Underwater Volcano

[ben_hamburg]

Hey guys,

I?m doing R&D for a couple of underwater shots of volcanoes and volcanic vents.

I?m going to start by looking at your volcano demo scene, but where should I take it from there?

1. Can I somehow increase air pressure to simulate the underwater pressure? Or would I simply increase gravity?
2. Should I increase the cooling factor? (The idea being that water is colder than air....)

I?m also thinking about how to add some lava. I?m not sure there will be any eruption yet, maybe I?m just gonna add some slow moving lava on the sides of the volcano.

I can do the high viscosity liquid, but I have no idea how to shade it properly, all I found was this video:
https://vimeo.com/101442124

Any ideas? Can I us RGB info from Particle velocity or something like this to create a blend material?
How would I add some displacement (lsome simple turbulent noise map), or rather can I use a proxy object for mapping or how would I do this?
It just needs to look good, it doesn?t have to be super accurate...

[georgi.zhekov]

Hello,

For the pressure you can use Max's Drag Force and that should do the job. You can play around with cooling as well and see if the result is what you're after.

For the shader - you can make a VRayLight Material and plug a VRayDirt map in the Color Slot. Then you can set the unoccluded color to black and drive the color with the occluded one.

For the displacement you can try with a noise map - just like you would add it for any other type of mesh.

[ben_hamburg]

Ok, thanks, I?ll try these tips.

I?ve got one more question though about resimulation:

I looked at the volcano demo scene and after simulating the base sim, I tried resimulation for the first time.

I?m a bit confused though: While the base simulation was pretty fast, resimulation took considerably longer. First I tried with amping the resolution by 3, but my sim quickly crashed, so I went back to 1.
But still: In the manual it says that a value of 1 should double the grid cells, but it seemed more than to be multiplied by a multitude.I?m not sure about the starting count, but I ended up with up to 200 mill cells.
Is the resimulation taking into account the final grid size (adaptive grid is enabled) and thats why it gets so high?
Because I?m trying to use the same setup for a different simulation and I?m not sure how many cells my bassim should have.

Another question about the underwater dynamics:

Besides cooling, I should probably also try and increase the smoke dissipation, right? This might also help with the simulation times, since the grid wouldn?t need to expand as far.

And one more thing:

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

How could I achieve that sucking effect? Would that also be done with drag or should I animate the emitter source to alternate between negative and positive values at the eruption phases?

Sorry bout the Question marks, somehow formatting is off after submitting a post...

[Svetlin.Nikolov]

Hey, will change the info on amp resolution in the docs - a value of 1 means doubling the resolution along each axis, which means 8 times more voxels. A value of 2 means tripling the res, which means 27 times more voxels, etc.

Maybe you can use dissipation, but I usually advise against using this for anything else but steam. You should better just use adaptive limits to cap the expansion in direction you don't care about.

For the sucking effect, I'd suggest you boost the conservation quality and definitely use PCG. With a strong discharge you should get that effect naturally. Drag will probably not help in achieving that - it could be a decoration, but not the main factor. Animation of the emitter should definitely help - otherwise fluids will naturally start to pulsate over time, but you could help that effect. In general, even apart from that specific case, it's always good to have some pulsation on the discharge strength as it helps get a more diverse result

Cheers!

[ben_hamburg]

Hey Svetlin, thanks for clarifying the amped resolution thing...I got a bit scared by the amount of cells....

But does resimulation with wavelets actually do much without also amping the resolution? I just did a short test and I couldn?t make out much of a difference, maybe if I had let it run longer.

For the bigger volcano simulation I am just going to use the expansion cap, since the top will be out of frame anyways. But for the smaller volcanic vents dissipation actually looks good, probably because its kind of like steam.... The default conservation quality in the volcano test scene is already at 100, is that anough or how much higher should I go? I?m gonna do some more tests with spikes in the discharge. I actually watched some more reference videos and the underwater volcano doesn?t really look that different from the normal ones...Adding some ash particles that start to float around will probably help to sell the effect best, together with smart shading of the smoke (looks much whiter underwater than in the air.).

[Svetlin.Nikolov]

Ah, yes - with amp=1 the wavelet won't work and it will have a visible effect getting close to 2. You can also boost the wavelet strength, but it's not hard to overcook it, so watch out After all it's just a fake noise that won't look organic with the rest of the fluid if it's made too visible.

If the conservation quality is already at 100, you need to just boost the discharge quite a lot - in the volcano sample, it's quite low, so that it would look slow and epic.

[ben_hamburg]

Thanks for clearing that up!

I still have quite some way to go, since I don?t have much time right now I tend to just some very quick "shoot and miss" over night sims and thus I?m progressing so slow.

It took me a while to get the lave to look right, but now that sim is on the right way, I was trying to use the actual lava as a source for the discharge of the fire/smoke sim.

Couple of issues I had there (starting of with the original preset):

1. The discharge seemed to be way too high and my sims looked like I just had fire and almost no smoke, because the fire was so intense. I played around quite extensively with the volumetric options, but I couldn?t get the fire to be les prominent, without it starting to look a bit weird (turned down opacity a lot etc.).
You said to get that "suction" effect from that youtube reference video I had to boost the discharge and the effect should occur naturally. But since its under water I?m thinking that maybe I just need to boost the cooling a lot in the sim to not have the fire be so prominent, but still get enough power and smoke. What do you think?

2. The second problem was that the discharge was too even: The lava in the middle should discharge the most, while the slow moving lava from the overflow, that is starting to harden, should of course discharge less, instead I got an even discharge everywhere.
I?m gonna try the grid tex again for that or meaby just use a Vraydistancetex to get the same effect (I used that for shading the lava as well).

3. The last problem was mostly speed of the setup: Animating the discharge to go in sync with the liquid eruption spikes was quite time consuming, since viewport was so slow.
Maybe I can fix this with the velocity based grid texture. But how would that actually work? Would I just plug in the velocity channel based grid texture in the discharge modifier map slot? And then set the discharge at the highes level and then the white values in the map would mean those high values, while the gray/black values in the map would mean the discharge would be modified to be lower there?

I?m als going to try and lower the PCG quality significantly, since I can?t really preview the effect of the discharge map on a low cell count sim...

[ben_hamburg]

Hey,

So I played around with the grid texture and I couldn?t get the Velocity texture to work. Well, I didnt find a way to remap the colors to be a representative black and white output.
I did however get the speed texture to work, I had to play around quite a bit with a color correction map, but know I have a nice black and white texture map to work with.

it already works pretty well for shading the lava, I?m not quite sure yet, if i can use it for the discharge.

I tried plugging it into the discharge map slot of the source of the fire sim (which is using the Liquid sim as source), but I can?t quite figure out how this works:

1. How do the black and white values correspond to the discharge amount? Is white and black some fixed values? Or do they just modifiy the discharge value I put in there?
And if so - how exactly are they modified? Lets say I have a value of 1000 I want as maximum discharge - completely white should be 1000 and black should be zero - and grey values are percentages of the maximum value?
Can I still animate the discharge value and the black and white values of the map change accordingly?

2. What exactly is the "Mod" slot right beside the Map slot for? Deos it have anything to do with the map?

3. Can I just plug in the grid texture including color correction as I described here, or do I first need to export it or something like that (I think thats only the case for the particle texture, right?)

[Svetlin.Nikolov]

Hey,

I think you found the answers to most questions, so just a couple of questions from me: Why is the viewport slow? Does increasing the reduction or enabling auto-reduction improve it? Also, if you have Show Mesh enabled, the reduction will not work on the mesh, so you might need to turn it off and stick to voxel and particle preview instead.

You should indeed use the grid tex as Speed, not as Velocity - we already mentioned that. Speed is the magnitude, while Velocity is a 3D vector with positive and negative values in the color components, so unless you need direction, you better use Speed instead

Any texture you plug in as a mask for general discharge, or smoke, temperature, etc, will work as a multiply mask - just the way you described it. Did you check the docs on the discharge modifiers? I will add a viewport preview of the effect of those so it would be easier to manage, but I need to find a little free time first.

You can just plug the grid texture including any maps following it, yes

Cheers!

[ben_hamburg]

Hey,

the viewport is probably just slow since the cache files have to be loaded from the network, turning on autoreduction doesn?t seem to help a lot in this case. I was curious why the viewport was slow even if I hid the simulator, I?m guessing its because I?m using the grid tex in the material editor (which notoriously slows things down in Max when using animated maps).

But seems I don?t really ned it right now, I?ve kind of abandoned the idea of the liquid lava as a source, since I will probably just use fire/smoke. After doing some test renders I noticed, that the lava was mostly hidden behin the smoke, which needs to be pretty dense to get that massive look. I also looked at the reference video again and you can barely see the lava, its really mostly the smoke and some hot debris floating around. I finally got that "suction" Effect going aswell, I?m using a couple of smiple spheres now as emitters with quick discharge bursts from +5000 and immediately after that -2000 or so. Looks pretty close to the reference video now.

Thanks with helping to get this off the ground (or rather: under water)!

[ben_hamburg]

Oh, just one more question about resimulation:

I?ve read somewhere that I could lower conservation quality for the resimulation, but I think that was for 2.2, right?

[Svetlin.Nikolov]

Yay, I hope you'll be able to share the animation soon I think I remember that conservation trick, but it won't matter to the resim speed anymore - conservation and all velocity-generating stages are getting skipped during resim now. Loading caches from the network is a horror though...

[ben_hamburg]

Yeah, since I will be doing a ton of sims for my short I?m still trying to figure out the best workflow in terms of speed. I?ll try simulating on my local workstation next time when doing R&D to see if the difference is worth copying cache files to the network only before rendering...
And I?ll also go back to lower grid size for testing purposes...I got fooled by the automatic preview quality reduction and thought the grid was already so coarse, so I didn?t want to lower it even further...

I?m still having some issues with the form of the smoke, I just did some resim overnight and I?m not sure if its because of the resim or because of the volemetric options, but the render looks different than the tests I did with the democ scene, sort of like if a noise texture was applied to the smoke or something like that.
I?ve only done some quick single frame renderings, maybe it looks fin in motion, I?ll get back here with the next update!

I`m guessing underwater volcanos are not a very common thing to simulate, but maybe someone in the future will have to do this again...

[ben_hamburg]

Ok, here is the latest Version:

https://www.youtube.com/watch?v=DZjh...ature=youtu.be

Dynamics worked out pretty nice with the negative discharge. The rolling off the smoke is still looking different from what I?m used to, but again, after watching the reference video it seems like this might just be what happens with some negative dicharge following the positive discharge.
I?ve kept the liquid lave, even though mostly only the part is visible, that is flowing out of the hole.

What I?m still gonna change is the amount of explosions: There are too many and they follow each other too rapidly. That might also be the reason why it looks a bit noisy, because there is just too much going on. I?d rather have fewer of them but make them a bit stronger (the more smoke there is the less they for those nice bubbles).

Then all I gotta do is add some debris/ash particles and of course switch out the volcano for my own and I?ll be very happy...