I’ve searched the forum for similar and only found 1 thread from years ago and I was the last person to respond to it. But I need to make the most accurate infinity pool possible for (relatively) close-up animations.
You know the type - where the water from the top pool rolls over the edge creating a waterfall effect and into a pool below.
These videos show the exact effect I’m trying to achieve, not the best examples, but the best I could find:
The second video isn’t the exact effect as the water is cascading down a more irregular wall and the top edge overhangs said wall, but it does show very clearly the variation of the water rolling over the top edge of the pool, and it shows the random small undulations and splashes you get in the cascading water down the wall and then the smaller splashes into the pool below.
What would be the most accurate way to recreate this with Phoenix.
I’ve tried methods using a single piece of geometry and the OceanTex in the displacement but it’s hard to control whilst avoiding clashes with the underlying geometry and not very realistic.
You could try a setup with a cascade connection, where the fluid flows from one Simulator into another. The option you will need for this setup to work is located in the Grid rollout → Cascade Simulator. This option specifies the source Liquid Simulator to connect the selected Simulator to, forming a cascading simulation. This allows you to join several Simulators into a structure with a complex shape. Here you could find more information about the Cascade Simulator option - Chaos Docs nSimulators-CascadeConnectionTransferringfluidbetweenSimulator susingaCascadeConnection and Chaos Docs
We also have a Template Scene - Liquid Cascade Simulation, that might be useful - Phoenix for 3ds Max
Also, both Simulators can have the Initial Liquid Fill Up option enabled, so that you will have liquid in the pool from the start of the simulation. And then you can have an emitter in the first Simulator to generate more liquid, splash and foam particles that will cascade to the second Simulator.
In addition, we have a Pool Example Scene - Phoenix for 3ds Max, that might be useful as a starting point. In this scene a Pure Ocean Mode and a Phoenix Ocean Texture are used to displace the pool surface. In your case you could try with a Wave Force with loaded Phoenix Ocean Texture in the Ocean Map slot. More information about the Wave Force - Phoenix for 3ds Max
Another perfect example of what I want to achieve, later in the animation there will be objects interacting with the main pool causing slightly larger ‘waves’ to flow over the edge INFINITY POOL
Ok so I’m getting somewhere, the problem i was having in the screenshots above was ‘fill up for ocean’ being checked. I’ve unchecked that and now the water does move across to the lower pool as expected.
My next problem is there is too much starting motion and i can’t seem to make the effect subtle enough.
When I first start simulating with initial fill, the liquid appears to have some initial motion as if it’s been dropped into the pool geometry, this creates a ‘bulge’ or large wave like effect that I don’t really want.
As you can see in all of the video references I’ve sent. The effect is really subtle. The water fill line is already lined up with the infinity edge and then it almost trickles down the wall. At the minute I’m getting more of a heavy gale effect.
The pools in the reference videos are almost flat, and it’s the gentle breeze and possibly the movement from any filters that cause the water to have a constant but subtle overflow effect.
Hi Dave! I’m also at the moment starting a job where I’m going to need this solution and I’ve come to anticipate looking at the forum if there’s anything. Luckily I found your topic, I hope it’s working out there, if you can share the final result here with us!
In order for the liquid to settle you have a few options - you can increase the scene scale in the Grid rollout - if it’s too low the liquid will splash like in a glass, though if you raise it too much it will behave like an ocean, so you have to see what works for your scene.
The other option is to increase the steps per frame - this will increase simulation time so be careful when increasing them.
Finally you can do a pre-roll, just sim a few frames at the start and let the liquid settle.
You can combine all of the three things I listed and you should get a good result.
I was interested in this topic, so I just made a scene.
It behaves generally as it should at a lower resolution, though as soon as I increase resolution
then the sim just behaves insanely, increasing the initial 30mil particles to half a billion within a couple of frames, due
to what looks to be just incorrect interactions/collision speeds, at least when you look at the very low emission speeds.
I don’t see any obvious way to anticipate this, or any method by which to troubleshoot to adjust in any meaningfully interactive way.
Probably there is a simple answer that hopefully doesn’t include ‘it’s trial and error’
Maybe you could investigate and suggest a solution which could help me and maybe the OP too
Here is the scene file so you will see what I mean. I left it set to the resolution where it goes mad, so just lower it once or twice to get what I originally set up. INFINITY POOL.rar (140 KB)
Thanks for the scene. I’ve made some updates to the setup and this is the result I got at frame 122.
What I did was to change the scene scale of the Simulator to 2 so that the liquid moves a bit slower.
Removed the viscosity, surface tension and the Sticky liquid - in this setup I think you can get away without any of these.
I also disabled the Initial liquid fill for the simulator and made a duplicate of the bottom face of the pool and extruded it. Then from the Phoenix properties of the object I enabled the Initial Liquid Fill option.
This way you get a precise control of how much of the pool will be filled without having to rely on the simulator size.
Then I scaled and repositioned a bit the box used for the underwater emission and also made it non-solid from the Phoenix properties, this way it won’t create any strange velocities and won’t interact with the simulaiton.
Here is the modified scene as well. Hope this helps
Thanks for taking a look.
Unfortunately this isn’t the correct look.
As the OP noted, the crucial thing is to have the water not cascade outwards from the lip of the pool but rather drip down vertically.
This was my reason for enabling sticky liquid, which does work up to a point, after which it collects at the top and starts again to cascade at an angle, as in the screenshot attached.
I’m guessing you didn’t try to raise the resolution, which was what I tried to do to increase detail and hopefuly get a smoother result.
I also lowered the emitter speed to try to avoid velocity issues.
However, as you can see in the second screenshot, at that frame the madness starts to occur, with super-fast particles shooting out
and raising the count from around 40mil to 260+, which continues to rise exponentially and needs to be stopped.
So my point would be that I thought if a low res sim worked generally as expected, then a higher res version should not start behaving in a completely different way.
Wouldn’t that mean that there is no point in doing lower res tests?
Generally, do you have a high degree of certainty that the specific referenced effect can be achieved, or is it just too specific a case for the solver to handle in one go?
That is, using a separate top surface and with additional emitters to represent the vertically dripping parts.
Thanks all. I’m currently in the process of buying a new PC that will be dedicated to real time applications and simulations. My current setup means Phoenix is really slow so trial and error becomes quite painful and laborious.
Thanks for all of the replies. I will test them in due course. This has actually given me a couple of ideas. Being the novice that I am, I hadn’t thought of using additional emitters to add more liquid which then forces it to overflow. I was trying to start with an already full pool and complete everything from a single initial fill object.
Agreed, the example scene provided is too out of control. Almost like a water pipe has burst under the pool
The desired effect is a lot more subtle and controlled. The water should simply roll over and down the pool edge quite gently. And only more aggressively if someone or something jumps into the pool.
I’ve attached the actual pool geometry if anyone fancies trying this properly while I wait for the new PC to arrive
And to reiterate, this is exactly the type of control and subtleness that I’m looking for INFINITY POOL and ideally in a continuous looping animation although that last bit isn’t essential POOL GEO.zip (76.1 KB)
I gave up on this I’m afraid. I’m not convinced it’s possible to do this in one sim at the necessary resolution to make it realistic, for all sorts of reasons.
Even spliting it up is causing similar issues, so it bored me in the end, as there’s no obvious route to go down to be able to iterate intuitively.
I think there’s a reason there are no existing examples of this scenario online…it is just hard.
Each idea I have unfortunately leads me nowhere, so I’ll hope that Georgi can come up with a solution that fits the problem.
Meanwhile I reverted to using the oceantex and think that’s probably more than realistic enough and without all the hassle of a sim
