Strong Surface mode

[magilla]

Strong surface mode causes the fluid to become sludgy and lack detail but without it turned on the surface is very blobby and unnatural. What can I do to reach a happy medium between the two?

The best result I have had so far is turning on strong surface mode after 100 frames of pre-roll but eventually the sim becomes too flat and boring.

[ttdww]

what resolution is your grid currently? the reason i ask is because if something closer to the surface tension of water is the goal, i've found that lower res grids will usually not yield satisfactory results with strong surface mode on, and then there's the tendancy to turn it off which can improve behavior in some ways but with the possibility of artifacts. with a higher res grid, s.s.m. seems to look a lot more natural (again for water sims at least - i can't speak to other liquid types). you could also try multi-pass simming. i've seen good results with it. perhaps mess with system scale. this is with limited testing. the senior users and dev team may have a lot more input.

[magilla]

5cm cell in a 4000x11000x220 grid - not high res but not low either. It should be high enough to do the job but I have had issues with increasing the resolution that causes the water to just disappear before it flows all the way through the grid, so I'm stuck at this resolution.

[ttdww]

yep that doesn't qualify as a low res grid regarding my comment about resolution. seems good enough.

[magilla]

yeah at 10cm though the fluid acted more like a blanket sliding over the surface.

[ttdww]

i guess there could be a number of things to try, however there are a lot of unknowns. if it's a river you're creating with rocks deflecting the fluid mesh, perhaps they are too smooth and the water flow not fast enough to cause the mesh to breakup - this is where multipass has helped me previously. perhaps an image sample and/or more info would help isolate the issue.

[magilla]

A very simple setup, the rocks are smooth but the river bed should throw enough turbulence into the mix. It's not doing too badly at the moment but I had to start with strong surface turned off.

[ttdww]

very cool. i don't use ssm if i don't have to. question - what's your total grid size (x*y*z) barring any units division. i may have misinterpreted what you posted.

[magilla]

just looks like marbles rolling down hill without the ssm

800x220x100 total 17,600,000 cells

[ttdww]

pretty good res for what you're trying to do - may be a bit low, but you mentioned vanishing water at a higher res. i may be at a loss at this point, short of me reviewing your scene params. certainly sounds like it should not have that behavior with ssm off. things to check: 3dsmax system scale (or scene scale in the dynamics rollup if you are on nightlies), conservation method and quality, uniform density on/off, advection type, liquid sharpness, viscosity, and surface tension. if those are all good, perhaps a senior member can be of some assistance.

[magilla]

how is 'good' determined for all those settings? I'm very new to Phoenix and the documentation is scarce and mostly out of date.

[ttdww]

click on the 'good' button next to each ;]

from my testing (so far), the following may give good results unless a senior member states otherwise. i've stolen these from the supplied sample scenes so they are close to a default setup i think. i've tried alternate params but found they seem to work good in most cases for standard water:
--conservation method and quality: symmetric, 30
--uniform density: off
--advection type: classic or multi-pass ('slow moving' sometimes)
--liquid sharpness: 1.0
--viscosity: 0
--surface tension: 0
--3dsmax system scale (or scene scale in the dynamics rollup if you are on nightlies): do a set of wedge tests

if you find one or more of the above params wasn't previously set to 'good', definitely do a very small localized sim around 1 or 2 rocks and lower the grid spacing to see if you get better results.

[ttdww]

also, if you're simming your splash/foam with the water mesh, also confirm you're not evaporating the water away with the 'affect grid' param under the splash rollup. safe bet is to keep it at '0' unless you've got a long waterfall or something to that effect.

[magilla]

cool, thanks heaps for that. Most of the settings are the same except advection was on forward transfer (I have no idea why, I think I was cycling through tests and just settled on that one).

I am emitting drag particles which I will frost later to create some gunk in the flow but I was going to try the resim to generate the foam since foam seemed to slow things down a lot. No point simulating foam on a crappy fluid sim.

The scale has me baffled. I am not using the nightlies and my scene scale is metric:centimetres system 1 unit = 1 metre

[ttdww]

sure thing. yea i know how that is.

good point.