What Makes Middle Gray, Middle Gray?

[d_poinsett]

What is the source of the 7-stop average range? It's a reasonable value but I'm curious if it is part of an ISO standard or equivalent? If so, is it just for testing or qualifying requirements in the standard?

 

Using the wonderful graphs and the apparent ease with which they seem to be manipulated, it would be interesting to see what happens to the two and three quadrant graphs when the range is more or less than 7-stops. I'm particularly interested in seeing the effect on exposure choice when using an averaging meter, (Of course, film CI adjustment will figure in there somewhere too).

 

Also, it would be interesting to see how to use the averaging meter to handle those cases where the subject is mostly black or mostly white.

 

Does the utility of a 12% or 18 middle gray reference hold up in these non-average conditions?

[stephen_benskin]

<i>What is the source of the 7-stop average range? It's a reasonable value but I'm curious if it is part of an ISO standard or equivalent? If so, is it just for testing or qualifying requirements in the standard? </i>

<p>

Dave, it�s actually 7 1/3 stops. The figure comes from the major test Jones and Condit conducted in the 1930s. The results from this test were used for the first excellent print test, the determination of the average luminance range (7 1/3), printing LER, and just about everything else in the photographic standards.

<p>

Jones, LA and Condit, HR, <i><u>The Brightness Scale of Exterior Scenes and the Computation of Correct Photographic Exposure</i></u>, Journal of the Optical Society of America, V. 31, n. 11, Nov. 1941, 651-678. The paper concludes the average LSLR derived from the 126 sample scenes was 1:160 (2.2 log units), with a standard deviation of 0.38. This means that 68% of the scenes fell within a 2½ stop luminance range of the average, while 95% fell within a 5 stop luminance range. The luminance ranges of the scenes tested were from a minimum 1:27 (1.43 log units, approx. 4⅔ stops) to a maximum 1:750 (2.88 log units, approx. 9 ½ stops).

<p>

<i>Using the wonderful graphs and the apparent ease with which they seem to be manipulated, it would be interesting to see what happens to the two and three quadrant graphs when the range is more or less than 7-stops. I'm particularly interested in seeing the effect on exposure choice when using an averaging meter. </i>

<p>

Being the world�s worst programmer (amateur status), I have yet to put that variable in. I feel it will answer many questions about tonal shifting and ones I may not have even thought of. I should really just bite the bullet and start working on it. Now, where did I put the Advil?

<p>

<i>Also, it would be interesting to see how to use the averaging meter to handle those cases where the subject is mostly black or mostly white.</i>

<p>

I�m not sure if this kind of program would work on that. I�m sure I couldn�t write a program to work for that. However, Jack Dunn�s book <u><i>Exposure Manual</u></i> does a very nice job on that issue.

<p>

<i>Does the utility of a 12% or 18 middle gray reference hold up in these non-average conditions?</i>

<p>

Excellent observation. No they don�t, but since I�m only showing proof for average conditions, I�m off the hook on that one. :-) In other words, it�s a big issue and I�m avoiding dealing with it for now. Where the hell is that Advil? Pounding in head, Pounding in head.<div></div>

[patrick_gainer]

I think I lost track of the purpose of this thread. It sometimes happens at my age. We are not likely to get a casual observer to point out middle gray in an actual scene. It is difficult for a person with normal sight to see the world in shades of gray. Middle gray can stand as a useful term in expressing the tonal range of a B&W photo, but I have not heard of its use as a psychometric term. I did study perception and learning during my employment at Langley Research Center of NASA while I was engaged in Simulation and Human Factors research. I think if the occasion had arisen we would have taken the log of "middle gray" to be the mean of the logs of extreme stimuli that the astronauts would be likely to see, say, on the Moon. The extremes, however, are of more importance than the mean. We could have the same middle gray with innumerably many values of the extremes. I find that to be true of photography as well. It is well to know where the bottom end of the film is with respect to what the meter is reading, but that can be known without introducing the concept of middle gray. It used to be that the bottom was 1/10 the reading of the meter. The trick then was to set the meter at ten times the ASA number and read the scene where you wanted the bottom to be. No discussion of middle gray was necessary. Now we worry about whether the meter is reading everything as an 18% gray card or what. If I point the meter at something, the meter is reading that something if it is selective enough. Now, if I use an incident reading, I have a different problem. Now I take it on faith that the meter will give me a correct exposure calculation for the average scene, But we still must be concerned with extremes because, as I pointed out, the same average can arise from many different extreme conditions. However, we do know what is likely in terms of reflectance of things in nature. The incident reading has the advantage that it is less likely to be fooled by specular reflections that we don't care about anyway. Still no mention of middle gray, except perhaps in the mind of the designer of the meter or of the ISO committee. Are they the ones who designed the Giraffe? I'd better quit on that one.

[Troll]

Wow! This thread lost me about 10-20 posts ago (I sort of lost count). The original must have been quite a duzy! Wasn't the concept of middle gray defined in the original H&D paper? The practical use of it has come down to us from the Movie Industry, (as so many standards have) because of the incredible cost of NOT having a standard for continuity.