I want to correct a misconception about the “Inverse Square Law”. It’s not as easy as it’s cracked up to be.
The ISL ONLY applies to point-sources of light, which do not exist in the real world. It’s a generalization that is useful, but only if you know what you’re doing.
Consider a white fluorescent light — one of the long ceiling ones.
Let’s say you put your hand close to one end. Let’s call the amount of illuminance you receive X. Now, move your hand to the other end of the bulb. Has X changed?
You’ve moved a significant distance from the original point, but the intensity of the light arriving on your hand has not changed a bit. Why?
It’s because you’re dealing with an area of light, which contains an infinite number of theoretical point sources. When you move away from one point, you are actually moving towards another.
The same holds true when dealing with distance *away* from the light, instead of just *across* it. Moving your hand away from the light, you are moving a different distance from each theoretical point source. Doubling your distance away from an area source does NOT halve the illuminance, especially when you’re very close to the source.
This is why it’s very important to use multiple-sampled area lights to simulate these kinds of light sources.
However, if you are far enough away from the light source, then the ISL is a “close enough” approximation that works really well, so long as you are using a linear workflow.