing happen when i sim using wavelet mode...its stop before finished all frames.
our previous ver.3.0.2...and i didnt have this problem (note: i use same scene max).
anyone had same problem like me with ver.3.5.1?
thanks.

[JohnnyRandom]

Does this happen with different scenes?

Does this happen with a default setup extended to 250 frames?

Does this happen with any newly created scenes?

Does this happen when you import the old Fume setup into a new scene?

Have you tried saving out the container as a preset, creating a new one, and apply the preset values to it?

I could be wrong of course, I suspect that it is just a fluke. Let me know if it still occurs after any of the above steps.

[Nickel0531]

hi all!

ive been met wierd situation,
when I let my container constrained to the dummy which has animation,
the simulate result will have horrible blocky noise.

Ive notice that opacity and density would cause that kind of problem ,
but no matter how i raise my opacity or dencity , it still has blocky noise.

another problem that i met before , i render my FFX image in perspective and cam ,
the cam one does have same blocky noise but perspective doesnt see any of these problem.

this is making me crazy, any suggestion ? :sad:

[JohnnyRandom]

Voxels are square. When a voxels density is very high, it will look much like a cube, while looking at it from at different angles will changes how it looks.

When blockiness is an issue you can try:
1. Lower the voxel density, either by emitting less fluid, use an effector to replace density if its too high, raise the dissipation min. density and strength (if constantly producing a lot of smoke).
2. Lower the render opacity(s), tweak the afc curve(s).
3. raise the render jitter amount to >=50 lower the step to <=15
4. simulate with more voxels, ie lower the spacing value.

[Daniel-B]

I think you guys will want to see this. :)

http://www.videocopilot.net/blog/2013/02/tutorial-series-explosive-training/

[AdrienSliver]

it's really cool

[JohnnyRandom]

Who's he kidding "bring it up to 80's ILM quality"? :wip: :D

[Nickel0531]

Voxels are square. When a voxels density is very high, it will look much like a cube, while looking at it from at different angles will changes how it looks.

When blockiness is an issue you can try:
1. Lower the voxel density, either by emitting less fluid, use an effector to replace density if its too high, raise the dissipation min. density and strength (if constantly producing a lot of smoke).
2. Lower the render opacity(s), tweak the afc curve(s).
3. raise the render jitter amount to >=50 lower the step to <=15
4. simulate with more voxels, ie lower the spacing value.

thank you johnny for those helpful advise
ill try those steps to Re-simulate my smoke:)

btw , does systemScale affect block noise too ?
cuz i tried adjust systemScale , and the result was different , even the blockiness still showed up tho , but somehow it became weaker , so i am wondering if it affects too

[finalight]

hi there...
last few days ago..my office ugraded our fumefx to ver.3.5.1 ...but i got problem..it seems my sim stop at certain frame..without error msg or crash. let say my sim range 0-250 frames. at frame 240..always stop..its like fumefx finished the sim. i checked that my sim range at general tab was correct, at range (0-250). same with time config anim bar 3dsmax. and RAM ok too.
and same thing happen when i sim using wavelet mode...its stop before finished all frames.
our previous ver.3.0.2...and i didnt have this problem (note: i use same scene max).
anyone had same problem like me with ver.3.5.1?
thanks.

I've had this problem in 3.5.x when simming to a network drive. Is that your situation, or are you simming to a local drive? To check for it, select the box to keep the simming progress window open after simming completes and it will throw a memory error on the frame it stops at.

[JohnnyRandom]

cuz i tried adjust systemScale , and the result was different ,

.systemScale is simply a global multiplier. It will increase or decrease all parameters equally by the given value.

[maganiac]

Hi people!

Im trying to do a windtunnel with an f1 but im having troubles to do the smoke. Im using fumefx but i don´t manage to have a dense narrow and long smoke column, i just manage to have smoke with mushroom, what should i do??

Just an example of what i whant to achieve,

http://www.listofcarbrands.com/wp-content/uploads/2012/02/pagani-zonda-s-windtunnel.jpg

seems easy thing to do but im running out of ideas to try...

[Nickel0531]

.systemScale is simply a global multiplier. It will increase or decrease all parameters equally by the given value.

i see, i was misunderstanding abt this .
thanks for reply Johnny ;)

[JohnnyRandom]

Hi people!

Im trying to do a windtunnel with an f1 but im having troubles to do the smoke. Im using fumefx but i don´t manage to have a dense narrow and long smoke column, i just manage to have smoke with mushroom, what should i do??

Just an example of what i whant to achieve,

http://www.listofcarbrands.com/wp-content/uploads/2012/02/pagani-zonda-s-windtunnel.jpg

seems easy thing to do but im running out of ideas to try...

Well look at it this way, how does a wind tunnel actually work? Thought about using a wind spacewarp? Or are you manipulating gravity in the the direction you want? Maybe even think a little different, how about just orienting your scene along the z-axis so you car is vertical then just turn your camera. Smoke rises right? :)

no problem Nickel0531, it is useful don't discard using it.

[finalight]

Hi people!

Im trying to do a windtunnel with an f1 but im having troubles to do the smoke. Im using fumefx but i don´t manage to have a dense narrow and long smoke column, i just manage to have smoke with mushroom, what should i do??

Just an example of what i whant to achieve,

http://www.listofcarbrands.com/wp-content/uploads/2012/02/pagani-zonda-s-windtunnel.jpg

seems easy thing to do but im running out of ideas to try...

Another option, if it's just a one-off and doesn't have to behave dynamically (meaning the emitters will be moved around) would be to sim smoke being pushed in straight lines by wind, lower the vorticity so mushrooming is minimal - wind may stretch them out anyway, and then use an FFD after the sim is complete to deform the streams around the contour of the body shape.

[maganiac]

Thanks guys!

Booth ideas seems writes, im gonna try first orient the whole scene in z axis so i can see if results are more realist, but finalight idea is great and very simple which sometimes is not easy thing....

What about smoke settings?

I promiss to put some renders on here, I´m looking forward to see result!! :D

[prieMAX]

Does this happen with different scenes?

Does this happen with a default setup extended to 250 frames?

Does this happen with any newly created scenes?

Does this happen when you import the old Fume setup into a new scene?

Have you tried saving out the container as a preset, creating a new one, and apply the preset values to it?

I could be wrong of course, I suspect that it is just a fluke. Let me know if it still occurs after any of the above steps.


hi sorry for late reply...
Q1: yup...it happen with diff scene..specialy from other old fume scene.
Q2: havent try
Q3: if the scene fresh new with default setting..yes..its seem ok.
Q4: yup..(i merged old one to new scene)
Q5: yup..same result ..fume stop sim.

but had interesting here.... it seem happen when i start increasing/playing with 'spacing' ..
my old scene actualy i set only around 800MB ram, with old fume ok..no problem. but after update to 3.5.1 ..fume cant finished the sim. its only finished the sim, if i brought down the 'spacing' until around 200-300MB ram.
my office computer with 8 gig ram and the cache at netwotk drive.

thanks John for asking..

prie

[prieMAX]

I've had this problem in 3.5.x when simming to a network drive. Is that your situation, or are you simming to a local drive? To check for it, select the box to keep the simming progress window open after simming completes and it will throw a memory error on the frame it stops at.

hi..
yup..i put the cache at network drive. havent try to local drive.

here i copy from log:
Frame: 80
Turbulence Effector is active (mm)
Turbulence Scale Effector is active (nn)
Vmax=6.460, dtReq=0.97, steps=1, CFL'=5.15
Resizing: from (450,456,80) to (452,457,81)
Memory used: 1244Mb
ApplyForces done in 3.87 sec
Advect velocities done.
CG done in 4.83 sec with 50 iterations (167.376 MVOPS).
CG res e:-2.380 m:-2.686
Advect fields done.
I/O done in 3.90 sec
Saving cache data OK.
Frame time: 00:23.26

Frame: 81
Turbulence Effector is active (mm)
Turbulence Scale Effector is active (nn)
Vmax=6.195, dtReq=1.01, steps=1, CFL'=4.94
Resizing: from (452,457,81) to (454,460,81)
Memory used: 1258Mb
ApplyForces done in 3.88 sec
Advect velocities done.
CG done in 4.40 sec with 50 iterations (185.802 MVOPS).
CG res e:-2.396 m:-2.726
Advect fields done.
I/O done in 0.81 sec
Saving cache data FAILED (low RAM).
Frame time: 00:20.23
Simulation terminated.
Flushing FFX Particle Src001
Flushing EFXTR
Flushing FFXt2
Flushing RAM

at frame 81 sim stop. it say low ram..abit weird... memory used only 1258MB.. my actual ram 8Gig.
so fail cos network drive?

[finalight]

hi..
yup..i put the cache at network drive. havent try to local drive.

here i copy from log:
Frame: 80 ...............
..............
I/O done in 0.81 sec
Saving cache data FAILED (low RAM).
Frame time: 00:20.23
Simulation terminated.
..............

at frame 81 sim stop. it say low ram..abit weird... memory used only 1258MB.. my actual ram 8Gig.
so fail cos network drive?
Yep, that's the 'simming to network drive' error I've experienced. Try simming the same scene to a local drive. It's resolved it most every time for me. I've put in an email to Afterworks about it. They replied, however at this point I'm not too sure of the status. It may be on their to-do list. For now, I've been simming to local drives across all farm machines.

[prieMAX]

Yep, that's the 'simming to network drive' error I've experienced. Try simming the same scene to a local drive. It's resolved it most every time for me. I've put in an email to Afterworks about it. They replied, however at this point I'm not too sure of the status. It may be on their to-do list. For now, I've been simming to local drives across all farm machines.


hi Brian ..
wow..its seem u are right...i just finished test with 2 option.
option 1: i put cache at local drive, same computer with fumefx installed, and its seem fume can finished all the frames.

option 2: i put cache at local drive too, but this time using maping network shared folder (consider network drive), same computer with fumefx installed. (cos normaly i put cache at other computer). and the result fume failed to finished all frames..stop at same frame, at frame 81.

so in my case its seem only work with local drive without maping/network link drive.
now i am trying increasing 'spacing' around 1500MB..(and next step will try wavelet sim), cos like my previous reply at Johny, i got interesting ...in my case fume only can finished all frames if only 'spacing' below 300MB, with network drive.

thks Brian

btw, i cant direct join to afterwork or fumefx forum or send email if i got questions, because when we bought fumefx using our CGsup's email.

prie.

[JohnnyRandom]

So out of curiosity, you are caching to a network server, what are the specs of that machine? Is it running a single OS or is it running multiple virtual machines?

[prieMAX]

So out of curiosity, you are caching to a network server, what are the specs of that machine? Is it running a single OS or is it running multiple virtual machines?

hi johnny,
yup..we are using network server, actualy i just used it for network drive..
hmm...i am not sure, i think old HP blade with xeon cpus, and single OS.
anyway like i said before..its ok using network drive before we updated to fumefx 3.5.1, even some scene with low 'spacing' setting (its mean high memory required).

my question is have u try before using network drive for caching with very low 'spacing' setting? or u didnt get this issue even using network drive.? (using fumefx 3.5.1 for sure)

thks

prie.

[JohnnyRandom]

Hmm I will test some tomorrow. I am guessing to say it is at a deeper level than I can dig. If I can verify and create a repro I will submit a bug report for it.

The only work done in the last cycle in relation to caches had to do with listing them in the Asset Browser, this may or may not have something to do with it.

[mcflait]

Hi guys, i am trying to do a galaxy, everything is perfect but i have a Little problem, i am trying to do Brown threads, and i can not do it, i am trying to emit smoke from a particle Meshed and rendered from a pflow and fumefxBIRTH and Krakatoa, but my final result is not the same, maybe my workflow is not the best.
Someone can help me?

[JohnnyRandom]

hi johnny,
yup..we are using network server, actualy i just used it for network drive..
hmm...i am not sure, i think old HP blade with xeon cpus, and single OS.
anyway like i said before..its ok using network drive before we updated to fumefx 3.5.1, even some scene with low 'spacing' setting (its mean high memory required).

my question is have u try before using network drive for caching with very low 'spacing' setting? or u didnt get this issue even using network drive.? (using fumefx 3.5.1 for sure)

thks

prie.

Ok, I am currently writing just under 400mb frame files over a gigabit network with zero issues. I am getting a solid 55% network utilization during the frame write. I am writing to a standard HP nas dl380e 30tb with a single xeon e5 4 core 2.2 and 12 gb ram, so pretty much a low end enterprise nas.

This is a guess but I suspect it is an issue with your frame store, what are the specs of that blade server? How is your network utilization? It takes processor power/mem to send/receive packets, that is usually a limiting factor on one end or the other. I am no network support specialist but I think I would at least verify your network connection to and from is solid.

[JohnnyRandom]

Hi guys, i am trying to do a galaxy, everything is perfect but i have a Little problem, i am trying to do Brown threads, and i can not do it, i am trying to emit smoke from a particle Meshed and rendered from a pflow and fumefxBIRTH and Krakatoa, but my final result is not the same, maybe my workflow is not the best.
Someone can help me?

Have you tried using a Krakatoa PRT FumeFX object? There really should be no reason to use pflow with something like this.

[prieMAX]

Ok, I am currently writing just under 400mb frame files over a gigabit network with zero issues. I am getting a solid 55% network utilization during the frame write. I am writing to a standard HP nas dl380e 30tb with a single xeon e5 4 core 2.2 and 12 gb ram, so pretty much a low end enterprise nas.

This is a guess but I suspect it is an issue with your frame store, what are the specs of that blade server? How is your network utilization? It takes processor power/mem to send/receive packets, that is usually a limiting factor on one end or the other. I am no network support specialist but I think I would at least verify your network connection to and from is solid.

hi Johnny,
thanks for did the test. like my previous reply. at my side, fumefx also running well via network drive if 'spacing' around 300mb.
like i said, the problem is with the same networks configuration (including network drive, giga switch and server,.ect.) the previous version (fumefx 3.0.2) didnt have this issue, even i pumped the 'spacing' around 1000mb. now with fumefx 3.5.1 if i start pump 'spacing' above 500mb fumefx failed to finish all frames.
and dont forget i did 2 test like Brian suggestion (using local drive).
(1) i did put cache at local drive, and fumefx ran well..can finished all frames. even i pumped 'spacing' above 500mb (yesterday test around 1200mb).
(2) i did put cache at local drive too, but using shared folder, consider network drive. so basicly its just looping network to my local drive. and fumefx was failed to finish all frames.(with 'spacing' above 500mb).

btw, i will ask to my IT guys about our networks configuration, cos i cant answer ur questions now.., i also not network specialist too... hehe. and i also suspect this issue maybe 'cos network problem too..but like i said with same network configs, the previous fumefx didnt have this issue.

again ..thanks alot Johnny..

prie.

[JohnnyRandom]

Yeah I read all previously. Not sure what to say.:shrug::sad: I typically always render/sim to a local network share. I am trying to reproduce it. I was simming up 1600mb, my grid was 150x50x150 at 0.25 without issue. I can fire one off tomorrow night when I leave work at a higher res but I am not going to wait around for something like that to cook.

You or Brian can email me a stripped scene file that exhibits the behavior and I can see if I can reproduce the issue, else I am just shooting in the dark trying to get it to fail.

[mcflait]

Have you tried using a Krakatoa PRT FumeFX object? There really should be no reason to use pflow with something like this.

Hi Jhonny, thanks so much for your answer.
Yes, i have tried your way, but with PRTFumeFX and with Pflow way, i have the same problem, the simulations are with the same density when i am rendering with Krakatoa.

[prieMAX]

Yeah I read all previously. Not sure what to say.:shrug::sad: I typically always render/sim to a local network share. I am trying to reproduce it. I was simming up 1600mb, my grid was 150x50x150 at 0.25 without issue. I can fire one off tomorrow night when I leave work at a higher res but I am not going to wait around for something like that to cook.

You or Brian can email me a stripped scene file that exhibits the behavior and I can see if I can reproduce the issue, else I am just shooting in the dark trying to get it to fail.

hi johnny..
i got new scene test (different scene but same problem with network drive), but this time i put all channels and 'sensitivity' using velocity. and guest what, even it say only take sim memory around 500mb...but it eat my ram more than 8gig...and cache file around 200mb -1gig per frame. specialy after the sim out the fumefx cage..i mean start using 'boundless', the cache files become bigger and bigger.
i will send the link for the scene file via PM ... i put at my dropbox. pls check ur message folder.
just load the scene file and change the folder path for cache(pls using network path).. and sim it....let see what happen..hehe.
thanks alot.

info scene:
i am using 3dsmax 2012, it shockwave smoke... drive by pFlow. this test scene actualy i want to see 'bondless' feature.. and i want to compare with N-sim.

again thanks alot

prie

[JohnnyRandom]

Well I am only on frame 50 of your scene and currently writing 2+ gb files straight to a network attached storage (the same one I originally tested on) without any issues. The max process is using 10 gb (private working mem) and ~90% cpu. getting about 65% network utilization at write time.

I tried another at home last last night to a local network shortcut without issue. I will try another tonight writing to a different machine at home and see what happens.

http://i.imgur.com/2zsQfkC.png

[JohnnyRandom]

I tested across a VPN network from home to work, slow as mud but still threw zero errors. I tested over wireless and wired connections.

I am either blessed by the gods of networking or there is just something plainly wrong with your guy's networks/credentials/scene setups/pathing something. I cannot get a failure UNLESS I run out of system resources, my network creds were bad, or ran out of disk space. You aren't running a "demo" version are you?

[prieMAX]

Well I am only on frame 50 of your scene and currently writing 2+ gb files straight to a network attached storage (the same one I originally tested on) without any issues. The max process is using 10 gb (private working mem) and ~90% cpu. getting about 65% network utilization at write time.

I tried another at home last last night to a local network shortcut without issue. I will try another tonight writing to a different machine at home and see what happens.
(image deleted)


thanks johnny for test...
hmm... i also speechless...why u got no issue..
anyway i think u can change 'threshold' sensitivity to above 0.6 (let say 1.0 or 1.2) to minimize file size.

thanks

prie.

[prieMAX]

I tested across a VPN network from home to work, slow as mud but still threw zero errors. I tested over wireless and wired connections.

I am either blessed by the gods of networking or there is just something plainly wrong with your guy's networks/credentials/scene setups/pathing something. I cannot get a failure UNLESS I run out of system resources, my network creds were bad, or ran out of disk space. You aren't running a "demo" version are you?

well..i believe we are running full version.
we bought ver 3.0.2 ..and we download update 3.5.1, the file name FumeFX35R2012.EXE
i think this one not a demo version.
there is anyway i can check..either demo or not?

and our network/IT guy said..there's nothing wrong with our network, its seem the issue come from the update it self. now we still doing tests.

thanks

prie

[finalight]

well..i believe we are running full version.
we bought ver 3.0.2 ..and we download update 3.5.1, the file name FumeFX35R2012.EXE
i think this one not a demo version.
there is anyway i can check..either demo or not?

and our network/IT guy said..there's nothing wrong with our network, its seem the issue come from the update it self. now we still doing tests.

thanks

prie

Hi guys,

I'm still leaning toward it being something with the update. All previous versions of Ffx were caching across various networks fine for me until the 3.5.x update. Btw, what OS are you running with both successes and failure? Win7? XP? Might have something to do with it.

[JohnnyRandom]

You can check in the UI header or the about box. That wasn't exactly what I meant by "demo" :) Sorry, no offense, I had to ask, I helped someone with another issue a little while back that was directly related to using a less than legitimate version.

Why would I want a smaller file size :D We are testing network related issues, the bigger the better.


Me win7 x64 SP1 on all machines. EDIT: Hmm, I have been testing in 2011&2013, I don't use 2012. Brian are you using 2012 too?

This is just the first I have heard about it. No one on the beta board has reported such issues, no one has reported anything in the bug tracker either, or anyone else but you two. That is why I am curious. I am not saying that is doesn't have to do with it, it definitely may be something with 3.5.x, I just can't repro it.


So here is another question, if you sim locally, copy the cache to your network drive, then render, do you have any issues?

[finalight]

You can check in the UI header or the about box. That wasn't exactly what I meant by "demo" :) Sorry, no offense, I had to ask, I helped someone with another issue a little while back that was directly related to using a less than legitimate version.

Why would I want a smaller file size :D We are testing network related issues, the bigger the better.


Me win7 x64 SP1 on all machines. EDIT: Hmm, I have been testing in 2011&2013, I don't use 2012. Brian are you using 2012 too?

This is just the first I have heard about it. No one on the beta board has reported such issues, no one has reported anything in the bug tracker either, or anyone else but you two. That is why I am curious. I am not saying that is doesn't have to do with it, it definitely may be something with 3.5.x, I just can't repro it.


So here is another question, if you sim locally, copy the cache to your network drive, then render, do you have any issues?

Heya John,

No worries - thx very much for the help - very much appreciated.

It looks like this is occurring on both WinXP x64 and also Win7 x64. When running local sims, network rendering works fine across the farm, with all nodes reading cache from the shared local drive. This has also occurred on RAID drives, USB connection, SATA connection, internal, external. Both UNC path and non.

I have an email into Kres about it. I also haven't seen anything on the Afterworks forum.

[prieMAX]

You can check in the UI header or the about box. That wasn't exactly what I meant by "demo" :) Sorry, no offense, I had to ask, I helped someone with another issue a little while back that was directly related to using a less than legitimate version.

Why would I want a smaller file size :D We are testing network related issues, the bigger the better.


Me win7 x64 SP1 on all machines. EDIT: Hmm, I have been testing in 2011&2013, I don't use 2012. Brian are you using 2012 too?

This is just the first I have heard about it. No one on the beta board has reported such issues, no one has reported anything in the bug tracker either, or anyone else but you two. That is why I am curious. I am not saying that is doesn't have to do with it, it definitely may be something with 3.5.x, I just can't repro it.


So here is another question, if you sim locally, copy the cache to your network drive, then render, do you have any issues?

hi johnny...
yup..no worry...hehe..., and i checked already...and our's legit.

ok..then, u can fire 'em up with big file size then.. hehe.

our system still using XP 64.

yup..just finished copy to network all the local files cache... and change the path to that network. and its seem rendering fine...no issues. (using the scene i sended to you, i sim over the weekend with threshold 2.0, 36gig space only, and local cache path., all 100 frames done, with no issues)

thanks johnny..


prie.

[prieMAX]

Heya John,

No worries - thx very much for the help - very much appreciated.

It looks like this is occurring on both WinXP x64 and also Win7 x64. When running local sims, network rendering works fine across the farm, with all nodes reading cache from the shared local drive. This has also occurred on RAID drives, USB connection, SATA connection, internal, external. Both UNC path and non.

I have an email into Kres about it. I also haven't seen anything on the Afterworks forum.


hi Brian,
lol..so only us had this issue.... hehe

prie.

[finalight]

hi Brian,
lol..so only us had this issue.... hehe

prie.

There are probably more, but some users I spoke to only sim to local so they just don't know they have the issue ;]

I haven't had a chance, but I should also test it out further such as:
--only one type of channel being written?
--tweaks on various params?
--stop/continue active or no
--etc.

[JohnnyRandom]

I know for a fact a few big studios sim to network storage, usually multiple sims concurrently, I would think we would hear something, eventually. We'll see what Kreso has to say when he gets back from holiday.

[finalight]

I know for a fact a few big studios sim to network storage, usually multiple sims concurrently, I would think we would hear something, eventually. We'll see what Kreso has to say when he gets back from holiday.

Yep, I'm sure they do - Final Light isn't a large VFX house but I've always simmed to network storage as well ;] (or wish I still could for that matter). My best guess is that something in the code changed and some of us, few though it may be, are getting this memory/network error. It all started with 3.5.x. Haven't had an issue before that all the way back to Alpha stage.

The last response from Kreso was that he couldn't reproduce it and that he'd keep an eye on it. I owed him a test scene and got tied up with projects, so the ball was in my court. Guess I'll ping him again since I see I haven't gone mad and other's are getting the error.

[FabianB]

Hey guys,

I've updated my script to work with Fume 3.5.1 and added some cool new features:

http://fumefx-training.com/training/IgnitionBeta044.zip

Here is a little tutorial on how to use it:
https://vimeo.com/61607996

And here is the old video if you haven't seen it already:
https://vimeo.com/24329699

The new version has full spacewarp support, so you can add in wind and increment the strength to see how it affects the sim.
You can also do the tests over multiple grids, just add them all in!

cheers,
Fabian

[Nickel0531]

hello guys ,

I was wondering if Fume could do simulation on thinkbox deadline ?

ive searched online but no result came out :argh:

[AdrienSliver]

thanks a lot fabian !

[JohnnyRandom]

Nice Fabian! :)

hello guys ,

I was wondering if Fume could do simulation on thinkbox deadline ?

ive searched online but no result came out :argh:

You need to submit to deadline via maxscript. Or from the help:

Can I Perform Fume FX Simulations With Deadline? Yes. To do so, follow these steps:
Your render nodes need to have Fume FX licensed properly, either with a "full" or "simulation" licenses. This requirement is the same if you were rendering with Backburner.
Before you launch the 3dsmax submission script, make sure that the Fume FX NetRender toggle button is "ON" in the Fume FX options in 3dsmax.
Before you submit the job, make sure the "Disable Progress Update Timeout" option is enabled under the Render in the 3dsmax submission window.

[Nickel0531]

Nice Fabian! :)



You need to submit to deadline via maxscript. Or from the help:





thank you again Johnny !
really appreciate with all the reply u did :thumbsup:

[frint]

hi all, I have a problem here.

I read through this thread but, i could not find the solution so, please let me ask new question.

Questions
1. Can I render fumeFX velocity channel with camera motion embdedd?

2. How can I render fumeFX velocity or zDepth channel, unclamped to openEXR.


I tried most of things but couldnt figure it out by my self.
Enyone got good idea? thnx

[frint]

I found out that, fumeFX velocity channel is opposite with VRay velocity.
Snd it is opposite with ReelSmartMotionBlur too(VRay and RSMB is same).

So you need to set the Ydirection -1.0 at RSMB.

Im not sure about this though...

[procanic]

has anyone come across the problem that an object w/ mr matte/shadow/reflection shader will not recieve any shadows on areas that are inside a fumefx container?

http://i1270.photobucket.com/albums/jj608/procanic/ffx_shadow_test_01_small_marked.jpg


(spotlight w/ raytraced shadows, autovolume on, ffx cast/recieve shadows on, atmosphere shadows on. standard mat recieves shadows just fine, the mr matte shader only on areas outside the grid.)

am i missing something or is the mr matte/shadow/refl really not 100% compatible w/ fumefx?

my fix for now would be rendering the ground shadow pass w/ scanline, so any suggestion on how i could fix the shadow using mr is most appreciated.

[JohnnyRandom]

1. Can I render fumeFX velocity channel with camera motion embdedd?

2. How can I render fumeFX velocity or zDepth channel, unclamped to openEXR.


1. Just be sure to include the velocity channel in your simulation. FumeFX supports Image mBlur only, enable it in the Environment Dialog->Atmosphere->Fusionworks Renderer->Create Channels->Image Motion Blur

2. IIRC FFX_Zdepth and FFX_Velocity Render Elements are only 16-bit, not full float, they do not render unclamped, all others should render unclamped (ie FFX_Fire, FFX_Smoke, diffuse, ect)

As for the direction, I think I have heard some different things, ie Nuke specifically the velocity direction is wrong. I can't confirm or deny that though :D


@procanic, hmm never use mental ray, if I were to guess, it is a bug. Have you tried different shadow types?

[frint]

hello Jhonny, thank you for reply.

1. Just be sure to include the velocity channel in your simulation. FumeFX supports Image mBlur only, enable it in the Environment Dialog->Atmosphere->Fusionworks Renderer->Create Channels->Image Motion Blur

I tired this but I didnt get the camera motion.
When my camera is panning fast, other velocities like Vray Velocity will create the vector of the camera pan. But at fusionworks Renderer, it wont. Will It seems to me though.
May be Fusionworks Renderer wont support camera motion.

About clamping, thank you for the answer! Ill stop testing these things.
And about Direction of velocity, im not familier to Nuke but the color of velocity element rendered was opposite besides Fusionworks and VRay.

thank you !

[procanic]

thanks for the suggestion, john. (and for the fx floater & ffx bbq as well on this occasion! ;) )
yes, i've tried all available shadow types. couldn't try AB shadows though, cause we don't have any AB licences :/ ...
from what i've heard/read, most people seem to use vray or fr with fume. i might be able to get a vray1.5 license for this project. just wondering if there are also any drawbacks using vray w/ fumefx, since i found that using scanline or mr w/ fume has pros and cons either way ...
does vray support the ffx smoke & fire render elements?

@frint:
i did a quick test and can confirm that fusionworks' velocity pass doesn't take camera motion into account at all. can vray's velocity pass give you a correct result with fumefx' internal velocities + camera motion?
really interested! would be one more reason to get a vray license ;) ...

[JohnnyRandom]

hello Jhonny, thank you for reply.
May be Fusionworks Renderer wont support camera motion.


The only way I know of to generate camera motion blur on the FFX Velocity render element is to use multi-pass in the camera.

[JohnnyRandom]

thanks for the suggestion, john. (and for the fx floater & ffx bbq as well on this occasion! ;) )
yes, i've tried all available shadow types. couldn't try AB shadows though, cause we don't have any AB licences :/ ...
from what i've heard/read, most people seem to use vray or fr with fume. i might be able to get a vray1.5 license for this project. just wondering if there are also any drawbacks using vray w/ fumefx, since i found that using scanline or mr w/ fume has pros and cons either way ...
does vray support the ffx smoke & fire render elements?


Sorry I missed your post, small bit of history, the only reason mr is supported was simply because it is a "shipping" render and many people asked for it. Its problems typically stem from the fact the everything from max needs to be translated into something mr can read. Other 3rd party renderers are more native, meaning they don't require a translator. It would be like using renderMan with max.

VRay works really well with Fume (probably because they have Phoenix) fR works well to, admittedly I haven't used it nearly as much as I have with Vray. That said I am still not sure you are going to get complete velocity vectors in the RE's.

[procanic]

thanks for the history update on why mr sucks w/ fume :D - totally makes sense & i think i'll have to try to organize a vray lic ...

The only way I know of to generate camera motion blur on the FFX Velocity render element is to use multi-pass in the camera.
yeah, that's what i thought as well. tried that while doing that quick test for frint. the thing is, that it just blurs the velocity channel, thus lowering the color intensity. here's what it looks like:

http://i1270.photobucket.com/albums/jj608/procanic/fume_velocityPass_multipass.jpg


should't it just alter/intensify the colors (motion information) instead while keeping the shape of an unblured object/fluid?
trying to achieve that i also set the duration in the multi pass effect to zero, hoping that it just reads the cam motion as information for the velocity pass without applying any motion blur to RGBA or any passes. result in the velocity channel was the same (color wise) as without the multi-pass effect ...

guess if you want to have camera motion blur you'll have to render it directly into the RGBA or do it in post without a vector pass. or am i missing something here?

[JohnnyRandom]

I suppose it should be using Additive when it adds each pass, instead it is more like screen or overlay but yeah it is just adding each subsequent pass on top of the old, which is the only reason why it "works" in the first place.

I suspect you may be able to do something fancy via script. Possibley try exporting a helper to the compositor your using, get the motion vector of the camera helper at time A & B and calculate the length and direction then apply a vector blur per frame, I do not know how acceptable vector blur plugins are to scripted input as I have never tried it, this just seems one way you could go about doing it.

BTW:
Tool for Automated Dust Trails (http://procanic.blogspot.com/2012/07/tool-for-automated-dust-trails.html)

Cool! :D

[procanic]

thanks for all the suggestions, john - really appreciate it! i talked to the comp guys about your idea of calculating the v-pass in post from cam motion. could be possible - at least they said they'll give it a shot & see what they can come up with ;) ...
glad you like that old dust trail tool - your post doubled my blog's daily visitors rate for the last two days. :D i'd love to extend it to use fluids as well, since right now it's all just particle based. been thinking about a private rewrite since quite some time (so i could make it public) but have no clue how to make time for that currently :shrug: ...

[frint]

procanic, thank you very much for the tests!

and for next step, ill try create vector with the helper in post.
with ReelSmartMotionBlur, you can make custom vector with the track point, so, Ill try that.
And composite the velocity with the FusionWorks velocity.
Or you can render some plane at the fume grid, and use that vector for merge with.

But, i think its not that worth....
Its easier to do with no v-pass. and it works pretty fine. As you say,
guess if you want to have camera motion blur you'll have to render it directly into the RGBA or do it in post without a vector pass.
And, for the question above:
can vray's velocity pass give you a correct result with fumefx' internal velocities + camera motion?
The answer is No, for 3dsmax VRay( Im not sure for VRay for Maya).
fumeFX would not appear to VRay Velocity pass at all.... :(
Also, VRay RaytraceShadow cant cast the object shadows to fume...

again, thank you Jhony, procanic!! I really helped.

cool blog ;)

[Glacierise]

Rendering with Krakatoa gives great motionblur too :)

[JohnnyRandom]

Very good point, if they could get faster speeds out of Voxel mode it would be much more viable IMO and getting the FFX Color channel piped into the PRT FumeFX would be f'n huge! As it is having Magma to tweak with is awesome and all of the other love that comes with Krakatoa, a few more things and that pipeline would be solid. :)

[Glacierise]

Stoke would help with that. You can pick up UVs from emitters, emit from separate filtered sources, add motion with particles, partition. KCM. Lots of options!

[JohnnyRandom]

Pretty sure Stoke would also benefit from access to that :D Only born new particles based on color ;) Color derived from temp, or velocity, or turbulence, ect. Dude oysters lots of oysters!

[Glacierise]

It would be useful :)

[tamagochy]

My R'n'D of the big explosion...

http://www.youtube.com/watch?v=lSq82DWyydk

[Debneyink]

I'm doing a plane flying low to the ground and wanted to add the dust being whipped up...

i saw a test recently on this thread of a helicopter it looked really good just wondering if anyone could give me a few tips...

do i sim the fume on the ground plane and set inital state?

any tips welcome!


cheers

debneyink

[Shletten]

My R'n'D of the big explosion...

http://www.youtube.com/watch?v=lSq82DWyydk
Nice job! I dig it. Kinda wish I could make something similar but I haven't tried FumeFX seriously yet.

[Tollman]

I want to find out what simulation mode is set in the ui.
Essentially what i want is when i (with our inhouse tool) send the file to be simmed in backburner, and if the simulation mode is set to wavelet, and the wavelet path is set to "c:\" it should give me an error before submitting.
But i dont want it to care about the wavelet if i just want to simulate the default grid.
I haven't found script solutions to this, or any other way to solve it.

Anyone has a solution for this?

[Kresimir]

Hello,

I am glad to inform you that FumeFX 3.5.2 is out and it fixes the &quot;Low RAM&quot; issue.
It seems that on some systems the file write operation can time out, so we made saving in smaller chunks to keep the connection alive.

Regards,

Kresimir Tkalcec


thanks johnny for test...
hmm... i also speechless...why u got no issue..
anyway i think u can change 'threshold' sensitivity to above 0.6 (let say 1.0 or 1.2) to minimize file size.

thanks

prie.

[JohnnyRandom]

I want to find out what simulation mode is set in the ui.
Essentially what i want is when i (with our inhouse tool) send the file to be simmed in backburner, and if the simulation mode is set to wavelet, and the wavelet path is set to "c:\" it should give me an error before submitting.
But i dont want it to care about the wavelet if i just want to simulate the default grid.
I haven't found script solutions to this, or any other way to solve it.

Anyone has a solution for this?

You can do somehting like this, it will print out the state of each of the given params. You could easily set the states to in the same script it should be straightforward enough as i have listed what each integer is equal to which mode.


for i in objects where classOf i == fumefx do
(
format "<---Grid %--->\n" i.name --Current Grids Name
curCache = case i.SelectedCache of --Current Cache in use
(
0: "Default"
1: "Initial State"
2: "Wavelet"
3: "Post"
)
format " |Cache - % \n" curCache
curMode = case i.SimMode of --Current Simulation Mode
(
0: "Default"
1: "Initial State"
2: "Wavelet"
3: "Post"
)
format " |Mode - %\n" curMode
format " |Paths: \n"
for p in 0 to 4 do --Current Grids Paths
(
pathType = case p of
(
0: (" Default - ")+(i.GetPath "default" as string)
1: (" Wavelet - ")+(i.GetPath "wavelet" as string)
2: (" Post - ")+(i.GetPath "retimer" as string)
3: (" I. Map - ")+(i.GetPath "illummap" as string)
4: (" Preview - ")+(i.GetPath #Preview as string)
)
format " % \n" pathType
)
curNet = i.BackBurnerSim --BackBurner mode
format " |BB Mode - %\n " curNet
)

[prieMAX]

Hello,

I am glad to inform you that FumeFX 3.5.2 is out and it fixes the &quot;Low RAM&quot; issue.
It seems that on some systems the file write operation can time out, so we made saving in smaller chunks to keep the connection alive.

Regards,

Kresimir Tkalcec

at last...hehe....thanks alot...
hopely now ..we/ i can go back using network drive....
and thanks to johnnyrandom too for testing our scene.
again thanks alot.

prie

[Tollman]

You can do somehting like this, it will print out the state of each of the given params. You could easily set the states to in the same script it should be straightforward enough as i have listed what each integer is equal to which mode.


Thanks!
The "i.SimMode" flag was what i was searching for.

[neomass]

Hello! I apologize in advance if this question has already sounded on this forum. I have a problem I do not know how to solve. I need to make animation were some burning object is passing trought the camera on a hi speed. In the middle of my movie i need to make a big slowmotion like in the film such as the matrix. Tell me how to do it correctly? I crashing my brain trying to slove it more than a week and cant find the answer by myself.
I tried to do the very slow burning object with key animation and a decrease in the parameter Time-stsale but in this case I disappear a trace of fire. And the slowing effect is lost. Thank you very much for your fast reply.
here is the shot http://www.sendspace.com/file/2w5sr8

[Phlok]

I've got an issue with Max 2013 crashing each time I want to make a preview out of FumeFX (Preview -> Make Preview). The studio here is on FFX 2.1c and I've checked if there were any problems with user rights, but there aren't. Has anyone else seen this behaviour?

[JohnnyRandom]

Not me sorry, I haven't even run 2.x in 2013.

Do you get to the pick codec window?

What codec are you trying to wrap the preview with?

If so, you are trying to use quicktime I would try a standard uncompressed avi or even xvid if you want mp4.

[Phlok]

I've tried Video1 as well as uncompressed. Neither did work.
I did tests on two different machines, both running Max2013 and FumeFX 2.1c and both are crashing when trying to create a preview.

[Snake Eyess]

create preview tends to randomly do that...
One scene it will make the preview and then in another it will just refuse

I found 2 things that sometimes solve the problem.
A) render to a Sequence (I.E. tga or JPG)
B) lower the output size under "Percent of Output"

[Phlok]

There was a reproducable crash on two different machines using Max 2013/ FFX 2.1c After I installed Max 2012, Making Previews out of FumeFX works like a delight.

[JohnnyRandom]

Really strange, sorry you had to go back a version but at least it is working.

3.x has been out for two years, out of curiosity is there a specific reason you guys haven't upgraded yet?

[Phlok]

They don't do this kind of effects here very often and there is no specific reason to update for the current project.

[Daniel-B]

I haven't done an explosion sim in a while, but I've finally almost completely eliminated the dreaded "bubbling" expansion at the beginning of explosion sims. This one isn't 100% what I want, but I feel it's almost there.

https://vimeo.com/64288638

Edit: Uploaded my latest sim scene file if any of you are interested in playing with it.

https://dl.dropboxusercontent.com/u/73226790/explosion_4_17_2013.max

[fireknght2]

I haven't done an explosion sim in a while, but I've finally almost completely eliminated the dreaded "bubbling" expansion at the beginning of explosion sims. This one isn't 100% what I want, but I feel it's almost there.

https://vimeo.com/64288638

It appears to be fine to me depending on the fuel you may or may not get a bubble effect, I like what you've done...:)

Rich

[Daniel-B]

It appears to be fine to me depending on the fuel you may or may not get a bubble effect, I like what you've done...:)

Rich

What I mean about bubble effects is that unturbed leading edge on fast moving fluid simulations that always looks really fake. Take my previous test for example. The explosion looks good mid way through, but the beginning is just 4 or 5 hollow looking fire shaded bubble shapes, that looks bad.

https://vimeo.com/47798297

And back to this one today, the beginning feels more natural I think, but I can still make it better I believe with a little bit of tweaking.

https://vimeo.com/64288638

Edit: Uploaded my latest sim scene file if any of you are interested in playing with it.

https://dl.dropboxusercontent.com/u/73226790/explosion_4_17_2013.max

[fireknght2]

What I mean about bubble effects is that unturbed leading edge on fast moving fluid simulations that always looks really fake. Take my previous test for example. The explosion looks good mid way through, but the beginning is just 4 or 5 hollow looking fire shaded bubble shapes, that looks bad.

https://vimeo.com/47798297

And back to this one today, the beginning feels more natural I think, but I can still make it better I believe with a little bit of tweaking.

https://vimeo.com/64288638

I'm seeing that effect your talking about appears to me more of rounded effect instead of the crisp presentation you have in the second video. Tricky thing fire and explosions, seems as if you got them under control though.

Rich

[prieMAX]

after my office downloaded the update fumefx 3.5.2 ...and try it.
so far so good.. can use back network drive for caching without errors.
fumefx can finished all frames.

thanks sitni sati for update, thanks johnnyrandom for testing our scene.

regards,
prie.

[adom86]

Strange! This week I get terrible viewport performance when having Fume preview window open under GPU and non-GPU previews. Anyone had this before? I have to set viewport update to 0 as it adds about 10 seconds on per frame!

Fume 3.5.1
Quadro 4000 (old one)
Max 2012

Shall login and try the latest fume see if any better.

[JohnnyRandom]

Unless you are prototyping, r&d, look dev, I leave the preview window closed. It has to read the data from mem or disk then draw. If using GPU you could be calling for objects and/or shadows too that all takes time to calculate.

If you are using the viewport visualization tools (ie VP velocity, smoke, fire, ect) you will make it even slower! You are throwing soooo much data at both the viewport and the preview window it just gobbles up CPU/GPU just to display. You are robbing power from simulation cycles.

Seriously you should never forget to close your preview window and disable VP display when you run a sim, particlulary if you are not going to watch the paint dry, it can cost you quite a bit of time.

All that said in 3.5.2 there were issues addressed directly related to GPU Preview window performance so you should see some gains.

[adom86]

Thanks for the tips! Never had this issue before but I have updated fume and graphics drivers, see a drastic improvement although I did install it on max 2014. Shall see if improved on 2012 tomorrow. Hopefully coming off 2012 at work, be glad to see the back of it :)

[lordyanting]

hi guys, I had come across this problem and couldn't find any solution else where, let say I had an animation of 45 frame, after I post processing it retime in 0.5, I got a end result of 88 frame instead of 90 frame, is there any way to fix it?tried with other scenes as well there's always 2 - 3 frame missing.

[JohnnyRandom]

Did you try re-timing with a value of 0.55? instead of 0.5.

Just a guess, I would expect with any value less than 1.0 you are creating frames based off of interpolated values from the frame before and the frame after, you might always be 1 less frame.

[lordyanting]

nope didn't thought of this since my fume is tied with an animation so it would probably offset if retime value is different, but no harm trying though thanks!

[Phlok]

Is it possible to make FumeFX interact with a particle system?

I've tried the following using TP, but without success:

1) Shot some particles as standard shapes into the fume conatiner, unfortunately, I couldn't select the TP system as collision Object inside TP's Objects Tab, so this didn't do anything.

2) Instead, I instanced a sphere to the particles, selected that sphere as collision object, but this didn't affect the fluid, either.

Is it possible somehow to influence FFX with a TP system? Did I maybe miss something in my setup? Any other ideas?

[silwerfeldt]

Is it possible to make FumeFX interact with a particle system?

I've tried the following using TP, but without success:

1) Shot some particles as standard shapes into the fume conatiner, unfortunately, I couldn't select the TP system as collision Object inside TP's Objects Tab, so this didn't do anything.

2) Instead, I instanced a sphere to the particles, selected that sphere as collision object, but this didn't affect the fluid, either.

Is it possible somehow to influence FFX with a TP system? Did I maybe miss something in my setup? Any other ideas?

Maybe you can use a mesher on your particle system. I´m not infront my Computer right now so i can´t test ut.. I think it should work.

[cojoMan]

Is it possible to make FumeFX interact with a particle system?

I've tried the following using TP, but without success:

1) Shot some particles as standard shapes into the fume conatiner, unfortunately, I couldn't select the TP system as collision Object inside TP's Objects Tab, so this didn't do anything.

2) Instead, I instanced a sphere to the particles, selected that sphere as collision object, but this didn't affect the fluid, either.

Is it possible somehow to influence FFX with a TP system? Did I maybe miss something in my setup? Any other ideas?



in the master dynamic setup - use the "group as objects" checkbox
they will be selectable by group as deformable geometry - also you will be able to pick them in the mesher, use modifiers on top /etc

ps.also, you can use instead of the colliding geometry - only the particles' velocities - you might only want that, so you can drive your simulation forces by those inside TP.

[Phlok]

Thank you very much, guys. Exactly what i was looking for!!

[Rareden]

Is there a way to control boyancy over age? like with a volcano it will rise a certain amount then slow its accent and get blown by the wind

[Nickel0531]

Hi guys!

I was doing reserch about rendering ffx with vray,
is there any advantage using GI render smoke only ? (no fire only smoke. )
cuz I put vray HDRI on GI environment,
and the smoke render result does not have any difference between Gi on and off.
am I get any wrong about this ?

[finalight]

Hi guys!

I was doing reserch about rendering ffx with vray,
is there any advantage using GI render smoke only ? (no fire only smoke. )
cuz I put vray HDRI on GI environment,
and the smoke render result does not have any difference between Gi on and off.
am I get any wrong about this ?

I haven't researched this in-depth however I believe that is currently only available in some form with Final Render.

[finalight]

Is there a way to control boyancy over age? like with a volcano it will rise a certain amount then slow its accent and get blown by the wind

This would be a perfect situation for using Effectors.

[Nickel0531]

I haven't researched this in-depth however I believe that is currently only available in some form with Final Render.


too bad to hear that ;s
i asked about this because theres a vray GI multiply on ffx render tab but only works for fire.
so i was wondering if GI effects smoke or not .

anyway, thanks for reply finallight :)

[finalight]

too bad to hear that ;s
i asked about this because theres a vray GI multiply on ffx render tab but only works for fire.
so i was wondering if GI effects smoke or not .

anyway, thanks for reply finallight :)

Very welcome - yes I agree. I've done a bit of testing with using Vray to illuminate smoke as an ambient occlusion type pass. Vray is excellent at casting AO shadowing from smoke onto objects as well as it's fire GI and looks wonderful for fire/smoke explosions comped into shots with very nice integration. Not that you asked, but both Vray and Mental Ray (others?) offer the above however they each provide a different look.

[Phlok]

Does anybody know of a way to eliminate parts of the FFX container after simulation? Like deleting voxel content via scripting.

I am asking because we are creating PFlow particles from FumeFX simulations. For some limitations in PFlow, we are limited to 10M particles per PFlow Partition, which is slightly problematic.

PFlow creates particles in areas off camera, so I am looking for a way to delete those areas out of the FFX container prior to PFlow creating particles.

Thank you in advance for any hints.

[3ak]

you can create second grid with the same (bigger) dimensions as first one, add fumefx source with first (simulated grid) as source, turn off all sim stages in script (forces, advection, projection, combustion) in second grid and add void source to delete all you don't need (or inverse).
BUT. this will work with smoke, temp, vel fields. No fire. there is no burn (fire) field in fumefx source so you can't bring it to the second grid. But if all you need to advect particles vel field is all you need

[JohnnyRandom]

This may or may not work in your case but you could also try to get rid of them on the pflow side. You can cull them with a camera cull op, or data op.

You can also thin the grid with pflow FumeFX test and a delete op. set an inverse of a channel specify min/max values for the voxel and the test will send them out to delete.

[Phlok]

I'll try the duplicate container idea tomorrow. The problem with the PFlow approach is the particles need to get birthed inside PFlow and being addressed an ID. So this ID is consumed and AFAIK it will not be issued again, so my problem (all PFlow IDs consumed, so no new particles born) isn't solved.

I need to kill the volume information in FFX, before particles are created in PFlow.

[JohnnyRandom]

Are you on 32-bit version of max? These are Per Source limits:


The limit was changed with the Creativity Extension for Max 2009 (and now effective in Max 2010 and up).
For 32-bit: 50,000,000 ( 50 million )
For 64-bit: 1,000,000,000 ( 1 billion )

Thanks,
Oleg B.


You could cull via KCM too, then pipe into pflow, all without overhead of running another sim.

[Phlok]

Some questions...

I have an animated object that is rushing relatively quickly through a fume container, emitting particles. From those particles smoke and velocity is emitted into the container.

Now, I use this simulation to drive a particle system via FumeFX Birth/ Follow inside PFlow. No matter what I do (increase amount of particles, have it emit smoke to some crazy extent, use a diffusor to twirl the particles, it all looks strange, boring and none like the basic FumeFX simulation. Right as if Pflow wasn't creating any more particles based on this kind of simulation.

I already cranked up and checked the Max Particle for the PF_Sources (the reason for my question further above).

Though, it seems like if I emit directly from geometry (instead of first emitting particles from that very geometry) everything is working more the way I was expecting it to work.

Is there any known issue with that?

[alexhalstead]

Hey Flo,
I've always found using FumeFX Birth quite problematic when it comes to controlling initial velocities of my particles in PFlow and have almost always opted to emit from geometry instead. The main reason being that while your smoke or fire may have great motion, the velocity data is usually dramatically different around the source. So I usually just emit from geometry that roughly matches my fume emission and control the initial velocity with the speed operators.

Some questions...

I have an animated object that is rushing relatively quickly through a fume container, emitting particles. From those particles smoke and velocity is emitted into the container.

Now, I use this simulation to drive a particle system via FumeFX Birth/ Follow inside PFlow. No matter what I do (increase amount of particles, have it emit smoke to some crazy extent, use a diffusor to twirl the particles, it all looks strange, boring and none like the basic FumeFX simulation. Right as if Pflow wasn't creating any more particles based on this kind of simulation.

I already cranked up and checked the Max Particle for the PF_Sources (the reason for my question further above).

Though, it seems like if I emit directly from geometry (instead of first emitting particles from that very geometry) everything is working more the way I was expecting it to work.

Is there any known issue with that?

[CAICARA]

Hey there,

i have two different smoke emitter with different colors. I want them to collide and influence their movement each other.

But everytime i start them to emit, they mixing up and create a new color. I tryed with higher density, but couldn´t get satisfiying results.

Does anyone have an Idea?

[Phlok]

Thank you. This totally makes sense. Another thing I am trying to find out is how to get density thats near the kind of density coming out of FumeFX Simple source emitters.

In my scene, some geometry moves through a smoke cloud that was originally created by a FumeFX simple source emitter. The geo has a FumeFX Object Source emitter attached to it, that is really spewing out a lot of density.

Rendering FumeFX, I cannot see any difference between the stuff created by the FFX object source and ond FFX simple source.

However, after PFlow Birth/Follow-ing, there are significant differences in the density of the smokecloud that was created by FFX Simple Source and FFX object source. The particle seeding in the smoke from the object source isn't anywhere near the simple source's density.

I tried playing with the PFlow FFX Birth Smoke min/ Max parameters but thet didn't seem to have any effect at all.

[fuat]

Now, I use this simulation to drive a particle system via FumeFX Birth/ Follow inside PFlow. No matter what I do (increase amount of particles, have it emit smoke to some crazy extent, use a diffusor to twirl the particles, it all looks strange, boring and none like the basic FumeFX simulation. Right as if Pflow wasn't creating any more particles based on this kind of simulation.


did you already check out stokeMX?
i beta tested it for a few weeks, and it gives great results regarding particles driven by fumefx. it is extremely fast (much much faster than PFlow following fumefx) and you can even add different attributes, increase particle count, modify velocities with the magma modifier, etc..
hope this helps..
fuat

[JohnnyRandom]

However, after PFlow Birth/Follow-ing, there are significant differences in the density of the smokecloud that was created by FFX Simple Source and FFX object source. The particle seeding in the smoke from the object source isn't anywhere near the simple source's density.


They are totally different emitter types, of course they won't emit exactly the same. Simple Sources have a directional velocity component, Object sources have a normals velocity component. By default Simple sources emit from volume while Object Sources emit from solid. Emitting from volume creates more fluid. If you zero out all of the emitter velocities and match the fluid channels and emission type you will get as close as possible, which is pretty close, that is the best that can happen.

Notice, turn on viewport display, raise the "Reduce" threshold to ~6.o or so (scene dependent), enable the Smoke channel, turn on Show Slice + (typically) XZ, adjust position, and enable numerical values. This will tell you exactly how much smoke density you have in a voxel. These are useful tools, use them (with the exception of a few, no one else does :sad: ), that is why they are there ;) This will give you a real good idea of what is happening.

EDIT: super simple example of what I mean: