An Honest Confession
A recent poll result revealed that we have an almost equal mix of photographers who have been shooting film all along, and those who have recently started. While I’m one of the former, I wasn’t always aware of all the technical behind-the-scenes mumbo-jumbo. And when I started to become aware of the reciprocal relationship between f-stop and shutter speed I found it hard to grasp straight away. (I began with an Olympus XA; an aperture priority rangefinder.) And so this is where this post comes into play. Whether you’ve been shooting for years, or only a few days, it never hurts to examine the theory behind the way things work as they do. Even if it is just revision.
Substance
Reciprocity law states that as you increase the intensity of light reaching the film you also need to decrease the speed it reaches the film by the equivalent amount*.
Translated, this just means that as the light hitting the film gets brighter, your exposures must get shorter in order to maintain the correct exposure. Breaking it down further, reciprocity basically means that two things are in a state of mutual dependence. In our case, the film is in a state of dependence with light, in order to create an exposure. Referencing the reciprocity law again, we can now say for certain that if it is true that more light = shorter exposure, then it must be equally true that less light = longer exposure. This much is self explanatory, but it is also the point at where things start to get complicated.
When Everything Falls
If you look at the table above (you may need to go to the original page for a better view), you’ll notice that there are three columns. The first column is for exposure times when using an aperture of f8. The second column is a translation of these times for an aperture of f220 (apertures of this size are common in pinhole photography). The third column is a list of corrected times for Tmax 100. You’ll notice that when the light levels fall (in other words, when the shutter time increases), the numbers on the 3rd column become much greater than the numbers in the first or second column. For example: Let’s assume that for a correct exposure your light meter suggests a time of 2 seconds at f8 when using Kodak Tmax 100. This sort of metering is common on street lamp lit scenes. If you use a pinhole with an aperture of f220, then your exposure time becomes 26 minutes and 43 seconds. Now remembering the reciprocal relationship that exists between f-stop and shutter, we know that 2 seconds at f8 = 26 minutes and 43 seconds.
However, film being a man made component means that it isn’t perfect. Since manufacturers at this point in time have limitations placed on them by the chemical construction of film. The materials will only absorb so much light before starting to fail; bearing in mind that this is a simple statement of a much more complex chemical reaction, the simple resolution is to increase the time to compensate for that weak point in the chemical reaction. In the case of the chart and our example above, we need to increase our exposure time by almost three times to an exposure time of 1 hour 34 minutes to gain a well exposed image. Why? Read on.
The Perfect Night

Each manufacturer includes information about the film in the box. Each manufacturer also includes the same information available as .pdf downloads on the net. Considering you’re reading this article, I feel it is safe to assume you have access to an internet connection.
Usually this information is referred to as a datasheet. Using Kodak Tmax 100 again as an example, we can see that on page 5 of the datasheet for this film a table is given for adjustment depending on the circumstances. However this particular table only goes up to 100 seconds. Where does this leave us? Generally I just double the exposure time for readings of under 5 minutes, triple for up to an hour, then take a stab in the dark for times over an hour. This doesn’t always work, as you can see above right. The photo was taken using an LSI Diana F+ and Kodak Portra 400NC. Contrary to the title for this section, it was far from the perfect night. Lightning was striking nearby and it was pouring down with rain. I’d forgotten any sort of timing device, so I just opened the shutter for as long as I felt was necessary. I’ve posted the raw scan, and you can see that the colours are starting to appear in the sky, but despite the shutter being open for several minutes when a few seconds would do (according to the light meter), the image has failed to record much detail. (I actually quite like the shot and have printed it at 30cmx30cm with the black level adjusted and it looks great.)
Outro
Basically as far as reciprocity failure is concerned, in terms of long exposures it pays to check the manufacturers datasheet before you head out. If you stick with one type of film you’ll eventually learn how it performs, although you may need to bracket exposures for a few rolls just to make sure you get the shot you’re after!
Additionally, when using colour film (particularly slide film) long exposures tend to create colour shifts. If you’re aware of this before hand, you can harness the creativity if offers as seen in the example on the left. I also cross-processed that roll to deepen the blue that I knew would occur.
Still left with questions and/or still confused? Ask for clarification in the comments below!
*Those of you who are metal fans out there should recognise the headings from Haste the Day albums….Just thought I might have a little fun.


![[22.365] sphere-itize me, captain](http://www.feelingnegative.com/wp-content/uploads/2010/06/2550349404_3251282109-227x240.jpg)




June 23rd, 2010 at 9:00 am
You say “The materials will only absorb so much light before starting to fail” as the explanation for reciprocity failure, but that’s not the reason at all in the circumstances you describe where there is not enough light.
Each grain in the film emulsion needs a certain number of photons in order to be “activated” and form a latent image ready for development. When it’s too dark the photons don’t arrive often enough for a stable latent image to form so the exposure requires much longer. This isn’t a chemical limitation due to the film construction, but just a matter of physics and probability.
[Reply]
Tomas Webb Reply:
June 23rd, 2010 at 1:13 pm
Indeed. I wasn’t sure how to cover this without going into too much scientific explanation – you put it very succinctly and it makes me think I should have tried to at least touch on the scientific reason. Thanks for the correction
[Reply]
June 23rd, 2010 at 5:33 pm
I have a general question to people here: say that you have a light meter or similar device that tries to take reciprocity failure into account for you. It may have data points for specific times and films, but – as Tomas exemplifies – only up to some time, say 100 seconds, for a specific film.
Now, you want to take a shot that is longer than that. Do you want the device to make a “best guess” based on the lower values (even though it will probably be off by a fair bit), or do you want the device to simply refuse and tell you it doesn’t have a clue what the right exposure time is?
Let’s for arguments sake say that you can add data points whenever you like. So on one hand, after taking that wrong guess you can go back and enter a more reasonable value based on the failed exposure; on the other, you can always get around the device’s refusal by manually entering a reasonable-seeming value for that kind of time.
Thoughts on that?
[Reply]
Tomas Webb Reply:
June 24th, 2010 at 2:59 am
There is probably a market for an iphone app that calculates published manufacturers data on reciprocity failure for individual films, then allows for user input for personal results in excess of the published data.
The only problem with the personal data is that it would be dependant on personal sets of variables (for the most part) that would vary from photographer to photographer.
[Reply]
Janne Reply:
June 24th, 2010 at 4:57 am
My idea is for people to really do _personal_ data; no sharing with others. As you say, there’s too many parameters to make it widely useful for others. What I’m getting at is if an extrapolated time value would be useful for people or if it’s better to not guess at all.
It’s an Android app I’m planning for, and a free one. I mostly just need it myself, and if I manage to finish it I may as well share it with others.
[Reply]
Tomas Webb Reply:
June 24th, 2010 at 5:11 am
If I owned an Android or application capable phone then I’d definitely be interested in something like that.