real-life lighting tools and techniques

[krembo99]

Well..
The way I see it, your experiments are a bit wrong..
You are using VrayLight (as I understood from the lights in the picture)
The Vray Light is not (as i understand - and maybe i'm wrong) a 100% accurate as to how light works. It's a raw emmiting light source.
In real world, the way light is being emmited is influenced by a lot of parameters, that are not reproductable with flat Vraylight.
One of them is the web which you mentioned (IES and the vector of light particles ) but others can be also the emmiter shape, it's material, the sorounding materials etc.
That is why , In my opinion, putting a Vraylight, that emmits in a 90 degrees angle, and uniformly, produces similar results in different situations.
Don't get me wrong, I'm a vray fan myself.. but the real force of vray is in SIMULATING real world, and not IMMITATING it.
Anyhow, I do think your topic is very interesting, but maybe you should also take into consideration the calculation method itself (what did you use here ??)
I belive that non-biased brute force will produce better results in terms of accuracy and similance to the real world. (and Vray has that as well if i am not wrong in the light cash non biased solution)
But these are just my thoughts .. :-)

[mister3d]

Thank you for reply, I appreciate it.
I used vraylight, because I thought vraylight is the most accurate in vray.
Is your point that there is a more accurate light source for vray? I thought that:
standard lights - simple, can have any decay
photometric lights - the same algorithm, but with inverse-square decay and more advanced color adjustments
vray light - area light for vray with inverse-square decay, produces the most accurate shadows(and GI), so that's why it is used as a skylight portal(the same as in new mental ray) - to produce smooth photon\irradiance\lightmap distribution.

Yes, my lights in test lack IES distribution parameters, simply because I don't know yet which IES lights to use for them. Unfortunately I'm not familliar with real-life lamps used for artistical lighting.
About emitter shape, I think IES light should be enough.About material of light, maybe it could be ignored due to a very high intensity, and IES actually shows caustic distribution of this material.
If to talk about surrounding materials, there are mirror-like surfaces in lamps, but to use them to produce caustic reflection is crazyness :) , will take ages(and I didn't reach anything serious with vray caustics for large area distribution, maybe it's just me).
I will try QMC method for GI, as you reccomended, thank you.
So IMHO I could try find IES lights for these and use QMC for more accurate calculations.

Thank you once again, Krembo99, I would love to see some results from these experiments and techiques in future.

[krembo99]

The way I see it , one of two things happened :

Either I did not fully understood your question (which is probable :) )
2. you have mixed up a little the technical data/info/terms in your head, or they are not fully clear to you...

There are too many things I do not understand in your last post , I will try to list some.

I used vraylight, because I thought vraylight is the most accurate in vray.

well.. depends for what purpuse you need it . it's a simulation like I said.

vray light - area light for vray with inverse-square decay, produces the most accurate shadows(and GI), so that's why it is used as a skylight portal(the same as in new mental ray) - to produce smooth photon\irradiance\lightmap distribution.
But in this experiment you are trying to produce light emmiters, and not skylight or sun, am I wrong ? what does skylight portal has to do with a sodium lamp for example ?

Yes, my lights in test lack IES distribution parameters, simply because I don't know yet which IES lights to use for them. Unfortunately I'm not familliar with real-life lamps used for artistical lighting.
Try to go here : http://www.bega.de/sprachwahl.htm browse the catalog, You will find some usefull things there :thumbsup:

About emitter shape, I think IES light should be enough.
yes, It might, But you did not used it here, right ?

About material of light, maybe it could be ignored due to a very high intensity, and IES actually shows caustic distribution of this material.
That is , IMHO , wrong .
Caustic ? in your experiment, as i saw it, there was no possiable caustic producing elements... and anyhow, IES WEBS not always take that into considoration..

If to talk about surrounding materials, there are mirror-like surfaces in lamps, but to use them to produce caustic reflection is crazyness :) , will take ages(and I didn't reach anything serious with vray caustics for large area distribution, maybe it's just me).
Again caustics :)
in optics, caustics are caused by ray reflection or refraction by a curved surface or object.
CURVED...
if i am not wrong, there was no such surface in your experiment. and in mirror-like surfaces in lamps, not always you have them . depends on the lamp. but like before, IES web could treat this issue IF USED !..

I will try QMC method for GI, as you reccomended, thank you.
I did not say to use QMC (I do not say NOT to use it either ) , I just said Vray has a non biased solution under the LIGHT CACHE options, not QMC...
But again, this is when you want to reach phisical accuracy.

As I said in the begining of this post, I am not sure I fully understood the purpuse of your experiment, after reading your second reply.
Using normal Vray light with no IES , no shapes or deflectors, and with no control over the GI method, will not produce accuracy.
But like i said, Is accuracy really what you need ? or simulating it to produce nice VIZ is what your'e after ?

[mister3d]

"Well.. depends for what purpuse you need it . it's a simulation like I said."
Could you please clarify what you mean? If I need a point light, I use a point light, if I need an area light, I use vraylight. I also did not understend the difference between simulation and imitation. :banghead: :)
In this test I am trying to produce light emitters and blockers, so the snoot, the eggcrate and the skirt are blockers, and the purpose was to see the difference between them and standard means. I used vraylight in these 3 cases because they(snoot etc) actually block large surface lights.
I mentioned about skulight portal meaning that for high-quality shadows the vraylight is used as opposed to skylight.
About caustics, or GI reflection, thank you for correcting me, I still can't imagine using it in the test because it's very slow and not smooth.
And what is a non-biased solution(for GI?) in vray? I never heard about it.
"Using normal Vray light with no IES , no shapes or deflectors, and with no control over the GI method, will not produce accuracy." I agree once again I should use IES lights, and thanks for the link , but are there fixtures for decoration but not for professional lighting? About the shape, how do you propose to improve it? I jusr thought what blocks light will contribute the most, so is enough.
Anw what do you mean "no control over GI method"?

[krembo99]

before I will unswer the questions, I would like to understand one thing.
Do you think that your experiment was successful in terms of what you expected from it ?

[mister3d]

before I will unswer the questions, I would like to understand one thing.
Do you think that your experiment was successful in terms of what you expected from it ?

krembo99, no, it's just a try. Making it successful requires some very knowligeable people, who are familiar with lighting in real life too. But for sure I will use these techniques on occasion.

[krembo99]

Ok, so if you see yourself the experiment was not so successful, you must now think and understand why...
I will take ONLY the eggcrate as an example to explain to you why that happened with VRAY light.
You see - Like i told you before , VRAY light spread it's light (as far as I know) in a perpendicular way, that means 90° , or the vector of it's normal. that means that in YOUR example, the light goes ONLY down, in 90°, and than has a decay of some sort.
A normal light bulb, no matter of which kind, will never be able to emmit light ONLY in 90° down. I know no such device that can emit light only in ONE direction , except laser (and if such device exist, I would love to know it..)
So in your eggcrate example, the light actually almost "ignores" the geometry you crated, and goes down in a normal way. the light does not "hit" the inner faces, it slides by in a manner (i am exagerating in order to explain). That is why there is virtually no difference.
A normal light emmiter, in real life, would have spread it's light in other directions as well, thus "hitting" the inner faces as well, creating reflections, and changing like that the casting geometry.
This is why a plane Vray light is not so good for this experiment.. It has also a special decay/spreading method (45° if i am not wrong), so you will also notice different results by placing it in different hights above or inside your geometry.

Now about your questions,

The difference between simulating and immitating is semantic, but what I meant is the difference between brute force phisically acurate non biased render method and a biased method (of any kind) that is based on simulation and thoughtful guesses of any kind (like IRR map)
One is calculating everything, one is calculating a part, and based on that part guess the other part. that is simulating in a way.

This is the main difference between non-biased renders like maxwell (first versions) and fryRender for example, as oposed to Brazil, Vray or fRender that each uses one method or another of simulation and guessing to speed things up.

About the non biased solution in Vray, as far as I understand, would be the light cache, both primary and secondary,set to progressive path tracing and sample size set to value of 0.0. As far as I know - that is some kind of a non-biased solution. If I am wrong, please correct me.

About the shape, have you tried to see what happenes when you use the light material with a different shape in the eggcrate ? I do not know if it helps, becasue the light material has little control, but maybe it is worth a try.

and last, what i ment by "no control over GI method" is that when you did the experiment , you have ignored the GI method you used. like i said before, different methods use different "guessing" mechanisems, and from my experiance, produce different results.
As a factor , it was ignored in your experiment.

[arza]

ok i agree...with mr Krembo...here, about alla these kind of differences between, immitating and simulating real light. but i would like to add,(and correct me if i am wrong) about the actual surfaces...as you might know...the only...almost perfect flat surface is a mirror (that's is why it is reflecting straight rays- it has no measurable bump). In 3d all surfaces are actually flat (what i mean is that den all behave in the reflection of rays lights,almost like mirror). so my question is? doesn't 3d geometry affects the light simulation in rendering?
i must confese that i have tried experiments with displacemented geometry...

well...as i said before...i am not quite sure, if what i am saying has a scientific base.

question....about vray lights? do they support IES data, or you just use the photometric light s of 3dstudio max?

[krembo99]

the only...almost perfect flat surface is a mirror (that's is why it is reflecting straight rays- it has no measurable bump).
Well... As far as I know, it's a bit wrong. In order for a mirror to work, there is one more condition that must be applied, and that is a metalic surface of some kind.
One reason for the direct reflection IS the smooth surface, but without a mettalic film/coating etc applied, the object will not be a mirror. For example you can look at a glass or marble which can be very smooth, but still will not have the reflection of a mirror.
Try taking a mirror and peeling off the metal coating - you will loose the reflection.
I do not want to get into the phisics of WHY it happens, but conductivity of metal is important to the process of reflection.
In 3d all surfaces are actually flat (what i mean is that den all behave in the reflection of rays lights,almost like mirror).
well... the surfaces are flat in 3D apps, but if you uderstood what I wrote before , you would understand that :
1. Flat does not mean 100% 90° reflection.
2. The SHADERS come to game here, and simulate the refletion amount,direction,kind etc of the surface, making null (so to speak....) the actuall geometry smoothnes.
so my question is? doesn't 3d geometry affects the light simulation in rendering?
Well.. my first instinct would be to tell you "of course that 3D Geometry affects light simulation in renderings" .
But thinking again of your question, I must confess I do not know.
This question must be answered by those who program the render engines, and the algorithems inside.
I must say that I ASSUME that the geometry has little or no effect on the light, because, like i mentioned before, the SHADERS take this job of simulating light interactions, making null the geometry (to some extend of course). but again, I hope someone more proficient than me in light calculation algorithems inside render engines could answer for that.
question....about vray lights? do they support IES data, or you just use the photometric light s of 3dstudio max?
About the Vray light and IES - I can only answer of versions prior to 1.5 , Yes, you need to use normal light and just set the shadow type to Vray shadow. Vray light itself does not support IES webs .

[mister3d]

Thank you krembo99 and arza, I agree with you and will try to update the test when I have a bit more free time.

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.