[feynman]

Hi,

my first attempts at IES lighting does not yield the results I was expecting. I think I did all what is suggested in the documentation regarding IES lighting (no other light sources, sphere, correct IES file, all objects at correct scale, etc.). I want to simulate the effects of various diffuser geometries which are lit with a single 6W high performance LED. But the images turn out way too dark. I have the real product here and a single 490lm LED pointing upwards into a diffuser lights it well and also a white matt surface 2m below.

If I keep the IES intensity to 1,0 which one should, this is the dim result



If I change the IES intensity to 100,0 which one should not, I get reasonable illumination, but I don't want to trust the effect anymore



Any hints very much appreciated!

[jomaga]

If I understand correctly your scene from the ies shape, the emitter seem to be pointing upwards. Try to rotate 180º the sphere till the IES cloud point downwards

[feynman]

Well, that's exactly the point - the LEDs point upwards into diffusers so as to be hidden from direct view (no glare). It works with the real prototypes, but I was thinking it should look right when simulated with Maxwell Render, too. I want to test different diffusers to see the spread of reflected light on surfaces below (floors, tables, etc.).

[brodie_geers]

An IES typically isn't just data about how a bulb emits light. It takes into account a particular diffuser already. Are you certain that this IES data you have is correct for what you're simulating?

Also, are you matching the Maxwell camera parameters with your real life camera parameters?

-Brodie

[feynman]

Camera, scale, etc. should be alright; checked it again. On CITIZEN, SHARP or NICHIA websites, there is no mention of a diffuser being taken into account. That would actually defy the purpose of IES, wouldn't it? You can never know what light the designer will use the lamp for; any lamp - whether CFL, HL or LED could be put into any light, into any diffuser. Otherwise, manufacturers would supply data of the diffusor used to arrive at the IES data? Lamp manufacturers state that "The IES data format is an internationally accepted data format used for describing the light distribution of luminaires". So, it's just the luminaire that gets measured.

But if IES don't work, maybe I can simulate the LEDs using their illuminated surface (4-8 polygons) and the output in lumens to get results that are close to reality...

[brodie_geers]

Can you post the site where you got the ies data?

I'm mostly familiar with Erco who does indeed provide the fixture data. If you go here for example ( http://www.erco.com/products/indoor/swf ... ownl_1.php ) you'll see that you can download a .zip with a bunch of IES files and a corresponding .zip with a bunch of 3ds Max files which contain the fixture associated with the various lights.

However, I don't see any reason why IES data couldn't be gathered without the use of a fixture (not that I'm terribly familiar with the equipment used to gather the data) so if the site doesn't mention a specific fixture it seems reasonable to assume it is indeed the raw bulb data.

Maybe you could do some more controlled tests to verify? For example, how about pointing the LED down and not using a diffuser and comparing the results to try and hone down where the error is?

-Brodie

[feynman]

Here http://ce.citizen.co.jp/lighting_led/en ... ution.html you would find IES data for CITIZEN high-performance LEDs. I made two more tests:

The 3D-file of the LED in question with an emitter material (490lm output, 3500K colour temperature) assigned to the emitting plane (a few polygons on top)


The same LED represented via its IES data (5000K colour temperature)


Both viewport renderings actually look very similar, apart from the colour temperature difference. In a blacked out measuring room, however, the real LED's output reflected off the exact same matt white diffuser looks much brighter.

So, that means that something is simply not set correctly with the camera. Need to check that out next.

Ah, maybe the EV was just not small enough; more like what one can see with one's eyes.


Now I just have to create porcelain as material, because of the translucency a cheese material with SSS would be a good idea to begin with :)

[brodie_geers]

Are you comparing Maxwell's output with what you're seeing visually, or with the output from a real camera? To make it apples to apples you HAVE to use a real camera with similar settings.

-Brodie

[feynman]

I'm comparing it to a similar photograph which was taken in a perfectly dark room with a digital camera from which I have EXIF data, since the lights I'm simulating don't exist yet outside of 3D space :)