View Full Version : Testing For Reciprocity

neil poulsen
26-Jun-2007, 07:36
What are different ways that one can test for reciprocity failure to determine the increase in either light level or exposure time that's needed for a correct exposure. I've not done this previously.

For example, do all zones experience the same reciprocity failure? Is reciprocity strictly a characteristic of time, or can the level of light also have an effect on the degree of reciprocity failure? If so, would one want to test reciprocity for one of the higher zones like 8, or for one of the lower zones like 2 or 3? I suppose one could accompany the increase in exposure time with a change in development time. But then, what's a good way to test for this as well? (Perhaps, test for reciprocity to get the correct Zone 3, and then alter the development to obtain the correct Zone 8.)

Lots of questions, so any ideas on how to approach this problem would be appreciated.

Donald Qualls
26-Jun-2007, 10:13
Reciprocity failure, as usually discussed, is a loss of sensitivity below a certain light intensity. As such, it'll affect the lower zones first, with progressively higher zones becoming indistinguishable from Zone 0 as caculated exposure grows longer -- effectively, the range of the film is narrowed from the toe up; if you're in the 4x region of reciprocity failure, you'll see Zone II look like Zone 0, while Zone V looks almost normal for the exposure given -- effectively a loss of film speed and increase in contrast.

That complicates the process of compensating, because when you reduce development to correct the contrast increase, you lose still more film speed.

I'd suggest testing with a step chart rather than a gray card, using aperture and ND filters to increase exposure times while keeping actual lighting constant (to avoid effects from color of the light that can happen if you use dimmers). You'll need to see how all zones are affected so you can alter both exposure and processing when you find yourself forced to shoot in reciprocity failure regimes.

26-Jun-2007, 10:25
I did some net-research on the fast end of the reciprocity problem while I was researching my EG&G sensitometer. I think I read on the Kodak site somewhere the shape of the H&D curve does change (so all zones do not respond the same) but I think this is a small change. My personal experience with the EG&G is that the H&D curves done with 1/100, 1/1000 and 1/10,000 all look pretty similar (suggesting similar response across the zones).

Zakia and Todd has a good chapter on reciprocity effects.

Personally I think testing for reciprocity effects for long exposures will be difficult because to do the tests you need a light source of varying intensity that can be absolutly quantified. Even using an enlarger and a step tab. with a repeatable digital timer falls prey to the fact that the F-stop numbers on the enlarging lens you are using may not be spaced evenly. Perhaps using varying neutral density filters over the step tablet may be whe way to go.

To avoid all the techinical problems, perhaps, just some bracketed exposures of dim scenes might be the best practical solution.

Sorry I don't have a graph of this but here is some of my own raw data from the EG&G sensitometer (which you may or may not know, was designed to test reciprocity effects):
The light intensity was normalized with special filters to just look at changes in the shape of the H&D curve with varying exposure duration (ie to test that all the zones are affected equally).

TMX film in TMax developer in a Jobo processor

[Step #] [1/10,000] [1/1000] [1/100]

18 1.32 1.31 1.31
17 1.38 1.38 1.38
16 1.47 1.46 1.45
15 1.56 1.54 1.54
14 1.65 1.62 1.62
13 1.74 1.70 1.71

Sorry I don't have the 'toe' data in front of me, but, as Donald pointed out, there may be some discrepency there, but from what I recall they were pretty similar.

I know you are inquiring about the other end of the spectrum, but I'm just sharing what I have observed.

Kirk Keyes
26-Jun-2007, 13:17
Neil -

Check out Kit Courter's web site, find the section on film testing: http://home.earthlink.net/~kitathome/LunarLight/moonlight_gallery/index.htm
He has a pretty nice "in camera" system for testing reciprocity.

By the way, reciprocity also occours with very short exposures, not just with long ones. It is a function of "time" x "intensity".

If you want to do it in the darkroom, try using ND filters as mentioned above, either above the step tablet in contact with the film and with a combination of using f/stops and dichroic filters on your enlarger head if you have them. You can calibrate them with an incident meter with a flat diffuser if you want to get fairly precise values for the exposure.

But you are right, different zones with experience different amounts of reciprocity during an exposure. THis helps to account for the increase in contrast that is experieinced with long exposures, i.e. the brighter parts of the image will have less reciprocity failure while the darker parts will have more reciprocity failure which gives the darker part of the image less overall exposure. Thus more contrast.

So it can make a difference with films that have really poor reciprocity characteristics as to which Zone you use for testing.

26-Jun-2007, 16:17
Think of it as simultaneously solving two interrelated problems.

1. Threshold exposure / zone 0 or 1. The immediate issue: what does it take to get some exposure at the very bottom of the scale?

2. Since that will involve, in a low light situation, extended exposure time beyond what the meter says, and since the highlights will not lag as badly as the shadows, contrst increases. Development needs to be reduced to correct the build up of contrast. Reduced development will weaken the density of the shadow areas of the film - so you will be adding a bit more exposure time to hold your shadows.

ND filters are the safest way to eliminate aperture errors.

Make a repeatable set up in the studio, or indoors, as a simplest start. Have it set up with the full range of zones. Shoot it first at normal speeds, and then at various longer times. Process and print to see what's really happening.