Brian wrote:
.
So, the much-touted accuracy of an 18% gray card exists only for a scene brightness range of 5 stops. Accurate exposure is best found by spot-meter readings of the darkest and the lightest parts to be recorded in the scene, and the exposure set accordingly. For a scene of 5 stops, the card will do it. For contrastier or flatter scenes, the card will be off.
This’ll probably start a war …
Colin
Colin,
I just read your last sentence again. You mentioned contrastier scenes. Let me explain it now. What affects exposure is actually how much light is landing on a scene, not how much light is reflected from a scene. <snipped remainder>
With respect, Brian, I think your point of view has been derived from taking for granted the 18% card as being accurate for all occasions. That is not surprising, given that probably 95% of photographers believe the same. At the risk of boring people – though I think that learning is never boring – I’ll put it another way that might be easier to get hold of.
Let’s say you meter your shots with a spotmeter, and you do it the ‘Ansel Adams’ way, first measuring the brightness of the deepest important shadow, and then the brightest important highlight, not counting specular reflections.
And, let’s say the meter indicates 1/100 at f/4 for the shadows, and 1/100 at f/22 for the highlights. That’s a 5-stop range. Now, the chosen exposure to be used is obviously f/9.5, the median stop between f/4 and f/22, right? So we shoot at 1/100 at f/9.5.
Now, in that scenario, the brightness range of the scene extends 5 stops, and the median stop is 2½ stops away from the deepest shadows, and also 2½ stops away from the brightest highlights. So, the brightness ratio between f/4 and f/9.5 is 2^2.5, which is, according to my calculator, a ratio of 5.66 to 1. Likewise, the ratio between f/9.5 and f/22 is also 2½ stops, i.e. 5.66:1. Magically, if we multiply the two ranges to get the total range, we get 5.66×5.66 = 32.03. The .03 is due to rounding errors, so drop that, and we have the brightness range of 32:1 that we started with.
Finally, the ‘grayness’ of the median exposure, since the shadows are black and the highlights are white, is 5.66/32 of white, and, as in my first post, 5.66*100/32 = 17.69%.
If we had used an 18% gray card, it would have given the exposure we calculated, 1/100 at f/9.5.
Now, if you apply the same logic to a six-stop range, the median gray will be 12.5%, and so on.
Colin