[Thomas An.]

.... leaving the cavity as air works and making it another dielectric mat doesnt :/...

dd,

I suspect you are treating the cavity as a seperate object.
The cavity is part of the shell geometry (singe object). Just do a dome shape and then subtract (boolean) a little whole at the base to fit the emitter.

[lebbeus]

It's possible to use colored light sources--the image I posted used R, G, and B emitters, I also didn't have a cavity, so I shouldn't have been able to get anything out

[jalokin]

lebbeus,

When i look at the image you posted, it looks like the emitting object is white and the colored light is from the material enclosing the emitter. what does it look like if you remove the material that enclose the emitter on one of the LED

[lebbeus]

remodeled, now with a cavity; white emitter, no sss, 20 minute renders, SL over 11

This one is a two-layer bsdf plastic (transparent lambertian layer and reflective layer)


This is a single layer bsdf, 0 roughness, cool caustics, but the plastic really isn't lighting up


Single layer bsdf with roughness of 30

[glebe digital]

roughness of 10 on the epoxy, small sqr cavity around diode, SL 16 but I think this needs to cook for way longer as the preview is showing more brightness to come.

[ivox3]

Thomas An wrote:

If you were to merely stick the emiter imside the dielectric dome (without creating a cavity) then the result is "nothing". A ray of light will start and then hit the inner side of the dome; signifying that exits something ... but it doesn't know what it exits from ... because it never knew it was IN something. A ray needs to pass TWO dielectric surfaces to be complete. It needs to pass an incident and a non-incident face to know that it entered and exited (this is how it is in real life as well)

This makes total sense. I'll have a model screenshot shortly for review ....

Any thoughts on the epoxy properties itself, ...as in, ....what exactly is causing the reference photo LED that's optically clear to 'collect' the glow as it does?

Physically there's something chemically happening here that's definitively different than glass.

[ivox3]

Okay, ...time to hand out homework to the testers.


Happy reading ......... and you thought this was gonna be simple-- bah!

http://www.display-optics.com/pdf/LensD ... Davis3.pdf

http://www.display-optics.com/pdf/tech_ ... ct2002.pdf

[dd_]

thomas my cavity is a boolean from the shell :/
anyhoo more tests if time permits

[glebe digital]

At sl19 it's 'collecting' much better, this solution would need to run for a LONG time though.........


Scene here if you want to have a look, you'll need to load iors for the stainless enclosure.
http://www.glebedigital.co.uk/LEDtestC.rar

[dd_]

i noticed that the longer they cooked the better the light came, ah well

[glebe digital]

i noticed that the longer they cooked the better the light came, ah well

Yes I think the sss just takes SO long to come through, until it does it just doesn't look quite right.

Chris, the 'davis' workbook has some very interesting material in there........must read up some more on those fresnel lenses....

[jalokin]

Sorry for my ignorance in regards to colored emitters, Just was told that i had to reduce the burn value. Just realized how importent the burn and gamma values are with LED lighting.

[jalokin]

First test,

the geometri of enclosure is estimated and gives a little more lambertian distribution than exspected but close enough.

will try to create a more exact model later.

[ivox3]

hey Stu, ...yeah, I'm going to start over with a new model that considers what Thomas/Tom were saying and use the Davis work to stay true to the optical side of things.


.....we'll see.

[lebbeus]

this is Glebe's scene, changed roughness to 30 and bumped absorbtion to 1.0…I'll re-run this today with the same camera/angle for a better comparison. I'm thinking that surface roughness is more important than sss…
(sl13.45, 8hrs)