The Inverse Square Law Experiment Done Right - Myth unBusted
Yesterday I posted a cheat sheet that tried to question the applicability of the Inverse Square Law (ISL) on the way we use portable flashes I called this post The Inverse Square Law Cheat Sheet - Myth Busted.
The post stirred up a great conversation from which I learned about Light, some physics and some in camera processing facts. But mostly I learned that it is great fun to experiment and to share your findings. It definitely helped me get my knowledge to a higher level (at the small price of throwing a way my totally wrong fringy and conventions breaking experiment.
Firstly, The Inverse Square Law Experiment Done Right
Indecently Tim Rogan conducted the very same experiment just a few weeks ago.
Images by Tim Rogan
Tim took a bit of a different approach to measuring light fall off. Surprisingly, his results were almost dead on with the ISL:
Measured the light with a light meter from a flash on manual mode.
Lines are whole stops. Even when it looks totally dark on the wall you
are still getting light. Light was measured by holding a light meter
flash head high about a foot away from the wall.Using the flash I put a piece of tape at every whole stop
This Make Me Wonder What Was Wrong In The First Experiment
Thanks for every one that cheeped in the discussion - I will bring the main points here, however, I encourage everyone to go to the original post, read the great comments, and join the discussion.
Location Of Measurement (Joshua Targownik)
Light has to be measured on Axis constantly. Measuring light off axis contributes to part of the falloff.
Reflected Light and Incident Angle (Joshua Targownik)
The light measured on the original experiment is reflected from the wall (which in itself is not ideal), but further more - the light is reflected in different angles depending on the distance from the flash. The deviation in the incident angle changes the amount of reflected light.
Surface Across The Wall Is Not Even (Anon)
This one is an obvious one - The wall has dents, planes in different angles, and color is not similar across concrete and block - all of which play a part when measuring reflected light.
Blown Highlights (Sean Phillips)
Any measurement that falls on the 255mark is inaccurate. basically the sensor got over exposed, so we don't know what the exact value of exposure would have been if the light was within the limits.
In Camera Curves (Matti Rintala)
Camera has built in curves that are meant to enhance mid tone, hence post-shutter-click pixel data cannot be used for light measurements.
Even So, There Are Things To Learn From This Exercise
We Still Get To See The Light Patterns (Roger Barnes)
While we cannot deduce about the light intensity, we can see in what shape the light travels with each of the modifiers.
Reflection on the ISL (olli Rinne, Max)
This is a good opportunity for reflection, how can it be that the light behaves the same when it is directed by a small reflector/ lens inside the flash. What factor is missed. Of course, as Olli indicated, light is not a Laser. However, "simple" flashlight, where the light goes through a lens does not obey the ISL. This is a point to think about.
On the other hand, Max came up with a nice metaphor - think about the flash as a collection of point light sources
Additional Resources (Matti Rintala, Ian Mitchell)
"Light: Science & Magic" by Fil Hunter, Steven Biver, and Paul Fuqua is a great place to learn about the physical qualities of light.
Zack Arias provides a great explanation on ISL.
Get the DIYP greatness via RSS, newsletter and TwitterConnect with the community: Readers Photos, Discussions, Readers Projects
Comments
Aw dang, they all beat me to it.
Yeah, the first experiment was pretty deeply flawed. One thing of interest however, is that the inverse square law holds only for a point source of light. At any reasonable distance a bare flash is a pretty good approximation of a point source; however with a softbox or umbrella it's not - that's why we use them!
When the subject distance is small compared to the size of the light source, there is no falloff at all - intensity is constant with distance. When the subject distance is large compared with the size of the light source, the inverse square law takes over. In between - when subject distance and source size are similar, it is more complicated but basically falls somewhere inbetween... so at short distances, (up to 1-2 umbrella diameters, say) you can expect the light intensity to be less sensitive to distance than it will be beyond further away.
To do the flash experiment totally correctly the best way would be to take multiple exposures (on manual with constant settings) of a white card with the face of the card normal to the axis of the flash and the lens axis as close as possible to parallel, at varying distances. I think you'd find very good agreement with the inverse square law.
Congratulation
That's great you can reconsider your conclusions about your experimentation. A lot of people can't. I give you an 'A' for your mea-culpa and good work to explain where you're wrong.
distance from light to subject
Here are some tricks related to light placement and f-stop numbers:
If your light is 8-feet from the subject and you want to increase the light on the subject by one stop move the light to 5.6-feet from the subject.
If the light is 8-feet from the subject and you wish to decrease the light on the subject by one stop move the light to 11-feet from the subject.
These distances are one stop increments in terms of light intensity on the subject:
4', 5.6', 8', 11', 16', 22'
Of course, they are also f-stops on your camera lenses.
My interpretation
It's always seemed easier to me to understand the ISL when I think of light as a quantity of photons spreading, rather than an "intensity value". So if a certain "amount" of light spreads from a single point to a certain area at a certain distance (lots of "certain" in the same sentence), the same "amount" of light will spread on an area four times bigger at twice the distance (think Thales theorem applied to 3D space), thus reciprocally, the light intensity can be considered four times smaller.
This interpretation implies that any beam of light that is expanding with the distance will be subject to the ISL, whether it really spreads sperically, in all directions, or only as a beam that gets larger with the distance. The only case in which I think the law would not apply was if all the rays of light were perfectly parallel (which is virtually impossible and mostly theoritical). The very same area would be lit at any distance, so the same area would receive the same amount of light : light intensity does not decrease at all with the distance (except for the part that is "absorbed" by the air and dust or moisture or any other kind of obstacle to the light in it). In practice, this approximates the results obtained with a very far light source (e.g. the sun): no light fall between subject and background.
Anyway, whatever the experiment, we should bear in mind that the ISL is an approximation of the reality. Light fall is much more complex to represent. Max's metaphor, which seems perfectly adequate, shows how difficult it is to determine precisely the light fall when we have an inifity of point light sources...
re: ISL
Thanks Spica.
It's the bit about the spreading of the light (no matter if spherical or not) that was the hardest for me to understand.
Zack Arias - OneLight Workshop
I highly recommend watching this DVD. He goes over the ISL, plus it's just an amazing dvd to watch.
http://www.zarias.com
Post new comment