Inverse Square Law and window-light.

[cjbroadbent]

The ISL is my enemy and I fight back with:
Fresnels with scrim flags,
Neutral density wedges worn sideways,
Very large window-lights with monstrous jagged-edge foam panel gobbos.
Window-size beehive grids,
Backgrounds (and foregrounds) painted dark on the near-side and light on the far-side.
Black flags, Barn-doors, cardboard and anything available that might gradually feather off the light without screwing the shadows (shadows get screwed whenever the flag or barn-door happens parallel to the far side of the window.

That's why round sources or sources with central hotspot are so much kinder to noses, boxes or anything else, for that matter).
In fact, an umbrella with the lamp stuck well inside to avoid spill, and pointed across the subject does a better job with the inverse square law. It is round, has a hotspot (better penumbra and brilliance) and the edges of the brolly feather nicely all by themselves.

But here is the question.
Using a real window and a clear sky. There should be no Inverse Square Law dropping a stop or two at between nine and twelve feet from the window. The sky is at infinity. But it does drop. Like when the window is covered with a sheet of frost. Why?

[Jimi]

The light coming in through the window from the outside is just a source of light, and the further you move from that source, the less light you get. No escape.

No, I can't explain it better than that, sorry.

[jp]

You are much better at lighting things than most of us; hopefully I'll be understood as just sharing my understanding rather than teaching.

Looking out my window I see more trees and branches than sky. So the trees are reflecting and filtering the light coming in. Someone else might be half lawn, half sky. If you have a sunbeam coming in, that is of course coming from practically infinity and will not exhibit ISL. I would pretend the normal diffuse illumination from the window is a big rectangular softbox with a light in it; a natural rembrandt lighting source.

[Policar]

The ISL is my enemy and I fight back with:
Fresnels with scrim flags,
Neutral density wedges worn sideways,
Very large window-lights with monstrous jagged-edge foam panel gobbos.
Window-size beehive grids,
Backgrounds (and foregrounds) painted dark on the near-side and light on the far-side.
Black flags, Barn-doors, cardboard and anything available that might gradually feather off the light without screwing the shadows (shadows get screwed whenever the flag or barn-door happens parallel to the far side of the window.

That's why round sources or sources with central hotspot are so much kinder to noses, boxes or anything else, for that matter).
In fact, an umbrella with the lamp stuck well inside to avoid spill, and pointed across the subject does a better job with the inverse square law. It is round, has a hotspot (better penumbra and brilliance) and the edges of the brolly feather nicely all by themselves.

But here is the question.
Using a real window and a clear sky. There should be no Inverse Square Law dropping a stop or two at between nine and twelve feet from the window. The sky is at infinity. But it does drop. Like when the window is covered with a sheet of frost. Why?

Hard light coming through a window does not suffer from ISL (well, except to negligible extent that across the room is farther from the sun than less far across the room is). Soft light through windows is 100% bounced light and suffers from ISL the same way bouncing any other source with a soft bounce card (white side, not mirror) would. I'm not sure what the physics are here and it's probably not strict ISL, but this is what's happening.

[ic-racer]

The closer the subject is to the window, the larger the angle of view out the window, which allows more light rays to hit the subject. As you move back, the angle of view diminishes. Eventually you will be so far back that you have only collimated rays from the window hitting the subject. These few rays are going to be only a very small subset of all the angled rays hitting the subject when the subject was close to the window, and therefore dimmer.

[Policar]

The closer the subject is to the window, the larger the angle of view out the window, which allows more light rays to hit the subject. As you move back, the angle of view diminishes. Eventually you will be so far back that you have only collimated rays from the window hitting the subject. These few rays are going to be only a very small subset of all the angled rays hitting the subject when the subject was close to the window, and therefore dimmer.

I believe this is it; it's an issue with fast fall-off from all soft sources due to the varying angles of incidence, not ISL.

[cjbroadbent]

I believe this is it; it's an issue with fast fall-off from all soft sources due to the varying angles of incidence, not ISL.

That is, Ic-Racer has it right, the nearest side gets more sky. So the further the subjects are from the window the less difference there will be. Just like Inverse Square Law unfortunately.

[Jim Michael]

Under some circumstances (e.g. w/fresnels) you may calculate the position of a virtual point source placed at a different distance than the actual source in order to see a result that maps distance in f-stop feet to falloff in f-stops.

[Tobias Key]

Don't be fooled with all this the sun is 93 million miles from the earth crap. The window is an aperture and as such behaves in the same way as your camera aperture or any point source of light does - the position of the aperture is your light source and all calculations start from that point not not the position of the sun. Diffraction works from that aperture or window in the same way it works in a camera lens, otherwise you'd never get any light fall off from any daylight scene shot at close focus, after all what's an extra couple of inches when the sun is 93 million miles away?

As an experiment take light meter readings from a window at doubling distances, the inverse square law will apply exactly the same as if the window was a flashgun set at that spot.

[jim kitchen]

Ladies and Gentlemen,

I mention ladies if they decide to contribute or have contributed, but from what I remember about my practical physics in university and any subsequent light wave propagation, diffraction propagation or reflected light studies that I participated in while studying architecture, happens to be that I was taught that the Inverse Square Law did not have a corollary, unless a new science surfaced in the last few years that now indicates how the projected surface area measured from any light source type does not change with distance from the light source. I then believe that your hypotheses, your personal modifications to the ISL, or your interim interpretations of the ISL, whether it is a window, a door and, or a reflected source that allows a light source to strike the subject differently than what ISL emphatically states are absolutely incorrect.

To my knowledge, ISL does not have a corollary, no matter what the light source happens to be, whether it is diffused, diffracted, reflected, a single point source and, or whether the light source happens to be fifty feet away, or equal to the distance from our friendly neighbourhood sun.

I may be wrong after all these years complete with an aging memory, and although you may be able to apply ISL to celestial bodies where a mass in motion might have a Newtonian corollary, I would happily read your ISL dissertation or another physicist's paper, regarding the modification of ISL physics with interest...


jim k