Here are a couple of excerpts from Jack Holm's paper, Exposure-Speed Relations and Tone Reproduction, IS&Ts 47th Annual Conference, 1994. I believe it sums up what has been covered so far and begins to put into perspective the relationship between film speed and mean exposure.
“The speed values obtained from these exposure values should place the exposure values so that the corresponding tones of the scene are optimally reproduced.”
“Film speed exposure determination standards are based on the correlation of two experimental values, the film speed and the average scene luminance. These standards describe a speed point exposure which is deemed to produce a density or densities of significance for a particular film type. The optimal ratio between this exposure is then determined and used in the calculation of a constant for the speed equation. Film speeds are obtained for specific films by determining the speed point exposure and plugging the value into the speed equation to get the film speed. Since this film speed was determined assuming a particular speed-point-exposure/mean-exposure ratio, if this ratio is approximately correct, setting the film speed value on an exposure meter and reading the scene should indicate the correct exposure. The trick is to select a speed point which allows for the maximum variation in the actual speed-point-exposure /mean-exposure ratio with the minimum variation in image quality."
“It is important to note, however, that mononumeric speed values provide information about the predicted output at only one point on the characteristic curve. If a speed value is based on a shadow density, it will predict only the exposure required to produce that density, and similarly for highlight and other speed point exposure criteria. Also, typical light meters can only read one luminance value for a scene and thereby predict only one film plane exposure. The actual tone reproduction is therefore dependent on: the speed point exposure, the desired reproduction of the speed point exposure, the metered exposure, the desired reproduction of the metered exposure, and the difference between the speed point and metered exposures.”
Last edited by Stephen Benskin; 01-05-2012 at 09:24 PM. Click to view previous post history.
So, what is this important ratio between the speed point, film speed, and metered exposure all about? This ratio is what defines the relationship between the speed point and the metered exposure point. or more specifically, the film speed and the exposure placement.
The idea is to find the optimum placement for an average scene because of it’s high occurrence, while at the same time considering maintaining the highest level of quality with variance in exposure and scene luminance range. In order to determine this placement, four variables need to be known: the film’s speed point, the metered exposure point, the average scene luminance range, and the ratio between the speed point exposure and the metered exposure.
The first three variables have already been discussed. The ratio is easy to determine. It’s Hg / Hm = k1. With the current B&W speed standard for a 125 speed film that would be:
0.064 / 0.0064 = 10 or a log 1.0
For color reversal it would be:
0.064 / 0.08 = 0.8 or a log - 0.1
What this means is the B&W ISO speed point is 1.0 log-H units to the left of metered exposure, and the color reversal ISO speed point is 0.10 log-H units to the right of the metered exposure.
Why is the ratio significant? If the distribution of the luminance range above and below the metered exposure point is known, optimum exposure placement can be determined by adjusting the ratio value. Changing the ratio between Hm and Hg is possible simply by changing the speed constant.
This is about the clearest "one graph" illustration of Zone System "Normal" I've seen.
Originally Posted by Stephen Benskin
What about how the shadow exposure falls 1 stop below the speed point?
Sponsored Ad. (Subscribers to APUG have the option to remove this ad.)
Well, I place my shadows on Zone II which is one third-stop above the speed point in your illustrations. And it is interesting that there "is" image information (that we don't expect to be able to print) all the way down to the 0.3G point (which is pretty close to Zone I).
Take another look at it. Zone I falls at 1.20 log-H units below the metered exposure, but that's 0.20 log-H units below the speed point which also happens to be at 0.10 over Fb+f. The two don't seem to match up. What affect do you think this has on comparable film speed values?
Last edited by Stephen Benskin; 01-08-2012 at 02:51 AM. Click to view previous post history.
The purpose of the speed point – metered exposure ratio is to link the sensitometric exposure of the film speed testing with the camera exposure. As film speed is derived from a sensitometric exposure, it is determined from non-flare testing conditions. As a certain amount of flare exists in shooting conditions, adding a flare value to the camera exposure equation will produce more realistic film plane exposure values.
Veiling flare has the greatest influence on the deep shadows and it’s influence progressively tappers off as the exposure increase. Because of this, I’m going to focus on the exposure below the metered exposure. The example here has a flare value of one stop. The flare value is determined using the last shadow exposure value. A one stop flare factor doubles this value and then adds it to each of the other exposure values. This will double the shadow exposure but will only add around 2-3% to the metered exposure.
Flare has effectively taken the scene’s luminance range of 2.20 and reduced it to an exposure range of 1.90. This change is down almost exclusively below the metered exposure. The exposure range has gone from a scene luminance range of 1.30 to a film plane exposure range of 1.0.
It’s done this by increasing the shadow exposure by one stop, which moves it to the right on the film curve. In the case of a one stop flare factor, it brings the exposure up to the speed point. In this situation, the metered exposure falls at a 10x ratio to the shadow exposure.
Flare isn’t consistent from scene to scene. It varies depending on a number of factors including luminance range and the distribution of the luminance values within that range. And flare is virtually impossible to measure in the field. It’s a variable that can’t be controlled. Let me repeat the Jack Holm quote from post #31, ‘The trick is to select a speed point which allows for the maximum variation in the actual speed-point-exposure /mean-exposure ratio with the minimum variation in image quality.”
This is a question of exposure placement and variance. Keep in mind that the film speed – meter exposure ratio is not intrinsic, nor is it arbitrary. It was chosen for a reason. With a flare factor of 2 (one stop) representing average flare, the shadow exposure for an average scene falls approximately one stop above the fractional gradient point. This point is the minimum point of exposure that can still insure a quality print. Even in a non-flare shooting situation, which is impossible, the shadow exposure will still fall at a usable point. This potential “safety factor” also allows for some accidental underexposure or scenes with slightly deeper shadows than the average.
As has been frequently discussed, overexposure within a certain range, doesn’t diminish image quality.
The following example uses the same film curve as post #33 with the exposure overlay incorporating the one stop flare factor.
There’s a argument to be made for a slightly higher flare factor in shooting conditions, especially with smaller formats. A 1 1/3 stop flare factor moves the shadow exposure to the right an addition 0.10 log-H units.
What happens when the film speed constant changes, effectively changing the film speed – metered exposure ratio?
Starting to hone in...
I keep confusing that while I place my shadows on Zone II, I should expect (non-flare) Zone I to be the 0.1 speed point.
In another thread I had this thought...
I can't reconcile the ideas of using statistical average when I could quickly spot shadow and highlight and know the actual average.