Maxwell Render | Official public forum

Okay, so I know nothing about writing scripts or what exactly all those numbers in an IES file mean, but would there be a way to create geometry from the IES file? From what I do understand the IES file is a table of points in 3 dimensional space and the lux value at that point. So wouldn't it be possible to create geometry and use the same points and use the weights of the lux values to figure out where it should go? I know it would kind of be backwards from the IES file because where the IES shows a spike you would want a flat point so more light is cast that direction but does anybody follow what I'm saying?

I only ask because in my latest Maxwell fight with my boxx he brought up how Maxwell can't use IES and that clients want to know what their lights are going to look like.

I do see alot of potential problems with this, like how to make the actual fixture itself light up correctly, but it could be a good starting point for getting IES into Maxwell.

Hey Misterasset,

I can definitely understand your boss' point about clients wanting to see exactly how their lights will 'cast.'

From what I've found you can achieve IES like scatters but this is strongly dependant upon two things: 1. the geometry of the fixture itself. 2. the physical 'correctness' of the materials and emitter plane.

IES is pretty much a convenience that we've all been spoiled by. Don't get me wrong it's definitely a time saver because, who really wants to model the actual fixture when, in the case of a recess can, it will never be seen. However, IES data, is a bit of a cheat per say.

Here's a great tutorial on the evermotion on this very thing:

http://www.evermotion.org/index.php?unf ... =exclusive

As you'll see, you can essentially create IES 'effects' by modelling the fixture and applying the proper, physically correct, materials (I haven't visited this tutorial in a bit so I'm not entirely sure how physically correct the materials are given the release of 1.1)

I'd love to see IES incorporated into maxwellrender. For now, this is the only way I know how to incorporate the effect into our physically correct renderer

Sorry for the longwindedness (and the spelling )

Luis

Yeah, I've seen that before and I have to agree with Mahaiuai (we all know who I'm talking about) that it would take forever to clear up if each light lit the room through a dielectric.

Does anybody know if that's the quality of Evermotion's models? If so it might be worth investigating.

But back on topic, does anybody thing IES Geometry would work?

lsega77 wrote:Hey Misterasset,

I can definitely understand your boss' point about clients wanting to see exactly how their lights will 'cast.'

From what I've found you can achieve IES like scatters but this is strongly dependant upon two things: 1. the geometry of the fixture itself. 2. the physical 'correctness' of the materials and emitter plane.

IES is pretty much a convenience that we've all been spoiled by. Don't get me wrong it's definitely a time saver because, who really wants to model the actual fixture when, in the case of a recess can, it will never be seen. However, IES data, is a bit of a cheat per say.

Here's a great tutorial on the evermotion on this very thing:

http://www.evermotion.org/index.php?unf ... =exclusive

As you'll see, you can essentially create IES 'effects' by modelling the fixture and applying the proper, physically correct, materials (I haven't visited this tutorial in a bit so I'm not entirely sure how physically correct the materials are given the release of 1.1)

I'd love to see IES incorporated into maxwellrender. For now, this is the only way I know how to incorporate the effect into our physically correct renderer

Sorry for the longwindedness (and the spelling )

Luis

Excuse me but how are IES files cheats when they are based on real world measurements? This is about as real as you can get.


Marcus

marcdevon wrote:Excuse me but how are IES files cheats when they are based on real world measurements? This is about as real as you can get.

Because they actually aren't true until a couple of feet out from the light source. It's great for pendant lights and fluorescent 2x4s and so forth, but wall washers and sconces aren't truly perfect.

But I will admit that it is close enough.

I don't know how it works or how accurate it is, but Max and several other programs which work with IES files can create a "light cone" based on the IES file data. Back when I was still using Radiosity in Max it was usefull, for example, to tell a spot light from a flood light by simple looking at the IES light cones from each emitter.

The cones look like a 3D version of the polar photometric charts that fixture cut sheets come with.

I mean, often .ies profiles take into account the bulb itself & the luminaire fixture, which eliminates the need to calculate irrandiance via caustics calculation.

So, technically speaking, if you model all of the geometry, source+bulb+fixture (reflectors gels etc) you should get really really accurate results, more so than an .ies file, which considers some tolerances for practical considerations. On the downside, you could turn 50 before getting a clean render in complex interior scenes.

In some cases maybe you could literally model an ies profile to simulate light distribution, might be worth a try, especially for complex scenes (in order to avoid long caustic calcs), but i doubt it would be really accurate as it wouldn't take into account caustics...