Maxwell Render | Official public forum

We're working on an affordable housing project and it includes a small pool. No matter what settings we try the water looks cloudy. Does anyone know what the right settings need to be to make it look more like a swimming pool ? Screenshots and example below-


Pool photo as reference


Studio settings


WIP test render....with murky water

I usually use the "water" preset combined with a simple plane with simple bump map, that I downloaded from the MXM gallery like 10 or more years ago
I've noticed that if you want to have "cleaner" pools you must consider the surfaces below the water and not only the ones to be reflected on water's surface. usually I tend to use blue materials on those submerged surfaces (like blue or greenish mosaic), and make sure to have distinct material-id so it will be easy to edit in PS.

ps, trying to add blue shades to the water material itself usually doesn't help, only frustrates

Hi Choo Chee.

Thanks for posting. I have used realistic finishes for the pool, in this case, it's more like a white Pebletex finish than tiled. We've got about 40 renders to produce for the project marketing whilst also documenting so I'll have a closer look at your ideas if there's time available.....the week's just seem to fly by these days~!
.

Hello Mark,

In order to get a super clean and super fast-rendering water material, just create a normal transparent refractive material, assign it to the water plane object and then, select the object and, in its properties, hide it to GI (or global illumination). This way, you'll have a material that refracts the bottom of the pool, but doesn't cast shadows. However, with this trick, the water won't produce the under water caustics.
This problem happens with many ray-tracers like Maxwell. It's a probabilistic issue: it's very hard for a ray that comes from the camera to refract into the water, bounce on the bottom of the pool (once or several times), refract out of the water plane, and find a very small and very distant light, such as the sun. So, if you just use a refractive material (not hidden to GI) and the default sun radius, Maxwell may need rendering up to SL 32 to see clear caustics through the water. With the method described above, you can get a clean pool at SL 13-14, but without caustics. If you increase the sun radius to 10 (if I remember correctly), you may get a nice looking pool with caustics at SL24.
Using the "hide to GI" method, you can fake some caustics by duplicating the water plane. You'll have to move it a bit up so it is not perfectly coincident; then assing an AGS material to it and apply a mask texture that looks like a caustic pattern (such as this one, but inverted: https://as1.ftcdn.net/v2/jpg/00/18/98/8 ... Y9RvZZ.jpg) to the material layer opacity, make the object visible to GI, but hidden to camera and reflections/refractions. This way, this duplicated object will generate some soft shadows where the the texture is white, and no shadow at all where the mask is black, creating a caustic-like pattern.

I know this is a bit convoluted, but it works nicely.

I hope this helps.
Best regards,
Fernando

Hi Fernando,

Thanks for posting this - much appreciated.

That renders instantly. Problem though, the water ain't blue, which, in high sunlight, in a white fibreglass pool, with 2 metres depth, it should be.

Hi Andreas,
Thanks for the render and heads up with the procedure. We're on other parts of the building at the moment so will check this and Fernando's suggestions when we get back to the pool area.
Thanks again.

What is most peculiar is that at any depth, also varying the attenuation distance, the water doesn't look blue.

This is where we got to with the pool render which was done prior to Fernando's tip and your test image posted above. Compared to the earlier test render, this one changed the Reflectance 0 and Transmittance to a light blue colour and looks better.

Before faking caustics to speed up render time, it's about the water itself. See 2 m, 1 m and 0,5 m attenuation and the "blueness" in a matte white pool, 5 x 3 x 3 metres. Pure water has, for all intents and purposes, no colour and does not attenuate, but what's in the water, minerals, particles, etc. does. I could not find a measured attenuation distance for reasonably clean pool water. I suspect it is rather higher than 2m.

Therefore, I have the impression that the pool's GRP or ceramic tile material would have to be blue to make the water look "bluer" than it would without colouring of the pool's lining. Of course, instead of a bump texture map, one would have to use a displacement map for the water's surface agitation.

Without simulation of minerals and particles, one needs a blue-tinted pool basin. If the water is slightly agitated, the fast rendering caustics trick works well. To fine-tune, best is to ask people who do not know that they're looking at a rendering.

For those interested in the real world physics and fine detail I can see knowing this is important. Architecturally, we just want it to look good (...and be close to being as real as possible)

I would say the above "fake caustics" approach looks credible and it renders very fast.

Hi Andreas
"I would say the above "fake caustics" approach looks credible and it renders very fast." - yes, you're right there. I've had closer look at the pool image above. I first thought it was a photo but now see it's a render -you've done a great job for something you've quickly put together. Impressive!