Smoke as bubbles - issues with dissipation

[Neilg]

I wanted to try something underwater with bubbles - the issue I am having is that when using smoke as the base it wants to expand and dissipate too much.
Dissipation and cooling are both on 0, vorticity is off completley, I have tried increasing the gravity while lowering the time scale to .1 (to 'suck' the bubbles up faster) and the best result I can get is still not really there.

As can be seen in the attached image - the smoke spreads out too much, becoming thinner and thinner until it stops being meshed. A long way from hitting the waters surface above us.

Any ideas for settings that will keep the smoke in clumps which mostly just travel upwards, instead of every direction all at once?

[Paul Oblomov]

U should play with isosurface value. Feels like u can find needed chunks

[Neilg]

That doesnt really help me - this is with it very low and it's already expanding a bit too much for my liking. Higher values means the mesh breaks almost immediately.
My problem is that the smoke expands and spreads out. I would like it to rise in a relatively straight line upwards and not break apart. Like a larger scale version of cigarette smoke so that it clumps.

In an ideal world I would fill my cavity with smoke and let it leak out of the hole naturally - as the effect is supposed to be an air pocket releasing. When I tested that method the smoke immediately dissipated to nothing on leaving the cavity it was trapped in.

[Paul Oblomov]

Then body force and all the fakes u like

[Svetlin.Nikolov]

Also, conservation type and strength is also really important - you better try either PCG or Smooth and raise the quality to about 50 or more.

Cheers!

[Svetlin.Nikolov]

Ah, and it might also help to switch to Forward Transfer - this will help the smoke dissipate less as well

[Neilg]

Ah, and it might also help to switch to Forward Transfer - this will help the smoke dissipate less as well

Thank you! this helped a ton.
I went back to my old idea of filling the cavity with smoke first, letting it settle, then opening the holes for it to leak out of properly and i'm getting a much better result. Got some tests to do with vorticity and it might be good!

[Neilg]

Inching closer....
https://www.youtube.com/watch?v=VXOb...ature=youtu.be

What I am finding interesting is how much the smoke expands still - when this video starts, there is no more smoke being emitted to the inside of the cavity yet it barely drains while those huge plumes pour out of it. Changing the isosurface level makes it look a bit more appropriate, but then I have the same issue where the bubbles dissipate and vanish before they get anywhere near the surface.

Is there anything else I can do to help this, or is this just a limitation of trying to use smoke with a variable density as air bubbles, something which has a mostly constant density?

[Svetlin.Nikolov]

The thing is that in the grid solvers, voxel values blue and dissipate as they travel through the grid, so the smoke that used to be all 1 when emitted from the source, will still get blurred and distributed in the voxels, eventually falling under the isosurface threshold which makes the mesh look ever smaller. The grid liquid solver before the FLIP solver used to apply a pass over the borders of the liquid which sharpened them and made voxels be either full of liquid, or empty, without the smoothed values that appear inbetween. Simulating bubbles with smoke is not really close to how bubbles work, so it's just a matter of getting it to look okay..

Will there be a difference if you emit smoke 10.0 from the source, instead of 1.0?

[Neilg]

I thought that might be the case!
I think i have enough to go on here and keep pushing it to look better, I was just wondering if there was a setting somewhere which I wasnt considering. I'll share when I have more!

[Svetlin.Nikolov]

Roger that, hope it works out Cheers!

[Neilg]

One last question - more of a feature request. Is it possible to get an 'initial fill up' but with smoke?
I cant be the only one that would find it useful. I guess it would also need a density value too to set how much smoke is in it?

[Svetlin.Nikolov]

Sure, yes - it would be a shortcut instead of using an animated Volume Brush source and would require the smoke amount too.

[squintnic]

Came here to second the smoke initial fill feature request