Submitted by Roger Breton on
Forums:
Still new to RawDigger and most of the concepts herein...
One of the striking features to me in the Histograms is the presence of an "EV scale".
I understand that EV stands for "Exposure Value", a legacy photographic term and related to scene Luminance.
What strikes me most is this "EV=0" reference?
Page 26 of 85 in the User Guide has a section of the "EV grid" where it is explained that "EV0" is automatically calculated from the image statistics: "Located at 3 stops lower than the maximum pixel value, rounded to the nearest power of 2".
Silly me! I thought, naively, that this "EV0" was an absolute "Reference point"?
And was under the impression that EV0 *had* to correspond to 18% gray reflectance?
For example, in my ColorChecker NEF, the N5 patch (18% gray) has these RGB values :
R: 869 of Max 4845
G: 1603 of Max 10279
B: 1138 of Max 7456
G2: 1691 of Max 7623
Silly me! I just noticed that there is place where EV0 is shown numerically, and that's 2048.
I can see that EV0 is indeed sitting at 2048 on the histogram for all four channels.
But 2048 does not correspond to any of the N5 RGB values... Sigh...
More reading and scratching my head...
-------------
Edit:
"Three stops lower than the maximum pixel value"?
I can see, on the Histogram, that in the case of my NIkon D810, the Max value is 16,000.
My Nikon D810 RAW is supposed to be 14bit so 16,384 value, maximum.
So could 2048 be 18% gray?
Please forgive my ignorance but I was under the impression that 18% gray would be an "invariant"?
I need to better understand this in the context of my particular camera response...
Dear Roger:
Submitted by LibRaw on
Dear Roger:
There is no standard reference point in raw. Most sources suggest to use calibration half a stop lower than 18%, that's, given 18% is about -2.5 EV below clipping ( log2(18/100)), leads to -3 EV (12.5% .. 12.7%, depending on how log2(18/100) is rounded).
First of all, thank you for
Submitted by Roger Breton on
First of all, thank you for helping :-)
BTW, you need to correct the formula on https://www.rawdigger.com/howtouse/calibrate-exposure-meter-to-improve-d...
You have the formula as log2(100/18) = 2.47, this should be log2(18/100) = 2.47 (as you used in the above reply).
Looks like I'm back to basics of photography. (I am curious to know why the photo industry has chose "Base2" for its Log scale historically?).
At any rate, I hunt down the shutter speed and aperture used to make the shot in the EXIF section and got :
f/10 1/640s ISO250 which, when I plug into the EV formula here : https://www.scantips.com/lights/evchart.html give mes an EV of 15.96 calculated in Excel as Log(100/0.0015625, 2).
EV 15.96? That's awfully close tp EV16.
(I wish EXIF would show EV somewhere directly?)
Please excuse my ignorance but is there a relationship between, say, EV16, in this case, and your suggestion of -2.5EV to -3EV? For the purpose of "calibration"? I guess not? Since, looking at your Histogram, the EV scale is graduated from -12EV to + 4EV.
I'm probably mixing "EVs" and "EVs", here...
www.graxx.ca
Dear Roger:
Submitted by LibRaw on
Dear Roger:
EV0 at 3 stops below clipping point (sometimes clipping point is referred to as "full exposure") is a relative, not an absolute (LV) value. This value of -3 EV was selected based on typical characteristic curves.
EV in the formula you've used is probably absolute (LV), and is used to calculate light in the scene, a different matter.
Base2: it was selected, I guess, because normal shutter speed and active lens surface area progressions follow 2x rule.
Use of Log 2 for Photography calculations
Submitted by Stephen Jackson (not verified) on
Roger, that struck me as a great question! I assumed it likely had to do with computers being binary calculators fundamentally speaking. A search on log 2 turned up Wikipedia (a good place to start) which has an article on the Binary Logarithm.
https://en.wikipedia.org/wiki/Binary_logarithm. The article has an outline, see item 4.7 Use-Photography.
Bottom line: EV is based on a human visual response which maps to log 2 as described by the Weber-Fechner law. It didn't hurt that Log 2 is also fundamental to computer science and information theory and thus two reasons for digital photography designers to choose log 2
Add new comment