Help with a feature water wall

[peterguthrie]

Hi Phoenixers

I've been playing around with phoenix a bit recently, trying to make something like the attached image.

I'm finding it very slow to make progress though, and was wondering if someone could give me some pointers, or even take a look at the attached max file? The foam and splash settings are pretty much on default, as they are still a mystery to me. I read in a previous forum post that increasing the water viscosity might make the water 'stick' to the wall a bit better, as without it it just seems to career off as if it is a massive waterfall. I also adjusted the gravity settings as it just didnt seem right. This is all despite my scene scale being in cm and my geometry is definitely the right size.

any help in making this look better would be very much appreciated.

Peter

[peterguthrie]

similarly with this one, I dont understand how to get the water to fall straight down the wall rather than cascade off as if it is a waterfall...

[Neilg]

I would definitely not use phoenix for this. If it was for a result I needed in animation i'd film a bunch of stuff (/buy stock footage) and make a collection of animated maps in fusion. Possibly use phoenix for the base to mix with something filmed to get a zdepth / displacement for the trough.

If it's a personal challenge then i'd have the emitter come out right at the very top of the wall pointing straight down it, not making a body of water that flows off the top as you currently have it. i'd also not really use much visible water geometry - 90% of the look & motion of this is coming from splash/foam particles.
i'd also probably end up making a ton of renders of all of phoenix's output to use as maps... p much the same workflow as if I was using footage but prerendering it all in a separate scene. in all honesty this can be rendered as a box with a good set of animated bump/normal, displace & masks for blend layers and look significantly better that trying to do it all in one go as a single pass sim.

[ynachev]

That's a pretty hard case to simulate, you will need a really small sell size, lots of steps per frame and turning on "wetting" should help the fluid stick to the surface, rather then jumping off of it like a waterfall. Also, as Neilg pointed out, most of the effect comes from the geometry of the wall in the reference picture, which brakes the water, producing lots of foam and splashes. Hope this is helpful, let us know how things progress

[George Barzinski]

Hi,
I tried to replicate the scene with the granite cube. So, I set the source to emit from Polygon ID. I set identical ID's to the top and the side polygon. Maybe this could work for you.

[peterguthrie]

thanks for the responses guys, going to put some more time into it and see where i get to

[ThomasHorsley]

Have you tried using the vertical part of the wall surface as an emitter? I would give that a go rather than trying to get all the particles to fall over the top edge.
OR try using the PHXTurbulence to fake the noise on the surface - you can quicky test this on a grid box - use the initial fill up option then try playing with the PHXTurbulence settings to disturb the liquid surface.
Once you have that working you could also add a side wind force - once your sim is complete, rotate the whole solution 90 degress and it could look like falling water!

[chadstevens]

Hi Pete, I've done a few tests and I think I have a base from which you can work.

This is a very simple render.





The main things I've changed

1. Your cell size was extremely small (188 Million cells), making testing almost impossible. I increased the cell size to 1.9cm which took the overall cell count down to roughly 4 million, I think your cell size will be fine for the final sim, but I would not use it for testing.

2.Set the emitter outgoing velocity to 200

3.Gravity back to default (1.0)

4. Steps per frame 8 the higher the more accurate and the more time it takes to simulate. This is one of the most important ones. Description from help file:

"Steps per frame | spf – One of the most important parameters of the simulator, with significant impact on quality and performance. To understand how to use it, keep in mind that the simulation is a sequential process and happens step by step. It produces good results if each simulation step introduces small changes. For example, if you have an object that is hitting the liquid surface with high speed, the result will be not good if at the first step the object is far away from the water, and at the second step the object is already deep under the water. You have to introduce intermediate steps until the changes of each step get small enough. This parameter creates these steps within each frame. A value of 1 means that there are no intermediate steps, and each step is exported into cache file. A value of 2 means that there is one intermediate step, i.e. each second step is exported to cache file, the intermediate steps are just calculated, but not exported. Signs that this parameter needs to be increased are:
Too many single liquid particles.
The liquid is torn and chaotic.
The streams have steps, or other periodical artifacts.
Keep in mind that by increasing the Steps per frame you directly decrease the performance because more steps need to be calculated between the frames. For more information, see the Steps Per Frame example below."

The help file is here:https://docs.chaosgroup.com/display/...iquid+Dynamics

5. I took the foam amount down to 2.0 just because it was taking longer to simulate.

6. Splash/mist, Splash amount = 2.0, threshold = 30, splash drag = 0.5 and mist drag = 0.25. I did this because it was interfering with the water falling, making it break up into different sections, I think wall geometry with all the facets is whats causing this.

NOTE: All of this was just a quick test, the settings are not set in stone. It's obviously not good enough for the final result, but I think with some more testing, you can get this to work much better.

I can't attach the scene for some reason, I think it's because I have restriction in this forum, If you need the file, I can wetransfer it to your email.

Hope this helps.

Cheers.

[peterguthrie]

please do! peter@peterguthrie.net

It looks very promising, will read through thoroughly and digest tomorrow