I do lots of night photography and shooting in dark abandoned buildings and have my exposure times nailed down fairly well, but recently I came across a situation which I don't know how to handle:
The time is the predawn just after the first hints of brightness start turning up in the sky.
The location is a cavernous warehouse type building open at one end to the growing light.
The exposure is started when the interior is still VERY dim and extends for at least 17 minutes (I'm shooting HP5, 4x5, f16).
My question involves the mechanics of reciprocity failure. When I first open the shutter, the lighting is extraordinarily low, but by the time the last few minutes of my exposure rolls around, the light in the building must be several times as bright (still too dim to adequately meter, though). So the majority of the actual exposure probably takes place at the tail end of the period of time that the shutter is open. Is there something about the long, inadequate, exposure up until that point which desensitizes the film to that later period of brighter light? In otherwords, would the film register those last few minutes in the same way whether it was preceded by a longer period of exposure to dimmer light or not? I hope I'm making sense with this question, it just suddenly occurred to me that I might get more to register on the film by just waiting for those last 5 minutes or so rather than going for those five and the loooooong, boring 12 that preceded them.
Hell. Did you take a look at Ilford's website? www.ilford.com
All the best
When do we get to the hard questions?
This is not completely "reciprocity failure" - a phenomemon directly linked to time. As we all know, "reciprocity" is the idea that an exposure of 1 second at f/16 is equal to an exposure of 1/2 second at f/11; and equal to 1/4 second at f/8; and 1/8 second at f/5.6. However, that "law" fails at *very* long times ... an indicated exposure of 10 seconds may well require 20; one minute exposure may require two minutes, etc. - depending on the characteristics of the film. The data sheet for the film usually specifies *rough* compensating factors for long exposures.
What you describe is something more... and could get theorectically *complicated*.
There is no assurance that the increase in light at "dawn" would be linear; if so, one might "strike an average" - but even so, reciprocity compensation is definitely NOT linear.
So - it is a problem. I was going to use the catch-all for exposure problems - "bracket" ... but that wouldn't work either.
Uh ... try it and build up experience? Take notes and ... hope for the best.
" Take notes and ... hope for the best."
That's pretty much what I've been doing. I guess I'm not looking for a formula (which I'd only ignore anyway) but for some idea of how the "failure" comes about -- whether there's some kind of progressive "desensitization" (for lack of a better word) at work which generally would make my way of shooting in those situations self defeating.
I haven't yet found anything on the Ilford site that helps much. I gather the issue doesn't come up often (!?)
What I would do in this case is to spend at least one morning at this site with only a spot meter and without the camera. I would note the changing light conditions at one minute intervals. Obviously in order to duplicate conditions, the sky must be clear. Taking a note from Les's book and also Ansel Adams as well, previsualization would serve one well in this case. From the notes that I would have made, I would be able to determine how this scene would render in a print and make my time of exposure accordingly. From this time of exposure, I would then work out the reciprocity considerations, since I have already noted the light conditions. I would use an "average weighted" exposure since to guage reciprocity on the beginning of the exposure would be inaccurate, just as guaging the exposure on the ending would also be inaccurate. One must note that the daylight is either lengthening or shortening each day depending on the season. The times of sunrise are published for most locals and dates. This would determine the relationship to actual exposure versus the test data.
To answer your question as it was posed, the film is exposing at an increasing rate because of the increase in intensity of the light conditions. If you are determining your reciprocity considerations on the light at the beginning of the exposure, you will have an overexposure. If you determine it on the light intensity at the end of the exposure, you will have an underexposure. Yes, obviously the light at the beginning of your exposure, so long as it was of sufficient intensity, would serve to break the exposure threshold of the emulsion. How much? I can't tell you precisely...obviously not as much as the later and more intense light.
Hope that this helps you. Good luck.
How much brighter is it at the end compared to the beginning? If it is more than, say, 4 stops, the beginning of the exposure might just be irrelevant, as the ambient light is usually ignored in exposure calculations with studio strobes, because the strobes are so much brighter unless a very long shutter speed is used.
Alternately, if the first part of the exposure is enough to bring the emulsion up to the threshold of exposure, then perhaps you might try to compute it like you would a pre-flashing exposure, except that the pre-flashing effect would only be relevant for the brighter parts of the image.
This is an absolute gem of a question and I have to say that I've never even considered this problem when I've made very long exposures so I'm sorry that I cannot give you a scientific answer. However, when I have given long exposures in changing light I've always erred on the side of over exposure for if it is on the negative you can print it. Even if you think the highlight is blown out I reckon it will print for film will hold up to 14 stops of contrast. On the other hand under exposure is not recoverable. So my advice is not to worry about the problem if you have information throughout the whole negative.
However, from the replies so far we are all clearly very interested in finding out the correct procedure to deal with this so I will talk to the technical boffins in Ilford UK to see what they say and report back, that is if I can understand the the complicated answer that I suspect I may get.
The cause for the reciprocity failure lies in the formation of the latent image. When a photon hits a halogen ion, an electron is freed and floats thru the crystal structure until it finds a suitable disturbance. This negative charged disturbance attracts free-floating silver ions. When they meet, the silver ion becomes a silver atom. This forms the first nucleus of the latent image. However, a single silver atom surrounded by lots of silver ions has a lifetime of a fraction of a second, depending on temperature. It takes a cluster of four of these silver atoms to form a stable latent image that can be developed. The lower the light, the more silver atoms will convert back to silver ions before a stable structure can be built. This is causes the loss of film speed.
Although this might not answer your question directly, it may give you an impression of what will happen. Although bracketing might be a silly recommendation in your case, you will hardly be able to calculate an exact exposure.
I can envision an exposure meter that would be equipped to handle the situation.
A real-time continuous photometer, similar to those built-in to the shutter mechanism of "Auto-exposure" cameras, with a processor capable of being programmed to the reciprocity curve of the film. Could be servo-coupled to the shutter of LF cameras.
Now to make it pocket-size, and figure a way to produce the thing for $3.98...
Thanks for all the responses so far.
In a practice, Les is correct that I have yet "over-expose" the situation. Over exposure isn't really the concern, it's how to achieve ANY exposure in the most economical way that's the question.
Also, I have bracketed the shot -- both by going out on consecutive days and on one memorable occasion, by shooting the P6x7, the 4x5 and 5x7 at the same time.
The end result is that I've gotten the shot I was after, but I'm still not sure I did so in the most efficient (and repeatable) way given whatever it is that's going on during reciprocity failure.
As for metering the scene, that's impossible. The lighting conditions are just too dim. For most of the type of things I like to shoot I've learned not to even bother taking a meter out with me.
Anyway, thank you all for your thoughts and if any more occur to you, please post.