Thread: Making Nodes and Transparency work together...

1. The files

As promised, here's the files. Also, here's a render of what the adjust node graphs do now.

He Who Forgot What He Was Going To Do Next.

2. Dave,
Very clear and logical explanations. Thank you. Useful to many people here I believe.

Originally Posted by Dave Jerrard
...He Who Forgot What He Was Going To Do Next.
Working on a Node-manual and tutorials maybe?

3. Very nice explanation. Though you took all the beauty away from what I felt for the nodal approach.

My gradient (glass-air) isn't in any way related to the poly side. It's just a nice way to simulate the fresnel effect as it blends the reflection and refraction according to the incidence angle. I set the transparency of the front poly to 0 as I thought my refraction node would take care of this. This works nice, as long as your main objects isn't transparent, but only it's upper surface, e.g. wet watery film ontop of skin.

Your approach looks correct (but complicated - to me), but it's more or less the old surfacing way. For instance I wouldn't have a clue of how to deal with two or three different layers of a car lacquer with each layer having a reflective, refractive and diffuse and specular amount going your way, as you seperate everything. If you split up each of these attributes, how can you simulate the order of these layers or even combine them in a way that you can feel what each layer does? My way has been efficient concerning these materials.

Example: Bottom layer with a diffuse paint, middle layer with metallic reflecting particles and diffusion and a shiny top layer with reflection, refraction and little diffuse amount.

Probably I just have to get used to the way you see nodal, as you just shattered my view of the world.

The only point that remains is that I have to ask myself, why all these inputs. It would be much more convenient to have only one input in the Base node and you can mix up all reflection, specular and other nodes in one tree. What's the advantage of seperating the 4 major shading models (refl, refr, spec and dif)?

Thanks you very much for the time spent to help me out.
Thomas

4. P.S.: Wouldn't it be better to set reflection and refraction to 100% and control the amount of both of them by the amount of the reflection or refration node (e.g. with mixer and gradient)?

5. Originally Posted by Thomas M.
P.S.: Wouldn't it be better to set reflection and refraction to 100% and control the amount of both of them by the amount of the reflection or refration node (e.g. with mixer and gradient)?

When using node based reflections and refractions there is a separation of
aspects you have to cconsider. One aspect being the Amount of refraction
or reflection and the other being the properties of the materials that are
refractive or reflective (e.g. diamonds refract and reflect differently than
clear plastic). This later aspect we'll call shading. So we have "Shading"
and "Amount" for both reflection and refraction.

Refraction and reflection shading is controlled by any nodes that are hooked
up to those inputs in the surface destination node - "Refraction Shading" and
from those respective "Add Node" categories but you don't have to. This
"Shading" thing for these surface attributes is a new thing for LW.

"Amount" is controlled by all of the usual mechanisms [T], [E], or Numeric
constant values as entered into Surface Editor. It is also possible to set
these Amounts via node networks that are connected to the green dots
in the "Surface" destination node, labeled "Transparency" or "Reflection".
These are the same controls we've always had even if now there is a new
way to define them via node networks.

Originally Posted by Thomas M.
I wouldn't have a clue of how to deal with two or three different layers of a car lacquer with each layer having a reflective, refractive and diffuse and specular amount going your way, as you seperate everything. If you split up each of these attributes, how can you simulate the order of these layers or even combine them in a way that you can feel what each layer does?
There probably wouldn't be any refraction in car paint but this kind
of multi-aspect surface would be achieved in nodes by using Mixers.
Additionally you could also use a different normal textues (bumps) for
each "layer" thus enhancing the effect of difference and depth.

The point is, with non-transparent surface I don't have any problems to use very complicated nodal trees with loads of mixed settings.

The thing which bothers me is that I don't need any of the green inputs in the base node while working with diffuse shaders. They control both shading and amount of brightness, so there's no need for green diffuse or color settings . Everything works great with Diffusion shading.

Same for Specular Shading. No need for the green inputs. All can be done with my Specular shaders (incidence, amount, colour, mixes, ...) in the Specular Shading slot.

So, if it comes to Reflection and Refraction Shading, following your and Dave's explanations, the red input Shading slots merely tell LW what shading model is used, but the amount needs to be regulated with an "old" green input node? That's pretty much confusing me.

7. close...With reflection you still need to use the regular scalar input, refraction you do not...You can set refration in the refracton shading node.

8. Doesn't make any sense in my opinion. One input in the base node would be much more convenient than splitting the four of them up like this. No advantage as far as I can see, only disadvantages.

9. Originally Posted by Thomas M.
The point is, with non-transparent surface I don't have any problems to use very complicated nodal trees with loads of mixed settings.
When you get into more complex node maps, you will start to run into a lot of problems if you're not aware of how those shader inputs work, as you've already discovered with the transparency issue.

The thing which bothers me is that I don't need any of the green inputs in the base node while working with diffuse shaders. They control both shading and amount of brightness, so there's no need for green diffuse or color settings . Everything works great with Diffusion shading.
The Diffuse Shading overrides the red Color input and the green Diffuse and Luminosity inputs, and those corresponding channels in the Surface Editor. (It also overrise the Diffuse Sharpness control.)

Anything in those channels is ignored when Diffuse Shading is used. The reason is those three channels actually define the settings for the internal Lambert shader that LightWave uses. If you look at the Lambert node, you'll see that it has the same Color and Diffuse settings, and that if you plug the same textures into these and into the Surface Node's Color & Diffuse channels, you'll get the exact same result. What nodes offer here that you can't get otherwise is that you can have multiple shaders blending together, so you can have multiple specular highlights, each with its own color, like you can with BRDF or Gaffer, but with far more control. In fact, you can texture each one differently, even adding bumps to one highlight while the rest remain smooth - which makes for a really effective car paint.

The Luminosity setting is missing from the Shader Nodes for a reason. A diffuse shader controls how light is reflected off a surface - what falloff it has at various angles, etc.. Luminosity doesn't care about the lighting at all. Luminosity is actually a sort of shading override - the higher the value, the more the shader is overridden. If you take a surface and make it 100% Luminous and 0% Diffuse, it will look flat. It will also look exactly the same if you it 0% Luminous, 100% Diffuse with 100% Ambient light only.

Same for Specular Shading. No need for the green inputs. All can be done with my Specular shaders (incidence, amount, colour, mixes, ...) in the Specular Shading slot.
You're kinda right here... When you're using the Specular Shading, it overrides the Specularity and Glossiness channels (and Color Highlights), so no, it doesn't need them. Those are just the settings for the internal Phong shader, and you'll find the same settings in the Specular Phong node, but with one important addition - Color. Without the nodes, your specular color is limited to the light color, or, if you use Colored Highlights, whatever the surface color is, or its negative (if you have a negative Specularity), unless you use another shader plugin (Gaffer, BRDF, G2, etc.). If you want to do something like a pearl, where the surface is white but the highlights are many irridescent colors, you now need to set up a second surface slightly raised above the white surface, just to get those colors (100% transparent, multi-colored, with Colored Highlights turned on).

So, if it comes to Reflection and Refraction Shading, following your and Dave's explanations, the red input Shading slots merely tell LW what shading model is used, but the amount needs to be regulated with an "old" green input node? That's pretty much confusing me.
You don't sound confused because you just nailed it. That's exactly what those inputs do - let you define what shading - if any - is being used. You don't need to use these at all, because LightWave has built in shaders that all the old channels have been using for 16 years or so. So you can do most of your surfacing without ever touching these four shading input - just use the inputs above them. But, when you want to change a shading model, like to get a more realistic rock or velvet shading (use Minnaert or Oren-Nayar), or a better brushed metal (Anistropic shaders), or Subsurface effects (Kappa, Omega), then you plug those nodes into the appropriate shading inputs, and instead of feeding all the textures these look for into the Surface Node, you feed them into the shader.

Specularity works in a similar way, but the Specular Shading input is added to the Diffuse Shading's color. You could do this yourself, using a mixer or math node to combine it with diffuse shaders, and pump it all into the Diffuse Shading to get the same result, but then you'd lose the ability to utilize the Specularity buffer. It's also usually neater to avoid this kind of mixing.

Again, if you really want to, you can feed any output into this input, but in most cases, this wouldn't make any sense. Whatever is being fed into this - a solid color, texture, gradient, etc. - will have its output added to the diffuse shading.

Reflection and Refraction are also color inputs. These are each blended (each in their own way) with the diffuse shading based on the amounts defined in the Transparency and Reflection channels. These also override certain other channels... The Refraction Shading overrides the settings in Refraction Index (if a refraction node is used), Color Filter, Additive Transparency, Refraction Options (Environment tab), Refraction Map, Image Seam Angle and Refraction Blurring.

Reflection Shading similarly overrides the Color Highlights, Reflection Options, Reflection Map, Image Seam Angle and Reflection Blurring settings.

Again, you can pretty much feed anything into these since they're only color buffers, but it makes most sense to use the appropriate shader nodes. For example, here's the same ball rendered three times. I've plugged the Bump output of a Crackle node into the Refraction Shading input. With Transparency set to 0%, the ball only renders as a solid white ball.

Transparency is set to 50% for the next image, and you can see the wild coloring showing up from the conversion from bump information to color.

This next image has Transparency set to 100%.

The bump colors are most intense here, and you can see that the shadow is more colored as well. Refraction Shading overrides the Color Filter, and since I've given it some intense colors, these are used in the shadow as well as to define the color of the surface. I've replaced the color of what would normally be seen through the ball with the bump colors of the Crackle Node, which will only show up wherever the ball is at least partially transparent, as you can see here, where I added a texture to the Transparency channel.

This example doesn't exactly show anything useful for regular applications - it just demonstrates what these shading inputs actually contain; color values. As you play around with them more, you'll get a better feel for them.

He Who Didn't Even Like Nodes At First.

10. I do think that they should gray the inputs out that are over ridden. They should still let you plug into them to access the buffers if needed. It would at least let people know that it wasn't doing anything.

11. It starts to get into my brain. I don't feel it yet, but I probably understand it (if I want to).

The question remains: If I want to have a car lacquer with several layers, e.g. each reflecting with a different blur and different amount (upper layers take away light from lower layers), how do I do this with "your thinking"?

I didn't have any problems to create this stuff in the diffuse shading slot, but if I split it up and need reflection (green) to set the amount of reflection, how does this one parameter can resemble each layer of the lacquer?

BTW, standing ovations for your post.

Thanks
Thomas

P.S.: Same problem with a wet reflecting surface. The wetness adds a second reflection layer. How to seperate the reflection amount with only one green input?

12. Originally Posted by Thomas M.
It starts to get into my brain. I don't feel it yet, but I probably understand it (if I want to).

The question remains: If I want to have a car lacquer with several layers, e.g. each reflecting with a different blur and different amount (upper layers take away light from lower layers), how do I do this with "your thinking"?
This is where you mix reflection nodes. You can blend two reflections in a mixer, any way you want, and you can mix the output from such a mixer with another reflection with yet another mixer. And each reflection node can have its own bumps, independent of anything else. The same for specular highlights.

I didn't have any problems to create this stuff in the diffuse shading slot, but if I split it up and need reflection (green) to set the amount of reflection, how does this one parameter can resemble each layer of the lacquer?
You don't have to limit it to this one paramter. The Reflection and Transparency channels basically turn thee features on and at what 'volume'. But you can tweak the nodes themselves by adjusting the color of each one. The color setting can be used to control how clear or reflective they are - black basically turning them off and white turning them to full strength.

He Who Will Whip Up Some Examples In A Bit.

13. Can I just state, officially and for the record?

You guys are making my poor old brain hurt.

Seriously, this is absorbing stuff and I think I'm gonna have to print it out to read it.

Dave J, what can I say? you da man.

14. Node & Transparency as discussed in this thread in a PDF.

15. nice one jmcinnis, i was just thinking there was a manual or faq in this ;-)

thanx dave, also helping me get my head around it too.

Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts
•