I have an example of this where I was trying to capture the insides of my father-in-laws wool shed. My 1st attempt was unprintable. I went back and took some more at a better time and light conditions. I think that's your answer! Go back real early in the morning or late at night, give it enough exposure for the interior and develop a lot less than normal. Print with every trick you can pull out of your sleeve...
One way of supporting shadow exposure and not have the highlights be so severely affected is to do a double exposure of the film. The first exposure would be done through a diffusion material (such as opaque acrylic panel) at a zone III exposure. This would be determined by metering through the diffusion material. This would then be followed by the second exposure that would be indicated by the scene brightness.
The effects of this can best be described by the effects of the following example:
1. Assign a numerical value of 1 to Zone I, assign a numerical value of 2 to Zone II, assign a numerical value of 4 to Zone III, assign a value of 8 to Zone IV...since each zone of luminance is a doubling of light value, the numerical value of Zone VIII would be 128.
2. The Zone III pre-exposure would effectively raise Zone I luminance in the second exposure to a Zone III 1/3 negative density, it would raise a Zone II luminance in the second exposure to a Zone III 2/3 negative density, it would raise a Zone III luminance in the second exposure to a Zone IV negative density, and it would have a minimal addition to a Zone VIII second exposure (4 units of light added to a 128 Zone VIII value). This may take the scene out of the reciprocity consideration (for long exposure, depending on the film)...additionally it would allow the highlight values to be kept within manageable negative density.
3. Lets assume that we have a scene which has a brightness ratio of 15 zones. By exposing the first exposure at a Zone III pre-exposure and followed by placing the deepest shadows in the second exposure at a Zone IV placement (effective Zone IV 1/3 considering the pre-exposure), we will have contracted the actual density as recorded on the film to 10 2/3 Zones. While this is still not optimum, it is certainly much more manageable at the printing stage.
3. Additional to the halation consideration that others have mentioned, the other aspect is one of lens flare. It is vitally important to have the lens absolutely clean and multicoated. Any haze or film on any of the glass surfaces will cause the light to scatter (especially in the example addressed). The place that it will scatter is to the emulsion of the film.
OK, based on all of the preceeding comments I propose the following actions as the most likely to solve Aggie's problem with this particular lighting condition.
1. Expose the film so that you can develop for less time. This will tame the highlights.
2. Use a multi-coated lens. Can make a big differnce in this type of lighting.
3. Use an efficient lens hood, even with a multi-coated lens.
4. Use a two-solution developer (D23, Diaxactol, etc.) to develop the film. And BTW, I have a Pyrocatechin based formula that is used in two solutions that works a lot like Diaxactol. If anbody is interested contact me directly by email.
I firmly believe this condition is a phenomena of staining developers. I have gone back through my negative files and compared a number of still life negatives of a magnolia taken against a black background in carefully controlled lighting conditions. I have negatives developed in PMK Pyro and XTOL and careful comparison scanning into photoshop shows for the film developed in PMK has a distinct "halo" around the edge of the magnolia. This effect is not present on the XTOL developed negative. It reminds me of the effect you when a subject that has high IR reflectance using Kodak HIE.
The PMK version can be found at http://home.pacbell.net/mkirwan/still_life.htm
It has plenty of highlight detail and the highlight bleed is not noticeable in the print but it is there!
My recommendation for taking pictures under extreme conditions would be to use a highly compensating developer, maybe even the water bath method but that effectively doubles your development time. I am a fan of diluted D-23 followed by an afterbath in either Borax or Kodalk. Cheap and effective http://apug.org/forum/html/emoticons/smile.gif and allows you to expose for the shadows and effectively develop for the highlights.
The reson I referred to the woodlands scene is that they show the highlights spilling out into the rebate. This shows that the light does NOT travel any of the "ordinary" ways, but must somehow spread inside the emulsion layer. The spillage is also very similar in all three examples, showing that the developer has very little effect on this.
Developing less would have an effect, but then the negative would be thinner overall which may not be wanted.
I believe I have some old 35mm negatives showing the same flare/halation problem - only worse because of the smaller negative. I'm pretty certain one of them is on Ilford XP1, a chromogenic film.
The main effect that ifferent developers will have on this henomenon is in the placement of the shoulder. A developer/film combination putting the highlights well into the shoulder will show the effect more clearly than one where the highlights are in the straightline-section. It would be interesting to know how the old Super-XX behaved in this respect, as I seem to recall that it has been characterized as "all toe". A film that is "all shoulder", such as XP2, will give denser flare areas.
I agree with Ole, seems the chromogenic films are more prone to this effect. I have plenty of them.
The thing about opinions is that everybody has one. In my opinion the type of bleed over from the highlights to the shadow we have been disucssing is in no way way a phenomna of staining develoeprs, as Bob K suggests.
However, if his opinion is indeed true it would be quite a revelation and I am certain that a well-documented study that could conclusively prove that staining develoers are more prone to bleed-over in this type of situation than traditional developers could easilsy get published in one of the national magazines. Some simple paramemters are the following.
1. The scene needs to be one that juxtaposes extreme areas of lighting, and it would be good to also have some fine detail present, such as tree limbs and leaves.
2. Both negatives (same film type of course) need to show the exact same scene, be exposed at the same exposure and within seconds of each other with the same lens and hood set-up. The position of the camera must no change between exposures, and the light must remain consistent.
3. The negatives must be developed to the same density range, and this DR should be the effective exposure scale (ES) for the process (or type of paper) being used. This could be determined by reading the densities with a densitometer, making sure that the correct color mode that is used for the stained negatvie is approriate for the process.
4. The result would have to be clealry visible on a print.
5. And since development artefacts are not uncommon the result would have to be confirmed with at least two or three additioinal separate tests.
Sandy, I absolutely agree.
The only correlation with developer I can imagine is that a highly compensating developer - regardless of staining - could possibly make it more visible. The reasoning behind this is that more of the flare would be in the shoulder area of the exposure/density graph, so the flare would have higher density compared to the shadow areas.
A scientific study of this would be very interesting...