# How does a honeycomb grid affect the travel of light?

A light modifier that is commonly used in studio photography is a honeycomb grid.

It narrows the beam of light to a circle with soft edges, as it can be seen here:

My question is: how is this happening?

A small reporter flash has a rectangular shape, if you place a rectangular shaped grid on it, it produces a "soft" circle of light. How is the light travel modified by the structure of the grid?

-
One thing I find helpful is to imagine myself sitting in front of the light (where the portrait subject would be). If you're looking straight down the "tubes" of the grid, you would see the blinding light directly, so you're brightly illuminated. If you move off to the side a bit, you no longer look straight down the tubes, and you just see softer light that bounces off the walls of the grid, so you're only weakly illuminated. – coneslayer Feb 25 '11 at 18:08
@coneslayer: That would have been a perfectly good answer worthy of a vote or two. – dmckee Feb 25 '11 at 20:47
Indeed, that would be a very concise and intuitive way of explaining this. – Colin K Feb 27 '11 at 4:05
@coneslayer: that's exactly how I imagined this indeed! But I wanted something a bit more "formal" to confirm or infirm my intuition :) – Olivier Lance Feb 27 '11 at 22:03
Also an important distinction is that a honeycomb grid does not "focus" or "direct" light. Those are merely secondary effects. In reality, it blocks light. – user7042 Jan 8 '12 at 13:38

## 1 Answer

From looking at the pictures of the grids provided on that website, I think I understand what they are doing. It's not actually very complicated.

It seems easiest to explain this if we consider only one of the hexagonal "tubes" formed by the grid. Normally, the light from a flash spreads out over some (relatively wide) angle, but now think of what happens to the light when it is confined to this tube: The light that would normally head off at a steep angle is intercepted by the walls of the tube, while light that was already propagating parallel (or at least close to parallel) to the tube axis is passed without obstruction. This results in a more confined range of angles over which the light can spread. Of course, the confinement is not perfect, as I will explain below, but you can imagine how the angular spread of the beam would be limited to consist only of those rays which can pass cleanly through the tube. This is simple geometry, and you could adjust this angle by varying the length and diameter of the tube.

However, the light that is intercepted by the tube walls is not completely eliminated. Instead it scatters into a random direction. It is this scattering that gives the resulting beam a nice soft edge. Instead of being sharply confined to the angular range allowed by the tube, the scattering allows it to "bleed" a little bit.

I should add that the hexagonal shape of the tubes probably has very little effect on the resulting illumination pattern. More likely a hexagonal shape was chosen for some other reason, like easy manufacturing or more mechanical stability. Optically, a square grid or even an array of circular tubes, like a bundle of straws, would work just fine.

-
Assume the honeycomb to consist of tubes of length l and diameter d, depending on the angle the visible/active cross section also get's smaller. Right now I'm too lazy to do the geometry, but I guess (educated / heuristicallly) that the distribution will be something like 1-sin^2(alpha*l/d) for the direct pass. Then there's the scattering term. – datenwolf Feb 25 '11 at 18:07
Yes. Isn't that what I said? – Colin K Feb 25 '11 at 18:10
Indeed, but you didn't quantify it ;) – datenwolf Feb 27 '11 at 14:25
@colin-k many thanks for your answer, it confirms the feeling I had on this! And you're right about the geometry of the tubes, one of the DIY ways of making a "honeycomb grid" is by using black straws, and grids are indeed square grids on large softboxes ;) – Olivier Lance Feb 27 '11 at 21:58
I'd suspect the softness of the shape comes from diffraction (around the edge of the hex) rather than scattering? – Yakk Aug 17 '15 at 18:02