Different way of doing lights

Feature requests, bug reports and related discussion
Post Reply
8 posts • Page 1 of 1
m1j
Posts: 45
Joined: Fri Sep 22, 2006 12:05 pm
Contact:

Different way of doing lights

Post by m1j » Wed Apr 18, 2007 8:19 am

I would like to see the light system changed from being a meshlight to being a material for models. As an example a material called light or perhaps any material with RGB values above 1 would illuminate the scene.

I have built scenes in obj format and had trouble because the object I wanted to be a light could not be used that way embedded all together.

So again rather than a meshlight or rectanglelight a light material would be better.

User avatar
Kram1032
Posts: 6649
Joined: Tue Jan 23, 2007 3:55 am
Location: Austria near Vienna

Post by Kram1032 » Wed Apr 18, 2007 8:30 am

That request already was made multiple times and is on the HUGE list of todos ;)

But it will be the same settings, as before, just the behaviour will change:
I.e. for now, you can't use a primitive sphere to light something: it's a primitive and not a mesh.

AND you can't blend them, yet.

RGB +1 only would be a material, that reflects more light, than it gets. Not a true emitter, but a "power-boost" or something like that... -> not very realistic and I guess, Indigo wont ever have that feature.

The mesh wont need to be called light. That's the costum behaviour of the exporter-script-writers. The material will be called light, though, I guess :D

m1j
Posts: 45
Joined: Fri Sep 22, 2006 12:05 pm
Contact:

Post by m1j » Wed Apr 18, 2007 9:04 am

RGB > 1 would be realistic when looking at how objects reflect or emit energy. A black object absorbs most energy while a white objects reflects most energy. An object with RGB > 1 would have an energy state at which it would emit more energy than it received. A light bulb works this way. The filament is a gray color emitting just part of the energy it receives. Once electricity is applied the filament now has enough energy to emit it as light. It is more than white. This is what I was talking about.

mrCarnivore
Posts: 517
Joined: Sun Mar 04, 2007 6:20 am
Location: Stuttgart, Germany

Post by mrCarnivore » Wed Apr 18, 2007 7:19 pm

m1j wrote:An object with RGB > 1 would have an energy state at which it would emit more energy than it received. A light bulb works this way.
That's not correct! You didn't think it through. If that were true, a light bulb would become even brighter when it receives more light. Which it doesn't.

A rgb value of above 1 would mean that the incoming light is multiplied with that value which is the same as for rgb values of <= 1) and therefore the emission energy would be proportional to the incoming light. A real light bulb just emits the same amount of light all the time + the amount of light it reflects (which is negligible).

So rgb values of above 1 are absolutely unrealistic!

User avatar
OnoSendai
Developer
Posts: 6241
Joined: Sat May 20, 2006 6:16 pm
Location: Wellington, NZ
Contact:

Post by OnoSendai » Wed Apr 18, 2007 7:42 pm

mrCarnivore wrote:
m1j wrote:An object with RGB > 1 would have an energy state at which it would emit more energy than it received. A light bulb works this way.
That's not correct! You didn't think it through. If that were true, a light bulb would become even brighter when it receives more light. Which it doesn't.

A rgb value of above 1 would mean that the incoming light is multiplied with that value which is the same as for rgb values of <= 1) and therefore the emission energy would be proportional to the incoming light. A real light bulb just emits the same amount of light all the time + the amount of light it reflects (which is negligible).

So rgb values of above 1 are absolutely unrealistic!
yup, mrCarnivore is correct.

User avatar
CTZn
Posts: 7240
Joined: Thu Nov 16, 2006 4:34 pm
Location: Paris, France

Post by CTZn » Wed Apr 18, 2007 8:20 pm

m1j, that idea you have sounds like how Final Gather works in some render engines including mental ray, am I correct ?
obsolete asset

m1j
Posts: 45
Joined: Fri Sep 22, 2006 12:05 pm
Contact:

Post by m1j » Thu Apr 19, 2007 3:34 am

To be precisely accurate the surface would have to be described in terms of atomic sub structure and temperature. The density and pattern of atoms determine the reflected color while the temperature determines the emitted color. A possible compromise would be to add a temperature element to the material section.

I was using the same compromise applied when using RGB to set surface color to set the emitted color. Neither is accurate but they are easy to set and understand. Indigo currently allows both methods to set emitter data. This could be moved into the material block and the light block could be eliminated. If the RGB of a material is above 1 then it is assumed to be the RGB of a light source -1. If below 1 and accompanied with a temperature block at a value to emit light it would again be used as a light source.

(I forgot to consider the fact of a filters effect on the color of light.)

Regardless of how it is implemented the ability to make any material an emitter expands indigo well beyond other rendering engines.

How does indigo use the RGB data of a HDRI to add emitted light to a scene?

User avatar
Kram1032
Posts: 6649
Joined: Tue Jan 23, 2007 3:55 am
Location: Austria near Vienna

Post by Kram1032 » Thu Apr 19, 2007 3:43 am

For "realism", you'd only need one material...
a real pure diffuse material doesn't exist, but it's the roughness of the material, that gives the diffuse effect...

You'd only need specular + SSS + temperature (emitting colour...) + light intensity + flourescense peak (I guess) as an other colour + possibility to define different IoR at different wavelenghs ( .nk-like) + cbc + a VERY highdetailed -> hires -bumpmap.
(Or you "just" model every single Atom xD)
Then, i should be possible to make ANY material...
except if I dunno one phenomena, that would be needed, too...

But high SSS, as needed for nearly diffuse, wouldn't render too fast... -> TOO much time for that kind of realism for nearly the same result...

Temperatur in materials would be nice, though ;)

Post Reply
8 posts • Page 1 of 1

Who is online

Users browsing this forum: Bing [Bot] and 30 guests