PDA

View Full Version : Luminosity Falloff?



IMI
12-30-2007, 04:21 PM
Is it possible to create a falloff for a luminous object, where, say, it's brighter in the middle, but less so outwards from the center? I thought maybe it could be done in the standard surface editor with a procedural, but to see any kind of effect, I had to turn the luminosity down low and the transparency up a little. Using a trans map or procedural does the effect a little better, and I can turn the lum up higher, but adds the transparency/render time factor to it, and doesn't really cut it anyway.
Nodes?

cagey5
12-30-2007, 05:17 PM
What about a gradient linked to a weight map?

IMI
12-30-2007, 05:31 PM
Oh man, again with the weight maps...
Thanks, it sounds like it would work. I never used weight maps for anything before besides rigging and I just never think about it for anything else. Seems like half the questions I ask get answered with "weight maps".
I have to do something about remembering to try that first - gonna put a big WEIGHT MAP sticky on my monitor now.
Thanks again. :)

oldtekerr
12-30-2007, 05:33 PM
you can do it with any texture...any type of gradient.

IMI
12-30-2007, 05:43 PM
And there's you always shoving gradients at me. ;)
That's it - I vow to spend the entire last day of this year studying gradients and weight maps in as many combinations and uses I can think of. Then I'm gonna go out and drink myself into a stupor. :D

oldtekerr
12-30-2007, 05:46 PM
Gradients are like the Force... learn them well Luke...

SplineGod
12-31-2007, 12:21 AM
You can also set the gradients input parameter to distance to object. :)
You can also use a procedural with sphereical falloff or other types.
The procedurals falloff can also be further modified by using a gradient with the input
parameter set to previous layer.

Cageman
12-31-2007, 12:48 AM
Weightmaps are, in my honest opinion, one of the best thought out features in LightWave. Multipurpose stuff is... magic! :)

IMI
12-31-2007, 11:16 AM
You know, the problem with gradients is that it's not documented well enough. There's alot written about gradients, yes, but considering the sheer amount of what all can be done with them, the information out there is way disproportionate. I've seen loads written on "how to do x..." with gradients, but I don't think I've seen anything which explains exactly what every feature is. I do know how to use them in the "old" surface editor, and have used gradients for color, bump, trans and so on, but when it comes to their more advanced potential, I'm clueless.
And the "preview" has never been much help for anything other than color, diffuse and reflection, IMO, and often what you see is way different than what you get when you press F9.

This is one of many reasons why LW needs an FPrime-type preview renderer, also IMO. One which works in 64 bit and works for all the nodes. VIPER sucks for anything more complex than hypervoxels (which it does really nicely).

For something as complex and powerful as gradients, I'm surprised nobody has written a 1,000 page book or ten-hour video lecture by now, specifically about gradients. I'd damn sure buy as much as I could on the subject of gradients, particularly gradients in the node editor.


And I can't remember ever having written anything with the word "gradients" appearing so much before. ;)

IMI
12-31-2007, 01:27 PM
Well, this is close to what I was looking for in that I figured out how to control the falloff of the luminosity, with a gradient in the node editor set to Incidence, and plugged into the Luminosity channel.
For the life of me though, I can't figure out how to make it become more transparent or fuzzy from the center outward.
I'm thinking along the lines of a fuzzy, cloud sort of object. Maybe not completely transparent at the edges, but...fuzzy.
The Thickness node only has one option - refraction index. I don't get it - I would think that it would adjust thickness of the object. I don't see how refraction relates to thickness. Well I do, but I don't see how refraction is supposed to control thickness. Seems it would be the other way around. Clearly I'm using it wrong. I tried it with double-sided and single-sided. I plug its Length into the gradient's Input and I lose the ability to select Incidence. How does that make sense? What does the refraction index have to do with the angle? I plug Hit Point into Normal, and I get weird refraction. If I use a negative refraction value, I get closer to what I want, actually, but end up with undesirable pinching at the top of the ball.
I'm truly stumped on this one. I'm sure I must need some of those fancy math/vector/function...whatever nodes somewhere in between, but haven't a clue where to start. :confused: