I was searching the renderosity forums for this topic with little success, so I thought I'd start one as a reference.

In my tests, I found that ice is an interesting challenge. Things I had to consider:

1. The type of ice I was trying to simulate (e.g. glacial ice vs the common ice cube). I tried to reproduce both but I eventually focused on an ice cube like material. I used this stock photo as a reference:

2. The material has to be a cycles-based material

3. The material had to work irrespective of the environment it is put in

I've had moderate success using a material consisting of 4 layers. This material works irrespective of the environment (assuming at least one light source is present):

Render with light grey background (ground set to visible):

Render with black background (ground set to invisible):

I searched youtube for some blender tutorials on making ice. I used one as a guide for making the material in poser. The setup is basically 4 layers (including the base layer):

I still need to tweak this setup in that the the center of the ice is opaque (as can be seen in the reference photo above).

Any suggestions would be appreciated.

Fun! I'll keep watching this topic.

Have you tried just to render it with the GlassBSDF with the IOR of 1.310?

Also you might want to try to add more ice cubes to your scene if you want to compare it more easily to the references image, as the ice most likely reflect in each other in that one.

Last I don't really see why you would add an emission to the material unless you want it to cast light?

3D-Mobster posted at 11:16AM Tue, 05 February 2019 - #4345343

Have you tried just to render it with the GlassBSDF with the IOR of 1.310?

Also you might want to try to add more ice cubes to your scene if you want to compare it more easily to the references image, as the ice most likely reflect in each other in that one.

In tutor4u's youtube Blendor tutorial, he doesn't really explain the reason for adding the emission material. My assumption is that it enhances whatever light is being refracted by the GlassBSDF material. If I only use the GlassBSDF, the result seems bland to me (see below) but then aesthetic is relative:

Dang, that looks really good! Nice work.

I see, however I do think that it is an incorrect approach for several reasons. You write in the first post:

---

3. The material had to work irrespective of the environment it is put in

---

Which I fully understand and agree with why that is important to you and is one of the main reason, why you would want to use PBR in the first place and stay away from Firefly. Cycles and the Physical material node is in Poser are PBR nodes. Even though the standard Poser surface node can render in Superfly, Im not really sure that its well suited for PBR rendering, maybe someone can expand on that, if they know more about it. But for me at least, I stay away from it, except when I export into 3ds max as materials doesn't seem to work or get exported or imported there "correctly" otherwise.

PBR in it self make sure that materials look correct regardless of lighting conditions, which is basically the environment as the reflection of light from objects determines how we perceive them. Which is why PBR is so much better than for instant Firefly rendering, because in Firefly if you change the lighting, lets say from day to night, you can basically screw up a material, which can cause you to having to adjust either the material or the lighting to get whatever look you are after, which is both time consuming and annoying. And then when you change the lighting again you might have to go through the whole process again. Which is also why I personally don't think there will be any major, if any further development of Firefly in future Poser releases to come, but that is just a prediction :D

However when talking about PBR and why I find the emission strange, is because its sort of the same approach as someone "faking" materials for Firefly would use. But it ain't necessary and advisable to do with PBR, unless you are deliberately going for a weird effect of some sort. Because it ruins the purpose of PBR as I see it. Very few objects in real life have emissions, lamps, the sun, some animals etc have it. But for the most part very few things. So when you add it to a piece of ice and as you say you want it to work in any environment, it won't work during a night scene as the ice will suddenly cast light.

So I think a good way to look at it and the image you rendered above saying it "looked to bland for you", which I agree it looks less interesting. But keep in mind and compare it to a real life setup. You have taken a single piece of ice and placed it in a light grey environment, with nothing to reflect. I doubt that a piece of ice in real life under such circumstances would look very interesting either.

So what I would suggest you to try is to hide the background and add a HDR map to cast light so there is something to be reflected. Even try to delete all other light sources in the scene, maybe put it on a floor, duplicate it a few times and place more in the scene and then try to render again to see if it looks more interesting. Also remember to turn on caustics for both refraction and reflection.

If you are not sure how to set up HDR lighting let me know and ill show you. But I have to eat now :D

Very good points. I investigated this further and eventually I realized I made a few newbie moves:

1. Poser material layers work primarily when each layer has some level of transparency or mask to allow the material layer beneath it to be seen in the render. So the initial setup I shared is completely wrong. The base layer contained an emission material that was completely hidden by the layer above it. A really basic mistake on my part but also now I have a much better understanding of how material layers work and how powerful they can be.

2. True enough, the emission node should not be used at all (for a real world material). I've removed it. To do a "real world" material requires high sample amounts in the superfly render settings (a relative pain depending on the power of the nvidia card). I've read some posts of people using some CRAZY numbers for sampling to get acceptable results. Again, crazy is relative. But here are my results using 50 samples (single light source, with medium-grey background and ground set to visible):

In this render, same as above except with black background and ground is hidden:

I also plugged an ice bitmap into the GlassBSDF's roughness which obscurred a lot of noise but at the cost of transparency (no surprise there):

So basically, yes "realistic" ice cube like texture is do-able but at the cost of high sampling in the render settings. One cheat is to plug the Emission node into a Poser Surface volume and adjust the brightness/color of the emission to your preference:

Very nice, to me at least these images looks much better and interesting than the first ones you posted. I still think you should try HDR lighting, simply to get those reflections into the scene and maybe an ice ground for it to sit on again just to further expand it. The last 3 image looks nice I think, despite being in a dull gray environment, and even though the last one with emission added, might not work in all environments and I think you could achieve the same effect simply using a "correct" light setup in your scene, its looks very cool.

In regards to pixel samples I don't think that 50 is all that much to be honest, a cheap trick is to render slightly larger than your final image and then scale it down, since that will force it to compact the pixel a bit and can therefore reduce the noise, but it can reduce the number of pixel samples needed slightly, whether its worth it or not depends, as it will reduce overall quality slightly. Also you could try to scale it and rotate it and place more of them inside each other to see if that give some more depth and effect to the ice.

Your right, 50 isn't that much and yes an HDRI map would be a big benefit. To be honest, I don't want to think about materials. I only want to apply them to an object with a reasonable expectation that they will render accurately and relatively fast. Most of my thought is dedicated to what I am trying to communicate in the scene. For the most part, this has been the case but like I said at the beginning, ice is not as simple a material as I had expected and I needed to research it. Thanks to your feedback, I think I have something fairly close.

I hope some of this info is beneficial to any of you that are interested in the topic.