Strange light emission

[Jambert]

I use neon light. I modeled neon with perfect circle extruded. I don't understand some shadows!?

Help needed, thx

Sorry for blured picture but it's a private job.

[fused]

thats weird.... can you show a wireframe of your lightsource?

[Jambert]

Some tests

[OnoSendai]

Maybe it's the light blocking other parts of the light

[Jambert]

Maybe it's the light blocking other parts of the light

That's it!!! thx

The affect looks a bit strong. It's the same mesh, but light emited from part 1 seems to be higher than part two

[Jambert]

A shot with mesh divided

[Jambert]

So, my first neon model was wrong, so I tried to model it in a more realistic way.

Second Model: The shadows still looks a bit strong because model is still wrong, it needs a blended emitter. Works fine because in real life it's a gaz, but it needs very high value to compensate blend factor (here 0.9 null)

Ono, light layer for blended material is still broken, will it be fix in next release?

thx

[pixie]

I think it has to do with the shape of the light itself

[Jambert]

I think it has to do with the shape of the light itself

You and ono are right, problem come from both, shape of the light and light blocking other part of the light.

edit: any way to fix it?

[dakiru]

Why fix it? It seems like a realistic light behaviour

[CTZn]

Yes it is, Jambert. You must build deflectors around the neons ! Or, prolongate the neons into the wall or an energy supply system, that will affect the resulting emission.

A "black body" is a solid object, blocking light. Seen from the profile, the emitting surface is way smaller than seen from above (in your tests with multiple horizontal shapes).

I'm not sure about your solution wich consists of blending a null with a diffuse material... Tubes are shooting light in pretty much all directions so the diffuse transmitter is not adding much, plus the 10% null introduces strong discrepancies in the lighting solution. I'm not fan about it personally, though the way using dt only is interesting.

Multiple edits

[Jambert]

I'm not sure about your solution wich consists of blending a null with a diffuse material... Tubes are shooting light in pretty much all directions so the diffuse transmitter is not adding much, plus the 10% null introduces strong discrepancies in the lighting solution. I'm not fan about it personally, though the way using dt only is interesting.

Multiple edits

Sorry my post aren't clear enough. I didn't blend diffuse transmiter, I used only one mesh, the tube, as a blend (10% emitter, 90% null). To me it's the more realistic way.

Using a mesh with a blackbody material, as pixie, give realy bad result to me (see shadows pict.2). It's physicaly correct because mesh block some lights, but in real life neon tube are glossy transparent and lights come from the gaz inside.

[FPSatan]

I guess it's the same effect as

http://en.wikipedia.org/wiki/Binary_sta ... g_binaries

[CTZn]

Nice pick FPSatan !

Sorry my post aren't clear enough. I didn't blend diffuse transmiter, I used only one mesh, the tube, as a blend (10% emitter, 90% null). To me it's the more realistic way.

Oh ok. No problem, I have a tendency to misread posts myself

Well that's some hack anyways

I wish true volumetric emission will be available someday !

[Jambert]

I wish true volumetric emission will be available someday !

An emitting smoke? Light emmision needs normal, don't think it's possible, maybe with indigo 5.0
First step could be particule for volumetric smoke simulation