[oz42]

Following on from my experiments here - http://www.maxwellrender.com/forum/view ... 3&start=15 - where I built a studio based on hyltom's LDR map so that light levels could be altered dynamically with MultiLight, I asked whether anyone knew of a way of also altering the colour?

I've modified my studio so that the Front, Left and Right light planes are split into vertical strips of Red, Blue and Green. With them all turned on 100% they, obviously, produce white light but because they are separate emitters, they can be altered to affect the colour of the light as well as the intensity.

The following set of images, starting with all the emitters on so that they produce white light, are all created from the same mxi by only adjusting the levels of the emitters in MultiLight.

The only disadvantages are that the MultiLight panel of mxcl gets pretty big (make sure your naming strategy is sound!) and the mxi gets pretty big (=size of a normal mxi x number of lights x three colours per light!!!)

Again, please excuse the very, very simple stand-in geometry - I'm just trying to prove a concept rather than produce a great render at this stage!

[tizxx]

Hi Oz your experiment is very beautifull... ... and much interessed ...

sorry for my bad english, I'am italian...

I try to reply your expreriment with unsuccessfull results...
please can you post your scene...


So many of us can understand better your expreriment...

many thanks

[Luca_Studioaltieri]

i did a similiar project for a lcd-lamp company and result was very good. and as real rgb-lcd-lamps, shadows have 'three colour border' if objects are near light source. i have to find final images of that work, and post it, so you can see application on a "real" scene.

[fractrix]

I did something similar SOME time ago and it works just fine,
if you are interested:

http://www.maxwellrender.com/forum/view ... highlight=

[oz42]

tizxx - no problem here's the scene although with reflective objects you get the problem highlighted on the second page of fractrix post http://www.RichardOsborn.com/maxwell/wh ... Studio.rar - enjoy

fractrix - great thread, sorry I didn't see it the first time around. You also spotted the problem with reflective materials. Nice solution with the diffuse reflectors.

Reflective material problem without diffuse reflectors: (there's also some weird cyan artefact at the top (no cyan lights used!))

[tizxx]

Thank you oz42 ...

I think that the emitter must be one only source, these esperiment
are very nice but has some artifacts in some situation.

What do you think about use 3 plane in stack, instead 3 plane in linear
where every plane has 1 light emitter (red, blue or green) plus 1 transparent layer without reflection / refllaction and custom Ior Nd = 1.0

your metod probably is more efficent if the planes (3) are very very little, like a pixel of a LCD pannel...

however very nice job...

probably this feature (Multilight Multicolor) we will see in the next major release of Maxwell Render - with RS2

When I have more time I try my idea of 3 plane in stack

bye bye

TiZxx

Tiziano Silvestri - Maxwell Render dot IT STAFF
www.maxwellrender.it

[oz42]

tizxx - if you re-read the old thread from fractrix he comes up with a clever solution that you may want to try.

[tizxx]

Yes, this is what i suppose to use...

3 plane = 3 emmitters

same position same normals

I don't remember how we can make invisible emitters

but probably if i blend emitter with a BSDF ghost layer i have the same effetc

Then i not understand the method to use a reflection plane

I'll try I'll try

very sooon I'll post some images.

[tizxx]

This is a very little experiment... but works fine

look at this link:

http://rapidshare.com/files/158174999/MM.rar.html

I made all with maxwell studio so the geometry is very very simple

for the emitter i make a plane and then i cloned it 2 times
for each plane i assigned a different color emitter (red, green and blue) +bsdf layer to make the emitter invisible...



try it and tell me what do you think...

the green emitter was 5 times efficent than others emitters material so we have:

red emitter equal 100W and 63.8 efficacy
green emitter equal 100W and 319 efficacy
blue emitter equal 100W and 63.8 efficay

or are the same

red emitter equal 20W and 63.8 efficacy
green emitter equal 100W and 63.8 efficacy
blue emitter equal 20W and 63.8 efficay

good night

[oz42]

tizxx,

pretty good test, although with only one light you could achieve similar results with post process colour correction. The method really come into it's own with multiple lights so that you can get results you never could in post.

This is a 3D method of something I remember from Paul Haeberli many, many years ago - http://www.graficaobscura.com/synth/index.html Have a look at the rest of his site, although it's old it is still very interesting.

[tizxx]

Hi Oz42,
the site is very cool ... ... but I think this situation is more simple
3 different light produce a photo with 3 different color...

the proble is other ...

create a light that can be dinamically change intensity and color without artifacts ...

if you clone the 3 plane in anothere 3 plane we have 2 multicolor multilight situation...

what i can't understand is Why the green emiter must be 5 times more power than red or blue emitters...

and another proble is that this kind of light are invisible and this situation is not a real and probably not phisically correct althout Maxwell can do that...

the I think another situation:

4 plane with 4 emitters:

1 plane has a white emitter or a dominant color emitter
1 plane has red emitter
1 plane has green emitter
1 plane has blue emitter

this evening i'll this new situation with 2 light source

I know the mxi file is more big but probably is the correct solution.

[oz42]

tizxx, the reason for the green light needing more energy is explained in tom's post in this thread http://www.maxwellrender.com/forum/view ... hp?t=15075, it's to do with the distance of the colours to the curve. fractrix found the same issue but it wasn't 5 times the value (see his first post in the same thread).

Good luck with the tests, I look forward to seeing the results.

[tizxx]

My God...

the color of light is not a linear solution as i supposed...

but if the range of each emitter is 10,000 or 100,000 we can simulate more colors because the intensity is more precise...

visually is very importat to know how to make a neutral light,
the correct distance between red green and blue,
if you render the scene with green emitter 5 times more power than blue and red we have a correct white light enviroment...

if these situation will be normalized to white, then is possible to make a table that correspond intensity of lights and correct color light

what do you think about that ?

[oz42]

tizxx, I'm afraid the maths is beyond me but tom or Mike Verta might be able to help. In Mike's material tutorial - http://www.maxwellrender.com/forum/view ... hp?t=22956 he makes a compelling argument for the use of 242,242,242 as a colour for reference white (as opposed to 255,255,255). I think it's in the second video?

[rob rhodes]

crikey this looks complex! Im surprised NL havent introduced multicolour yet as it would just seem to be some sort of RGB filter on each layer which doesn't sound that tricky compared to creating multilight in the first place. Fryrender has had this from early betas and NL have had quite a head start.