Translucent plastic silicone tubing material, what is a good approach?

[SharpEars]

I am referring to the type of silicone tubing showin in the attached pictures (soft milky tubing and other soft white or off-white shapes made out of milky translucent silicone).

Here is a good video showing it in use, which is helpful to see its translucent/transparent characteristics when fitted on to objects (especially starting at around 1:15 into the video): https://www.youtube.com/watch?v=Jq55YGcunrg

I am not sure if this is a good case for SSS, or if it requires full transparency and a Volume material. If anyone has created anything like this and/or knows giw a material that has these characteristics can be created, I'd appreciate the details.

[burnin]

Alike "Translucent Milky Plastic"?

[SharpEars]

Tried what was suggested there. Yes it looks like translucent hard white plastic, but I was looking for something that looks like translucent milky silicone (i.e., soft looking, but not sure what material characteristics are needed to give this impression).

It is the difference between the below two pics (hard plastic on the top (first example), soft silicone tubing below it

The bottom silicone example is in a light blue color to help separate color from other optical features of the material. It clearly shows that color saturation plays some sort of role in making it look "soft and silicony" and not like hard translucent plastic.

Compare this with the closest (in color and translucency hard plastic item I could find at the ) second from bottom sample.

Also a rolled silicone sheet is at the very bottom for a good reference of depth, layering, shadows, and highlights. Plus, something about it makes it look unmistakingly soft, and it's not just its geometry (i.e., the fact that it is rolled up and imperfect). I think it's how the saturation of the material works into the translucency, but trying to simulate that with the Scattering Directionality and Color of a Volume material I have found nearly impossible.

[SharpEars]

My best attempt so far is attached (4k/very long render time with caustics on - noise level 2.28%). Click the image and then the magnifying glass to see the image in full 4k resolution, if you have a 4k monitor.

Width of scene is 41 cm, as a reference (for the material). Camera focal length is 75 mm, to minimize geometric distortion.
Lighting: Large disc light from top (somewhat weak and distant, see faint shadow cast by the coil around the bottom of the metal pole in the center). Key light from front right, casting clearly visible shadows.

All of the translucent white plastic is just one material type, including the plastic rings around the metal pole's bottom, broken apart individually to the left of the pole to illustrate thickness. The material I came up with just looks too hard and plasticy.

[burnin]

Silicone: IOR ~1.43, Abbe ~ 44.44.
Also, use anisotropy (consider manufacturing process).

[SharpEars]

Silicone: IOR ~1.43, Abbe ~ 44.44.
Also, use anisotropy (consider manufacturing process).

OK, will add that in to see what affect it has and repost.

[pokoy]

I had a go at this... my thoughts were:
- it's clearly transparent/refractive so you need to with the refraction route with absorption and scattering (vs fraction)
- there's some rayleigh scattering going on (at least that's what it look slike) where certain wavelengths within the media get scattered more, hence the yellow/blue coloring. For this, absorption color needs to be slightly yellow (if you want this to be more or less pronounced, increase or decrease saturation in absorption color to control this effect)
- the internal scattering is pretty visible so scattering color has to be bright (the brighter the longer it renders, unfortunately)

Attached is a scene for max 2017 and 2020, give it a try. There's also another material done with the fraction route (it can't do refraction but it turns out pretty milky, you can also mix refraction/fraction in one shader if you want).

Play with:
- glossiness
- absorption distance
- scattering color (this one controls the scattering within the material, also, if you want the material to have a tone instead of the yellow/blue absorption, use this one, set absorption color to have no saturation then)

I also did a 10 minutes render with caustics. While it obviously took longer and is still quite noisy, the material looked much better since internal reflection were much more realistic, particularly visible on the tubes.

I had to remove some of the objects and HDR used in my renders, but the scene includes some lighting and the rolls, sheet and tubes.

One thing - the material is really sensitive in terms of lighting so keep in mind that any test you do may change considerably with different lighting ;)

[romullus]

This looks very nice and convincing material. Could you please attach screenshot with material settings for those with older 3ds max.

[pokoy]

Sorry, not in the office right now, I'll post a 2016 scene later tonight. Or do you need 2014?

[romullus]

Thanks, 2016 would do fine for me, but if you have such option, then 2014 would be even better, since i know there's few people still using it.

[PROH]

Looks very convincing :) Since OP uses C4D, I think that a screenshot of the material settings would be great also.

Regards

[pokoy]

Somehow missed that OP uses C4D. In that case I'll post a screenshot with color values... and 2014/2016 versions of the max file.

One thing I forgot to point out is that in Corona, surface and shading normals render differently with any sort of sss/translucency/absorption, so that you'll have to keep your geometry quite highres to avoid seeing faceted 'artefacts'. It's not visible in my render above, it's too low res and too noisy, but it can become visible in a cleaner highres...

[pokoy]

I might not be able to post anything today anymore, I'm afraid it's going to be tomorrow... Sorry.

[SharpEars]

I might not be able to post anything today anymore, I'm afraid it's going to be tomorrow... Sorry.

Thank you very very much for your post. Although the MAX scene file you provided doesn't have lighting that matches the image sample and is missing half the geometry (the interesting ones, too!), I was able to tweak it to get something comparable.

I then made a copy of all geometry and threw the best material I had come up with as of right now on the second set, to do a side by side comparison (for now with caustics enabled, for the utmost realism).

It's being rendered right now at 2160x1440px on my 💥BEAST💥 of a workstation (Crazy overclocked/water cooled 18/36 core 4.9 GHz constant speed Skylake X (i9-7980ex) with -1 AVX2 and 128 GB of RAM). and I will post the image (and the settings for both our materials) when it's done, so other can make use of them.

It's moving pretty slow due to caustics, I took the liberty of increasing the max trace level to 50, just in case, oh and of course I removed the ten minute time limit and set it to zero (= test of my patience). I'll wait for it to drop into the single digits before I kill it and give my PC a break (one of the cores is sitting at 90 degrees C, right now).

A wonderful scene file for testing things out.

I just wish all of the objects in the sample images were actually present in the scene. The following objects are missing:

- The large stroller cover looking thing at the top
- All three of the trinkets in the middle that are hard to describe as to what they actually are
- The antenna looking object protruding from the coil of hose (top left)

If you have a scene file with everything, and better yet with lighting that matches the sample images, I'd appreciate it and will re-render both your material and mine.

Updates, below:

- 10 Passes, 17 mins, 13.23% noise


Just realized that you had DOF enabled, now disabled (or my material set will be blurred and I don't want to play with the aperture/distance settings), and restarting the render...

Dropping to a region render of just a pair of coils, with the two different materials, but at a larger size. It's taking forever to clear with caustics turned on...

[pokoy]

Hey! I mentioned in my first post that I removed some stuff for sharing. The geometry you are missing is from a model on GrabCAD, I can post a link tomorrow. Obviously, I can't redistribute the data, but you can download yourself, and there are many other well suited models over there.

The HDR used for lighting my test scene came from hdrihaven, these again are public domain but I decided to remove it nonetheless... will post a link, too.

Glad you found it useful!! It's really very sensitive to lighting but I think the general parameters are acceptable. Then again, lots are left to play with and I had a hard time deciding which exact settings to go with. But that's the beauty of the scattering feature.
The only thing that's odd is how much scattering differs with caustics with exactly the same settings. I'm not sure what's going on there.

Your machine sounds about 10 times faster than mine :D

[SharpEars]

Hey! I mentioned in my first post that I removed some stuff for sharing. The geometry you are missing is from a model on GrabCAD, I can post a link tomorrow. Obviously, I can't redistribute the data, but you can download yourself, and there are many other well suited models over there.

The HDR used for lighting my test scene came from hdrihaven, these again are public domain but I decided to remove it nonetheless... will post a link, too.

Glad you found it useful!! It's really very sensitive to lighting but I think the general parameters are acceptable. Then again, lots are left to play with and I had a hard time deciding which exact settings to go with. But that's the beauty of the scattering feature.
The only thing that's odd is how much scattering differs with caustics with exactly the same settings. I'm not sure what's going on there.

Your machine sounds about 10 times faster than mine :D

Yeah, I noticed that as well. The caustic completely change the material's appearance. I think I almost prefer the version without them, but since caustics are supposed to add realism, I am trying to work them in to my material.

That said, our materials are somewhat different, although based on the same principles. I suppose you can judge for yourself once I get this render done.

[pokoy]

So here's the updated zip containing Max files for 2020, 2018, 2016 and 2014 with Corona v5, plus a screenshot of the material setup (all colors black/white except for absorption and scattering).

The HDR used for lighting the example was this one:
https://hdrihaven.com/hdri/?c=indoor&h=aft_lounge

The missing geometry is something I can't trace back anymore so I can't provide a source, unfortunately. There are tons of good CAD models suited perfectly for this over at grabCAD.com

[SharpEars]

So looks like I won't be posting comparison images of our materials, because at least with Corona 6/3ds max, the caustics appear to be "broken." The image can never clear, I tried doing region renders on small regions to be sure.

The photons being shot into the scene for caustics are large, there aren't enough of them, and there is no or not enough interpolation between them, causing large bright spots (and I don't mean fireflies, those could be removed with denoising) to be present in the scene and never clear. The denoisers are useless for this, because they only clear small "pixel size" noise, not large bright caustic photon splotches (see attached image for a better description of this effect in general, this is not what I am seeing specifically, just an example that illustrates what bad photon handling looks like, although in the sample image shown it was done intentionally).

I tried toggling adaptive caustics mode, even tried playing with the developer/experimental caustics settings, including maxing out the number of photons (per pass?) to 99,000 and playing with _many_ permutations of the other settings - well as many as time would allow, since rendering with caustics on is as slow as watching Pluto rotate around the sun, and I did want results some time this century, especially given that the hardware I am using, although not a 64/128 core threadripper, is far faster than what 99.99% of people are using (Skylake X 18/36 cores at a fixed 4.9 GHz with -1 AVX2 (also fixed) and custom loop direct [cold plate to actual silicon die - no IHS] water cooling to keep temps ≪90 ℃ while using AVX2 full blast, like Corona does). Nothing helped, the caustics mode is completely useless for this scene and Corona does not allow enough control over photon settings to fix it, not to mention that having to twiddle with these is reminiscent of trying to get caustics to work in VRay/RedShift/Corona Physical, etc..., with the million knobs you have to turn to get caustics right (or to appear at all in the scene!).

We love Corona for its simplicity when it comes to render settings - its greatest selling point! But, as this scene clearly shows that the caustics simply do not work for a scene like this. I don't know if this is a bug, a deficiency in the specific algorithm that the Corona team chose to use for caustics, or what.

If someone from the Corona team wants the 3ds Max Sample scene that demos caustics completely failing due to very poor caustic photon handling/calculations, I can post it as an attachment. It is just a variant of the scene already posted in this thread by pokoy with some tweaks.

Or is this a known issue with caustics - that they just don't work for some scenes with partially transparent objects? I mean the scene is not overly complex, by any means and the lighting is trivial, with two area lights.

[pokoy]

You should definitely let the devs know - best through a support request - that caustics failed in your scene.
I have just rendered my scene for 3 hours with caustics and it did work like expected - in v5, that is, and with HDR lighting, so that's a different setup - so make sure to provide the scene to the developers and see if they can figure out what went wrong. I could do the same in v6 tomorrow, just to test if v6 in general works differently.

[SharpEars]

I believe that the reason you are not seeing the photon issue is possibly due to rendering at fairly low resolutions. Try a high resolution render with a region select (of say one of the lower coils), so it finishes in a reasonable amount of time and see if you see the issue.

I'll post an image to show what I am getting, soon - trying a few more attempts to get to the bottom of it.

[GeorgeK]

You should definitely let the devs know - best through a support request - that caustics failed in your scene.
I have just rendered my scene for 3 hours with caustics and it did work like expected - in v5, that is, and with HDR lighting, so that's a different setup - so make sure to provide the scene to the developers and see if they can figure out what went wrong. I could do the same in v6 tomorrow, just to test if v6 in general works differently.

I believe that the reason you are not seeing the photon issue is possibly due to rendering at fairly low resolutions. Try a high resolution render with a region select (of say one of the lower coils), so it finishes in a reasonable amount of time and see if you see the issue.

I'll post an image to show what I am getting, soon - trying a few more attempts to get to the bottom of it.

I Will be checking this right away in case there is an issue.

Please submit a ticket or upload an archive of your scene, thanks

[pokoy]

I believe that the reason you are not seeing the photon issue is possibly due to rendering at fairly low resolutions. Try a high resolution render with a region select (of say one of the lower coils), so it finishes in a reasonable amount of time and see if you see the issue.

I'll post an image to show what I am getting, soon - trying a few more attempts to get to the bottom of it.

The scene seems to render fine here with v6.
I now see that you mention using region rendering - I think this is generally not supported and can cause unexpected results with caustics.

[kadir]

thanks for the share
i did a test with some tweak to achieve the look I wanted