Crystals with depth

[GManB]

I am trying to create some crystals in SL. Did a tourmaline firs and that looked... fair to mid.. https://gyazo.com/9b47e6de1c9ad39b5d63037b7dd22a22

I've tried a frew different techniques

1. Separate thin (1mm) objects for each facet (outer face and edges transparent, inner face black and high shine). Each facet is a projector. The tourmaline above is this.
2. Single object with normals reversed so invisible from the outside, black high shine inside. External projectors surrounding the crystal. ok, but still a bit weird
3. Single object normals normal (hehe) black high shine. External projectors. Ok, facets show up nicely but no depth 
4. Single object normals normal (hehe) black high shine. External projectors. Procedural generated pattern normal map to fiddle with the light. Maybe a bit better but still no depth and still a bit weird.

Here is the diamond in Maya rendered with Octane.. Just a simple Octane Specular material and a light in the center.. IoR = 2.42 for diamond. Looks nice.

https://gyazo.com/b6bde8123148520cccf7676aada5e701

 

Here is what I have in SL. Using 4 above.

https://gyazo.com/c7ab046381217040588ed31646dce679

 

Anyone have any ideas about a technique that I could use to get more depth and refraction in the crystal?

 

Thanks,
G

 

[Macrocosm Draegonne]

Looking good so far, I've thought about this myself, i bought some crystals from Candle & Cauldron a while back that are nice, their salt lamps are sweet too. She seemed to use copies of the mesh shrunken inside to create more depth, having the outermost layers be a bit more transparent stepping inward to inner sections adding more opaqueness.

Looking at your final resulting test there, perhaps make the black area inside more transparent, and your specular map conform to the texture of the normal, or at least not a solid gradient on the inner shines, then on the outer flat "carved & polished" layer, have the solid gradient shine, but less pronounced, so it doesn't wash things out as much.

Not sure if that helps any... lol... but it was something i was thinking about recently myself, so I thought I'd share.

Maybe we will get more shaders to work with once the transition to vulkan happens? Probably.

Side note: this reminds me of the bling craze that swept SL many years ago lol, people went overboard with it. 😂 Not saying you are, or even that it matters, just a funny memory.

[GManB]

Thanks Macrocosm!

I did some experiments with smaller versions inside and playing with the transparency. Didn't achieve much. Even with 1% transparency the outer layer loses all reflectivity and doesn't see to help with depth.

I did stop by C&C and there are some interesting crystals there. More toward translucent (which I am thinking about trying next as well).

Any idea on when vulkan might be available? I checked around the web but didn't see any hints on timelines for viewers.

Yes, I'll admit to liking shiny sparkly things..

 

Thanks again,

G

[Macrocosm Draegonne]

On 8/22/2020 at 2:11 PM, GManB said:

Thanks Macrocosm!

I did some experiments with smaller versions inside and playing with the transparency. Didn't achieve much. Even with 1% transparency the outer layer loses all reflectivity and doesn't see to help with depth.

I did stop by C&C and there are some interesting crystals there. More toward translucent (which I am thinking about trying next as well).

Any idea on when vulkan might be available? I checked around the web but didn't see any hints on timelines for viewers.

Yes, I'll admit to liking shiny sparkly things..

Thanks again,

G

Yeah, it will take lots of iterations and experiments, seems like one of the more difficult things to pull off, but you're making good progress. The translucent effect they used does look quite nice, im sure they tinkered for a long time to achieve it.

On the Vulkan front, well im sure that will take a bunch of time, who knows though, they've already placed detection code in the viewers to ascertain whom among users already have a vulkan capable machine. Given that Vulkan was created by the same folks who created openGL (which SL is based on) there are a lot of useful ways to transition, keeping older stuff workable as its evolved and such. Who knows how long it could take, im sure they've got their sights set on a timeline, but wont release such info until its solid and certain. I always like to hope for the best, and be pleased with whatever is, this way im not often disappointed.

Given the unique nature of SL and mega data transfer concerns, I would wager using vulkan to bring in more shaders would be high on the list, this way far more texture/appearance variations can be made using less and even smaller actual texture files, depending on the effect. The other biggie is the lighting IMO, that has a huge potential to make things look quite amazing with less resources. Ex. I put a request on Jira a while back for underwater caustics lighting (the way light goes thru the water surface and dances on things below), so people dont have to put up tons of spinning projection light-prims, you'd get a better effect, using far less resources.

Anyhow, im rambling on again lol, I wish you good luck in your crystal making adventure!

[GManB]

2 hours ago, Macrocosm Draegonne said:

Yeah, it will take lots of iterations and experiments, seems like one of the more difficult things to pull off, but you're making good progress. The translucent effect they used does look quite nice, im sure they tinkered for a long time to achieve it.

On the Vulkan front, well im sure that will take a bunch of time, who knows though, they've already placed detection code in the viewers to ascertain whom among users already have a vulkan capable machine. Given that Vulkan was created by the same folks who created openGL (which SL is based on) there are a lot of useful ways to transition, keeping older stuff workable as its evolved and such. Who knows how long it could take, im sure they've got their sights set on a timeline, but wont release such info until its solid and certain. I always like to hope for the best, and be pleased with whatever is, this way im not often disappointed.

Given the unique nature of SL and mega data transfer concerns, I would wager using vulkan to bring in more shaders would be high on the list, this way far more texture/appearance variations can be made using less and even smaller actual texture files, depending on the effect. The other biggie is the lighting IMO, that has a huge potential to make things look quite amazing with less resources. Ex. I put a request on Jira a while back for underwater caustics lighting (the way light goes thru the water surface and dances on things below), so people dont have to put up tons of spinning projection light-prims, you'd get a better effect, using far less resources.

Anyhow, im rambling on again lol, I wish you good luck in your crystal making adventure!

Macrocosm,

Thanks for the rambling! Good to hear someone wanting more graphics capabilities in SL. The difference between what the VFX folks can do and what we have here is striking. We'll never I think get real-time, online worlds close to that but some games now are closing the gap a bit with greater use of PBR.

For now, I'm going to focus on other projects while considering what other techniques I might use. Here is an animated gif of the technique that looks the best.

https://gyazo.com/44426a337de0a164ca1ad7a070000a22

I used a single-object mesh for the diamond. Normals outward. Black and High Shine for the texture. Created  a normal map from a greyscale of tight, narrow, vertical-oriented shapes. But but stretched the map by setting the texture 'repeats per meter' to a tiny value. Around the diamond are 22 projectors. 11 with colored, tight, narrow, vertical-orineted shapes and 11 with just white light.

 

Cheers,

G

[animats]

3 hours ago, GManB said:

The difference between what the VFX folks can do and what we have here is striking. We'll never I think get real-time, online worlds close to that but some games now are closing the gap a bit with greater use of PBR.

Yes. The encouraging thing is that how to do good rendering has become more standardized. For a while, people were coding custom shader programs for each object and wiring them together. Pixar had a big operation in San Rafael, California doing just that. Eventually, people at Pixar/Disney figured out a standard set of shaders that could do almost everything needed for movies. That became "Principled BSDF", which is a specific way to do physically based rendering with a standard set of layers. Then that made it into all the tools, such as Blender, into graphics APIs such as Vulkan, into GPU shader programs so it goes fast, and into game engines. This made artists much happier, since they could draw instead of program. So today it's pretty clear what direction to go

[Kyrah Abattoir]

16 hours ago, animats said:

For a while, people were coding custom shader programs for each object and wiring them together.

We still do to a degree but you usually write partial shaders nowadays.

19 hours ago, GManB said:

The difference between what the VFX folks can do and what we have here is striking.

VFX People don't mind spending a few hours on a single screen frame. Despite having dedicated graphic cards, you have to keep in mind that we are trying to get the best possible result that can be achieved in a rendering time of just a few miliseconds.

Edited August 29, 2020 by Kyrah Abattoir

[Macrocosm Draegonne]

On 8/28/2020 at 11:28 AM, GManB said:

Macrocosm,

Thanks for the rambling! Good to hear someone wanting more graphics capabilities in SL. The difference between what the VFX folks can do and what we have here is striking. We'll never I think get real-time, online worlds close to that but some games now are closing the gap a bit with greater use of PBR.

For now, I'm going to focus on other projects while considering what other techniques I might use. Here is an animated gif of the technique that looks the best.

https://gyazo.com/44426a337de0a164ca1ad7a070000a22

I used a single-object mesh for the diamond. Normals outward. Black and High Shine for the texture. Created  a normal map from a greyscale of tight, narrow, vertical-oriented shapes. But but stretched the map by setting the texture 'repeats per meter' to a tiny value. Around the diamond are 22 projectors. 11 with colored, tight, narrow, vertical-orineted shapes and 11 with just white light.

Cheers,

G

Thats looking pretty fabulous, great progress! I wish I could finish something myself... 🥴 I keep starting different things and bouncing around, life distractions & work tend to pull me away, but I love 3d arts and have been consistently learning & tinkering at least. Its such a massive array of skill sets, and I cannot settle into just one area,.

I have finally started some SL projects I think I'll complete this time though. If I could only keep myself from drifting around in all the complexities and what ifs, planning, iterating and such. Not sure I can do that anyway lol, "its an INFJ thing" they say.  I have become far better at meshing/texturing/lod and scripting though, so hopefully my creations will be decent enough to inspire me forward on to more. At least im having fun, thats the most important part by far.

I just need to keep in mind that 'art is never finished, only abandoned' as DaVinci said.

Good luck on your other projects, and crystal developments. 💚

[Macrocosm Draegonne]

Kyrah is right, its the variations we'd get.  Indeed though, it would be amazing to have PBR in SL, and those variations made available to creators would be rather epic when you consider how that would transform new creations going forward, and of course they'd need some legacy shaders for whats available now, those using plain textures, those with advanced materials, and the auto-normal/specular settings.  Could still maintain the 1024 size limit, but those shaders could have masks, detail maps, blend multiple textures, sub surface scattering, etcetera, etcetera. 😍