[Petero]

Jason,

ok I'll use spheres but i get the sense that the plane has less noise and might calculate quicker, and therefor be more useful in certain rendering circumstances. in my test i actually am not sure if the sphere gives more noise...i might have to let it go to a higher SL..

as far as my sphere is concerned, could i go with less facettes to go faster?

is there a minimum limit?

also would it be a good idea to have one facette facing directly downwards in my sphere's orientation?

again, many thanks for your thorough replies!

Peter

[Half Life]

That last test is interesting, does this happen with spheres?

My concern here is while testing is very useful, Maxwell has very specific parameters for how to properly use IES within the render engine and any bad results you get will be misleading because you are starting with faulty data. To come to any conclusions about whether IES works right or not while ignoring how the data was created in the first place and not understanding what an IES really is, is impossible due to your intentional misuse of the tool (planes and other geometries).

You can use low poly spheres (as I said earlier) the poly count and size of the sphere are not really terribly important (I've tried smoothed and faceted as well)...

The "shape" of the falloff will be correct (more or less) regardless of geometry, but as they even say on the lux render link you provided the intensity will not be correct and will likely exaggerate the lighting characteristics in unfavorable ways if you use the IES outside of the specified way:

However, to have the physical intensity of a specific ies scaled correctly the ies must be used with pointlights (mesh and planes will still make the correct 'shape' , but the intensity will be altered by their power, efficiency and size)

In their instance they are using pointlights (which as I said is the similar approach to how JD implemented this in Rhino), but whichever way it is implemented it is always best to follow the proper protocols so that render engine will deliver predictable results.

To me it makes no sense to start with a measurement of a real world light fixture then twist into something that it is not meant to be. The goal with Maxwell is maximum realism -- they were hesitant to implement IES in Maxwell to begin with because of the potential for people to get unrealistic results due to misuse... there is certain amount of trust here that we will use the tools properly.

Best,
Jason.

[Petero]

Jason, the notion of 'correctness' is a limited one for me since i am not creating renders to help a lighting designer spec a project. i want it to look good, be manageable, and be as fast as possible. but i am now fully forwarned of the dangers of any 'shortcut' i may use-that's very good to know. i will do more tests later regarding noise as related to varying geometries, and post results.

regards-

peter

[Half Life]

In any scenario, you are of course free to do it how you like... and knowing you are doing it "wrong" but for a particular purpose is the artists prerogative.

But you may one day find yourself with a need for the most physically accurate results -- knowing how to achieve that is important as well.

BTW published scientific studies are often torn apart by peers in the community due to poor testing methodology... I have had my testing methodologies and presuppositions destroyed in this forum as well -- which is a good thing as it keeps me honest

(I forget to use the emoticons sometimes, my typing can come across poorly as I am usually multitasking)

I am very interested in this noise issue -- and am eagerly looking forward to seeing your future posts on that because you may very well be onto something important there. What may be interesting is arranging the IES in different patterns to see if it always goes from first to last or if there is some other reason (like distance from camera) that is the cause.

Speed is one of those Maxwell issues that rarely enters into my consciousness , but I do understand how that is a big concern of others... for example I loathe AGS because it undermines the point of using Maxwell to me -- but at the same time I fully realize its utility for most users. It may be your experiments lead to an "AGS of IES" type result which I'm sure many users would appreciate

Best,
Jason.

[Half Life]

So I did a test of my own -- this one for polygon count and smoothing to verify what I has said earlier (it had been a few patches back since I ran a test).

There are 18 spheres at a radius of 1cm each -- the poly count is as follows.

From left to right (in longitude by latitude count):

1. 48 x 48 smooth
2. 48 x 48 faceted
3. 24 x 24 smooth
4. 24 x 24 faceted
5. 12 x 12 smooth
6. 12 x 12 faceted
7. 6 x 6 smooth
8. 6 x 6 faceted
9. 180 x 180 smooth

for the second half I repeated the first set in the same order.

Here's the MXS file if you'd like to use it... make sure you turn on devignetting at 100% to see accurate results from edge-to-edge with these types of tests.



To me these look (more or less) identical in noise, detail, and intensity... I rendered to SL 14 and I saw no difference at any point in the process.

Best,
Jason.

[Petero]

Jason,

cool test. i am away from computer today but will try some test versions of my own this week. i have old, slow computers, so i really have to keep an eye on render times.

i have never noticed the devignetting checkbox...

cheers-

peter

[Petero]

Hi Jason,

After a long hiatus I am back at my IES research. Working with small spheres as you suggested, but cannot get them to point anywhere but down...looking around in my 3D software for the answer(i work in Form Z).

I have a huge rendering project with lots of eitters and reflections that will also require an investment in a new machine for the Maxwell renderings...

http://www.formz.com/forum/discus41/mes ... 1308252074

best regards-

Peter

[pylon]

Hi Peter,

Sorry I didn't see this thread earlier, I normally only check the formZ and bonzai3d forums.

Unfortunately, the trick with rotating the IES-mapped sphere you describe does not work with formZ 6.

In bonzai3d (and the forthcoming formZ 7) there is an IES light type that I have added. You don't have to make a sphere, this is handled automatically. You can point the IES wherever you want, spin it on its axis, view the intensity distribution, etc.

http://www.formz.com/forum/discus41/mes ... 46004.html

Ben

[Petero]

Hi,

I have successfully followed Pylon's instructions, and - using a trial version of BONZAI - have introduced IES lights which I can now easily control. I create an MXS from this which i then import into the MXS created by my FormZ file, which works perfectly.

The only minor hitch: Bonzai doesn't seem to have the "Composite" choice in multilight so that it creates a slider for each emitter object. I simply delete the repetitive emitter materials, leaving only one, and reassign this material to all the lights.

thanks-

Peter