These latest experiments were mainly to do with cracked & broken ice… which requires a balance between believable ice shapes and a semi-convincing ice shader that works at both a distance and for extreme close-ups.
Ice comes in a vast number of varieties with all kinds of looks influenced by temperature, environment, impurities, etc. My experiments here represent examples of only one or maybe two types of ice… mainly high-end food service ice which is known for its lack of impurities.
I plan to keep experimenting to get different looks which may be closer to what you imagine when you think of ice or something that more closely resembles the ice that comes out of your freezer. Plus lots of other kinds because I enjoy working on these sorts of things.
Nice, I liked the broken ice pieces! Did you make them by hand or using the cell fracture addon?
For the ice cubes inside the glasses this looks great, but I think for ice to look nice in the foreground you would have to use volumetrics. The inside of ice is quite complex and is what makes it look like ice.
I used the cell fracture addon along with subdivision and remeshing and decimating (various times). In the future, I will likely model by hand because I prefer all quad geometry and it’s a lot more efficient.
As for the look, ice comes with all kinds of looks depending on environment, temperature, etc. So it would be very difficult to make an ice shader that applies to all circumstances. I agree that adding volumetrics would enhance the look for some kinds of ice for sure.
The ice here resembles my reference. But I’ve seen all kinds of ice in my research (and some a lot more interesting than this). Thanks for your feedback. I’ll continue my experiments and maybe try to get the look of some of it.
More procedural ice experiments… I tossed this scene together as a lighting experiment, but I have plans to build out an interior space which feels a bit more unified & esthetically intentional.
*When I say “tossed this scene together”, I mean used models and shaders I’ve already created or created new ones (such as the ice tray/holder) to “make a scene”.
Every material is procedural except the book cover.
I’ve found that it’s relatively easy to make ice “work” in a specific scene under certain lighting conditions, but making believable ice that works in many different lighting conditions and at different distances (such as extreme close-ups) is much more difficult.
Its been a process, but I’m enjoying trying to get it right. And I’ve learned a lot along the way.
It’s also fun because ice comes in so many different varieties.
I’m very curious about the ways you choose to experiment things in this whole process… I mean, how you organize your process and how do you know from where to start…
I totaly agree about the lighting conditions. Lighting can save or destroy a scene, depending the way its used…
That being said, I am enjoying to follow your discoveries here. Looking forward for more updates…
Whether it’s modeling or creating a shader (or a combination of both) I usually just start with a challenge or something I think would be fun to do (preferably it’s both). And then I try to accomplish it.
Creating realistic procedural shaders is still new to me. Most of my prior professional experience is in modeling and texturing. Before I started learning Blender in 2022, I’d never really used node-based software before…
Then I discovered I enjoyed it (I still have so much to learn).
I started with wanting to learn how to create convincing glass as it involves more than just making a material transparent. Then convincing liquid. Then ice. And then procedural fruit. They’re all somewhat related.
I’ll visit something for a while then go back to something else using what I’ve learned in various areas along the way.
I’ll also model things specifically to apply shaders to or to help build out a scene. So it all just kind of feeds itself.
Ever since I started learning Blender, I’ve been more on a photorealism kick. I don’t always succeed (but that’s what learning is about), but I enjoy trying to get things to look semi-convincing as it is quite challenging. Although, sometimes the solutions are simple, but you have to experiment a lot to find them.
But a lot of those failures end up being great learning experiences.
