Michael R 1974, I am in the camp where I don't change EI based on conditions. So for simplicity, I would NOT lower my exposure when faced with a scene that is buzzing with flare.
CPorter, I totally support your idea to reduce exposure slightly in such a scene. Maybe you are holding your highlights off the shoulder? Maybe you see the shadows are badly compressed anyway so you aren't making things (much) worse than they already are.
But after staring at this flare-included camera step wedge test family of curves (thanks Rafal Lukawiecki for the data), I think Michael R 1974 is right. You are well-advised to get up out of the very flat toe. I would keep Zone I above the little diamonds (0.1 speed point) or at least above the little 'c' marks (they happen to be what I used as Contrast Index measurement endpoints, meaningless on their own, especially meaningless because of the flare, but I have a hunch these little 'c' marks are a good place for Zone I exposures).
On one side we have the toe that will compress important tones if we let placement fall too low, on the other when we add exposure excessive flare adds more and more generalized exposure that flattens the toe and reduces the the overall contrast of the image especially in the low tones.
For me, given my metering and shooting techniques, I normally have enough latitude to reduce my exposure by a stop in almost any situation without losing important detail. In a high flare situation I could give that latitude away to get better overall contrast. I peg my shots to the middle somewhere and very much live by what Stephan expresses here.
You have more density and theoretical speed increase over the unexposed film where there is no additional light.
Now if you want soft shadows like mentioned before thats fine but if I want to keep the contrast in the shadows I would have to eventually expose more.
I've never reduced exposure more than 2/3 of a stop in an effort to counteract flare density in the low zones (nothing earth shattering), the notion that it causes problems is simply not realized in my experience. The very few times I have done it, only necessitated a quick re-evaluation of the important high value and planned developmet time.
Stephen points out that flare is always present to some degree, I simply do print through it as suggested, it's precisely why I rarely have ever made any attempt at compensation for it. But in the case of "excessive" flare possibilities, I have reduced exposure to no apparent problems----I would rather have a negative that minimizes it to some degree (which I believe that it does), while also printing through what remians. So, if after the shadow placement is made, it is believed that there is added density to it due to flare, then reduce exposure slightly to counteract it, re-evaluate the important high value, maybe adjust the planned development time, maybe not, make the exposure and move on. Some point of theory may suggest that this is detrimental, but I have some instances of photographing that suggests it is not.
Anyway, good or bad, that' how I do it, off to work now.
Also, let's not forget the loss of local contrast and reduced log-H range can be compensated by printing on a higher grade of paper. The increased slope of the higher grade of paper should help restore the local shadow contrast to a pleasing level.
The ISO film speed standard incorporates a stop flare into the equation. The contrast parameters for development is to make the fixed density method correspond to the results from the fractional gradient method which bases speed on the shadow gradient. So gradient is a factor with ISO film speed. I'm not really sure what effect additional flare would have on the ratio between the fractional gradient point and the fixed density point of 0.10. Increased processing will reduce the ratio, which is why the fractional gradient method and Delta-X Criterion have effective film speeds that tend not to move with increased processing (a likely action with excessive flare).
The frightening thing for me is the more I think about flare as pre-exposure, the more my view (expressed in the "challenge" to Chuck's reduced exposure with high flare) seems consistent with Barnbaum's take on pre-exposure of the negative (ie that it doesn't really help) :laugh:
Perhaps a decent way to think about it is consider paper flashing. This effectively increases the "speed" of the paper in the highlights (analogous to what happens in the shadows when exposing the negative under flare conditions), so that it takes less image-forming exposure to get some highlight detail onto the paper. But while it is easier to get that first hint of highlight tone, local contrast in the highlights is compressed, which can end up looking a little muddy. When exposing the negative this is what we have in the shadows. The effective "speed" is increased which means less image-forming exposure is required to get the first hint of density above FB+f, but local contrast in the shadows is reduced.
So I suppose I can summarize my view by saying that an increase in "effective" film speed due to non-image forming light should generally be ignored. This is because for me speed (or exposure index) is little more than a means to an end - ie getting contrast in the shadows as close as I can to the contrast of the rest of the curve. If the extra speed caused by flare/pre-exposure decreases shadow contrast, I'd argue it isn't "effective" speed at all.
There's another factor that hasn't been address and that is the perception of the scene. The viewer expects a certain look from a backlit scene and that usually includes a flattening of the shadows. If the object of the photograph is to be natural looking, then snappy shadow contrast may not be appropriate. Personally, one of my favorate photographic approaches is to play against these kind of conventions.