[AlexP]

For emitters, to show that light is emitted more efficiently to front and less to sides or vice versa simulating e.g. semitransparent glass or paper cover without affecting performance.

[bograt]

I have wondered about this for a while, like a roughness for emitters. Obviously more relevant for planes than spheres

[polynurb]

I have wondered about this for a while

me too.

using .ies allows for this, but some kind of interface to tweak such behaviour in real time would be great.

[Bubbaloo]

You can make your own using gradients and apply it as 32 bit exr/hdr/mxi based emitter.

[AlexP]

Still gradient is fake solution and for spheres and cylinders unusable IHMO... Gradient doesn't know where camera is right?

[bograt]

You can make your own using gradients and apply it as 32 bit exr/hdr/mxi based emitter.

What I meant (and I'm pretty sure Alex meant) is to be able to control the distribution of light from an emissive object, at the moment light is emitted evenly in all directions.
An emissive plane acts like a highly diffused softbox, it would be useful to be able to control the level of diffusion to simulate many different lighting scenarios (light enclosures, reflectors , how light enters a window etc.)
It is possible to create these scenarios realistically in Maxwell but it is way too slow.



You can use grids to achieve a similar effect but its not ideal

[AlexP]

Thats exactly it, thanx alot bograt!
Very rare example for this is LCD panel when is very dark, it spreads light to sides, not to front.

[Gary Bidwell]

+1. Pixar has this option in recent PRMan release.

[dariolanza]

Hi,

I guess that what you are referring is already possible and could be achieved with an IES file.

In fact, an IES file contains the description of the light lobe itself, being it very directional, widely spread, etc.

I know there are several IES generator applications, that you can use to directly paint the lobe itself with your cursor, and the application writes the numbers in the IES file.

Then simply apply the file to an emitter to get that exact light emission you designed.

Cheers

Dario Lanza

[AlexP]

But is e.g. large ceiling L-shaped light possible with IES? I mean IES is mapped from sphere right?

[Gary Bidwell]

I think something simple, like PRMan and Vray use for Area lights.

Make the light emit more parallel in the direction of the surface normals of the emitter. 0 default emission to 1 perfectly parallel emission.

Cosine power:
specifies a power for cosine power. Larger values make the light emit more strongly in the direction specified - giving a focusing effect. A Value of 1 is default and specifies that the light contribution is simply due to the form factor of a portion of the light, and is not further focused.

[AlexP]

That's it!
BTW didn't know vray has is, can't find it in its docs...

[Gary Bidwell]

http://vimeo.com/16905038

Check at 55 seconds.

[AlexP]

Indeed, maybe that feature is not enabled...

[bograt]

http://vimeo.com/16905038

Check at 55 seconds.

Bingo!