View Full Version : Maya Fluids Fire Help

03 March 2010, 09:26 PM
Hey everyone -

So I've been working on this fire sim. for the past couple days and thought I'd throw it up here to see if I could get some advice / help on getting it to look a little nicer.

Some background on my sim.:
Container Size: 20 x 8 x 2 (working in a (z)slice for now)
Res: 280 x 112 x 28
Density, Velocity, Temp., Fuel set to dynamic
High Detail Solve: off
Textures: none
Emitter Type: Surface
The only turbulence I'm using is from the emitter

One immediate issue I'm having is with the bubbling I'm getting on my sphere / surface.
Not sure what's causing this to happen and how to possibly tune it out. I'm guessing some of it may have to due with turbulence speed and or the opacity graph.

Any help would be greatly appreciated.


03 March 2010, 12:10 AM
If you want to get smoother emission from a surface, try making the emitter minDistance slightly non-zero (0.00001 for example) then adjust the max distance density+heat/sec and dropoff for best effect. The higher the max distance the further from the surface the emission happens. Additionally you can turn off jitter, although one tends to get much smoother emission just by using the max distance. The max distance is not triggered unless the min-distance is non-zero. (note that we've changed this confusing "toggle" in 2011 and added an explicit "use distance" toggle)

Also if you might try lowering the turbulence speed on the emitter if it is still too noisy.


03 March 2010, 06:43 PM
Hey Duncan!

Thanks for the insight. It really is a balance. You can alter one parameter and the whole sim will change.

I tried playing with the min / max distance and after some trial and error I didn't see a whole lot change. I did find that it's pretty sensitive and without finding the sweet spot your sim. will explode. Out of curiosity, if I change the min / max distance why would I need to possibly change the fluid drop off, density + heat/Voxel/Sec. What is that relationship?

I did find that by increasing my sampling to .1 the sim became much smoother and helped to get rid of some of the bubbling. I put up another image for comparison.

Fire comparison (

03 March 2010, 04:54 PM
...Out of curiosity, if I change the min / max distance why would I need to possibly change the fluid drop off, density + heat/Voxel/Sec. What is that relationship?

When minDistance is set to zero then surface emission simply emits into fluid voxels that intersect the surface(dropoff has minimal effect because the distance to the surface from the voxel centers is low). This can cause non-uniform distribution of fluid and staircasing. The emitter jitter helps a little with this but can make it noisy for heavy emission. When minDistance is non-zero then for each intersection voxel it emits out to the max distance but no closer than the minDistance( generally it is good to simply keep the min distance near zero unless you need a special effect ). One can think of it sort of like doing an airbrush stamp(splat) where the center of each stamp are the voxels that intersect the surface. Note that this is a lot more emission than with minDistance at zero( which toggles off the splats ), so one typically needs to lower emission when doing this. The distance of the voxels with the intersecting ones are used for the dropoff effect. Without the dropoff a large maxDistance will create a sharp outer boundary. The dropoff helps create a smoother falloff with distance from the surface. Also note that dropoff is in worldspace so scale of the fluid can dramatically change the emission( for larger fluids one should use proportionately lower dropoff ).

I suppose that instead of the splat method for min/max distance from the surface we could have computed a distance to surface grid, but this might not have been as efficient for the case where the emission distance is relatively small. (the spat method is very slow when the maxDistance is large, but normally one does not want it to be large) If we used the distance grid technique then changing the min/max distance would not affect the emission amount and computation speed as dramatically.


CGTalk Moderation
03 March 2010, 04:54 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.