Question about Luminous Flux

Deke

I can't get the hang of Iray and setting lights. In one seen I have a linear point line with lumenous flux cranked to 200,000.  Seems to work fine and cast a nice fill light. Then I have a distant light as the main light source. It has a luminous flux of 75 and is baking the scene...way too much light. I'm fussing with tone mapping trying to modulate things, but why is this so tricky? Is Luminous Flux the same as the lumens the light is spitting out or is there some other way to modulate that?

Richard Haseltine

Neither distant lights nor linear point lights will work well in Iray - linear pointlights in particular are not real (the light falls off linearly with distance from the source, instead of the correct fall off with the square of distance). Try using regular spot and point lights and set them to Photometric mode.

Deke

These are linear points lights with the photometric option activated. So I assume that makes them Iray friendly?

mjc1016

Deke said:

These are linear points lights with the photometric option activated. So I assume that makes them Iray friendly?

Sort of...they are still 'not real'.  They will still have linear falloff. 

Deke

So a linear point light can never be an Iray-style light with real-world fall off, but spot or point light (with photometrics activated) is an Iray-friendly light?  That still doesn't explain to me why one light requires a huge number of lumens and the other doesn't, but then I'm not sure luminous flux is actually about lumens output or something else.

mjc1016

Correct on the light types.

Because luminous flux is a real world measurement...and if the light doesn't behave in a real manner, then all bets are off as to how it will actually perform, when rendered.

Tobor

True about linear point lights, but distant light is designed like the sun  - to Iray, the sun and distant light is the same. Iray treats distant light (and the sun) as parallel rays from an "infinite" source. Like in the real world, parallel rays do not behave in inverse square; that's a property of spreading, uncollimated light; So Iray's interpretation of this light is physically correct.

Specific light fixtures are going to have different profiles, and their light is emitted differently. The same luminosity value will have different effects. Specify 1000 lumens for a point light, and those 1000 lumens are spread in all directions, with onkly a portion falling on the scene. So, 1000 lumens is not very bright. The same 1000 lumens in a spotlight looks brighter, because the same light energy is focused to a smaller area.

From appearances, distant lights work differently, and not because they are "un-physical." Rather than measure light intensity at the source, which doesn't really make sense for an "infinite" parallel-ray light, it very much seems that distant lights are measured as the light that falls evenly on the scene. For this reason, only low illumination is required when using distant lights. 

Deke

Thanks for the explanation. I've found I can control a distant light and render as "scene only" to achieve nice directional lighting. Using the Sun Dial setup and rendering Sun and Sky is perhaps better for exteriors.

mjc1016

Tobor said:

True about linear point lights, but distant light is designed like the sun  - to Iray, the sun and distant light is the same. Iray treats distant light (and the sun) as parallel rays from an "infinite" source. Like in the real world, parallel rays do not behave in inverse square; that's a property of spreading, uncollimated light; So Iray's interpretation of this light is physically correct.

Specific light fixtures are going to have different profiles, and their light is emitted differently. The same luminosity value will have different effects. Specify 1000 lumens for a point light, and those 1000 lumens are spread in all directions, with onkly a portion falling on the scene. So, 1000 lumens is not very bright. The same 1000 lumens in a spotlight looks brighter, because the same light energy is focused to a smaller area.

From appearances, distant lights work differently, and not because they are "un-physical." Rather than measure light intensity at the source, which doesn't really make sense for an "infinite" parallel-ray light, it very much seems that distant lights are measured as the light that falls evenly on the scene. For this reason, only low illumination is required when using distant lights. 

Now whether or not Iray should be doing that with a distant light, is another question.  And it comes down to how the light was assigned/categorized when being translated.

Tobor

Not sure Iray can handle light from an "infinite" source any other way. Light power is a function of the geometry of the output, not just intensity. Whether an infiite light has one ray or a billion, the rays are parallel -- effectively one ray that covers an entire scene -- or city, or planet, or whatever. Calculating the light falling on the scene, which is what photographers routinely do, seems the only plausible way to handle the situation. I dson't know how the lumens measurement is applied, but it may be per cm^2 or m^2. That seems in keeping with the way other parts of the program work.

As I noted in another thread, the infinite light source is a built-in feature of Iray. D|S doesn't really have to do much to implement it in an Iray fashion; just define the light type as LIGHT_INFINITE, set the direction and luminosity, and it's done.