Lighting and Scale

[1armedScissor]

Preferences->settings->working units->Linear

There is a drop down menu, change it to something more appropriate for the scale that you're working at.

Also what are the intensities of the quadratic lights you're using?


Can you post a sample scene?

[drossxyu]

Ok, but changing my working units to something else after my scene has already been modeled doesn't effect the lighting whatsoever. I've always kept it at centimeter because I remember reading somewhere that mental ray was more predictable (with GI + FG) when treating each unit (1 cm) as 1 ft. The intensity of my spot lights w\ quadratic have typically been between 50-100. I'll try to post a scene tommorow..

[drossxyu]

So one thing I discovered that could help my situation is to use the 'intensity curve' option in the light effects tab. It allows you to basically make a custom graph of the intensity, which pretty much allows you to control falloff. However, this option is only available with spotlights. I'd like to be able to also do this with omni lights.

[1armedScissor]

You can use this options on pointLights as well but there's a trick to it. First create a spotLight, then set it to have an intensity curve. In the attribute editor, change the light type from "spotLight" to "pointLight" and you will now have an intensity curve to control your light's falloff

[3DsIn]

Keeping everything to scale is important.. Why not just scale your scene up? I usually just keep the standard metric units and do the simple conversions.. (1inch = 2.54cm) Working like this is going to require you to scale your scene way up so keep a couple of these things in mind..

a) Set the far clip on your camera to an approiate value
b) Setting the locator scale up on your light will make it bigger.. (so you can see it)

As far as lighting, totally depends on your renderer.. Average values for a physical light in mental ray are usually up in the thousands (degrees kelvin)

-Ryan

[drossxyu]

So I did some experimenting with scaling my scene up. My working units are set to 'centimeter', and I scaled my room up so that it measures 141.7 centimeters high ( approx. 4.6 ft). So even at this scale, which is probably half the height of a normal room, I had to set my photon intensity to 900,000,000 to really see anything start to bounce around. Is this common? I know you mentioned setting a physical light into the thousands, but this seems a little bit ridiculous.

-f

[3DsIn]

Your photon intensity does not need to be set that high. Keep it around 8000 and try tweaking the exponent, lower it to 1.2 and see what happens.. Light intensities are measured in degrees kelvin. Jeremy Birn has documented all of this:

http://3drender.com/glossary/colortemp_F.htm

All of the lighting parameters and global illumination techniques have been documented extensively over the years in this forum.

-Ryan

[DrYo]

b) Setting the locator scale up on your light will make it bigger.. (so you can see it)



Oh man, three years of working with Maya, and I never knew that attribute existed. Thanks!

Now, if Autodesk would only add a constant screen size option....

[achoury]

... Is there a way around scaling the entire scene?
-f
yes you can, just select all the objects in your scenne, the fastest way to do that is to go in the Menu of input line operations and hit star in your keyboard then hit ctrl +G to group all objects together than you can exploit the scale on the whole of all objects of the scenne at one time.
sorry for my english
rachid

[drossxyu]

Excuse me if this is completely off , but does that then mean that 'Photon Intensity' is measured in degrees kelvin and it should be set to the appropriate color temperature? I've seen so many other examples and tutorials that have people bumping the photon intensity up to tremendous values. I've never thought to use the exponent to multiply/divide the value, so I'll give that a shot.

[Olegr]

If you change the exponent on photons they will no longer behave in a physically correct manner.

[drossxyu]

So back to that intensity curve..

It seems as if the MIA material disregards this. In other words... a spot light that should terminate before it hits a wall remains as expected if the wall has a lambert, but acts if it's being hit at full blast with an MIA material. Any ideas?

[Kako]

So I did some experimenting with scaling my scene up. My working units are set to 'centimeter', and I scaled my room up so that it measures 141.7 centimeters high ( approx. 4.6 ft). So even at this scale, which is probably half the height of a normal room, I had to set my photon intensity to 900,000,000 to really see anything start to bounce around. Is this common? I know you mentioned setting a physical light into the thousands, but this seems a little bit ridiculous.

-f

If you scale your scene by a factor of x, your light intensity and photon intensity should be scaled by x^2 to yield the same result. (Considering you're using quadratic decay for both direct and indirect illumination.)
So I'd say that the high photon intensity in your test wouldn't be that odd. Imagine if you scale your scene back down by 100. Your room would measure 1.417 unit high (Let's think in units... Forget that's centimeters), and your photon intensity would have to be divided by 10000 (100^2), thus 9000, which is kinda normal.
Now... Why aren't you having the fine control you're looking for? Is it because the light intensity is lower than 0.001? Lowering the light intensity and the photon intensity should give you the same result as scaling the scene up, shouldn't it? :curious:
Also, in this scaling transformations, remember to scale the photon radius by x (not squared) to keep the GI quality.

Kako.

[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.