Aesthetic issues with PyroFX smoke

[Lundgrenuk]

I am new to Houdini and trying to finish my graduate showreel, however I am having some issues with the look of pyro FX smoke tools.

 

The main issue I have is that when the smoke emitter moves too quickly (which is not very quick at all) then I get these clouds of smoke rather than a continuous plume of smoke following the emitter.

 

Secondly, the smoke does not seem to continue rising a dispersing like smoke normally would. Instead it just seems to hover where it got emitted.

 

If anyone on here could give me some advice so that I can finish my showreel and begin the daunting task of trying to find a job as a junior 3D generalist that would be greatly appreciated.

 

Edited April 12, 2016 by Lundgrenuk

[3iart]

Your abc. file wasn't included  so it's hard to know what exactly you are actually talking about.

 

Probably you want to familiarize yourself with adding details into a sim.

 

http://forums.odforce.net/topic/18397-get-rid-of-mushroom-effect-in-explosion-and-add-details-in-the-opening-frames

 

I'll summarize the most important part.

 

Fluids, as we expect to see them in the real world, do not exist in a vacuum.

[Lundgrenuk]

Thank you for the quick reply.

 

Unfortunately, even when I zip my .abc file it is still too large to upload here (826 MB). I guess that because the emitter is parented to a piece of geometry imported from the alembic file, the emittor is not moving at all when you open the hip file. Which makes sense now that I actually think about it.

 

If I understand the gist of the what the people are talking about in the link you sent me then adding some turbulent force in the shape of some slight wind and increasing the maximum substeps should go a long way to potentially resolving some of my issues.

[fathom]

when you get discrete blobs like that on a moving emitter, you should blur your emitter based on the speed it's moving.  depending on what you're trying to do, you might also include some of the velocity of your emitter in your sim as well.

 

typically i'll create an emitter based on points that are rasterized into a volume.  the points need to have a reasonable velocity so either they're trailed or they're simmed or they're scattered on something that has velocities.  then you add a randomized fraction portion of a frame's worth of velocity into each point (using a wrangle: "v@P += float(rand(i@ptnum/4.5))*f@TimeInc*v@v;")  then you rasterize into a debsity volume using a vdb from points.  add an additional volume called "vel" that comes from the "point.v" attribute (in the vdb from points sop).

 

then use that as your source and add both density and velocity...

[Ratman]

Can't open your file right now, but I just ran into a similar issue and the main things that helped were.

 

1. Your emitter MUST have substep frames, so if you're writing out bgeo files you need will need to use a TimeBlend reading back the points or write out bgeos with subteps. The catch is you can't timeblend a Fluid Source volume, so you'll have to generate it on the fly reading into your pyro sim.

 

2. On your pyro solver, in the advance tab you'll have to raise your Max Substeps to at least 2, and lower the CFL Condition, starting with 0.5. You may have to also raise the Min Substeps as well to 2, but experiment with those and it should help fix the steps issue.

 

Hope it helps!

-Rick

[Yon]

Try scaling your sim via source volume tab. Like 4,4,4 for the scale volume,temp and velocity parms. Your sim looks really thin and if your sim is large enough it will fill in the gaps for you without substeps. Fuel inefficiency will leave a trail of fuel that will constantly emit. Try setting this to 0 and obviously a higher dissapation.

More importantly, Smoke II by Peter Quint on youtube/vimeo covers this situation, I recommend you reference that.