rmalarz wrote:
Scotty, the situation you describe here leaves out one major component...the sensor/camera as a unit and how it handles exposure. ... --Bob
I am very familiar with all of that. I have examined many images using RawDigger and I know when highlights become irretrievably blown.
To recap, testing with RawDigger on two cameras (D610 and A7 II) has shown me that both of them place the raw values for middle gray (the middle of
exposure Zone V) between raw values of 1000 and 2000 on a 14-bit scale. That means that the top of exposure Zone V is at about 2000 (note that these are exposure zones, not print zones, which range from 0, maximum black, to X paper white) Zone VI reaches 4000, Zone VII to 8000 and Zone VIII to 16000. If there were a usable Zone IX it would reach 32000 but a 14-bit raw file blows out at 16383. For all practical purposes there is no exposure Zone IX in either of these cameras. That leaves only 3-1/2 stops of latitude above middle gray. I would be surprised to come across any other Nikon or Sony camera that behaves differently. I don't have access to a camera with a 16-bit raw file but I bet it would have more latitude on the high end.
The bottom of
exposure Zone V is 1000, Zone IV is 500, Zone III is 256 (switching to powers of 2), Zone II is 128, Zone I is 64, Zone 0 is 32, Zone -1 is 16, Zone -2 is 8, Zone -3 is 4 ... This explains why there are a lot more zones available below middle gray than above it. It also explains why ISO Invariant cameras can handle a great deal of underexposure before the darker portions become irretrievable banded and ugly.
So there is more latitude for underexposure than for overexposure in in the raw file of a digital camera just as there is more latitude for overexposure than for overexposure with negative film.
Of course, underexposure followed by a correction in post processing is likely to magnify the visibility of noise. But we both know that the visibility of noise is a function of exposure, not ISO. If you expose normally at a reasonable ISO, noise is not the issue.
I realize that not every image includes something as reflective as an egret and that there are cases where you can get away with moving the histogram to the right until the JPEG looks somewhat washed out (and maybe even greenish) only to move it back to the left in post processing to restore a more natural looking image. I prefer to be able to review the image on the LCD while it is still in the camera.
ETTR begs the question, "Why bother?" We have seen all the demonstrations we need of how it is done. I am still waiting to see a demonstration of a visible benefit.