Metering, some myths explored

Myth 1: One myth we have is 'Meters are calibrated to 18% grey'.

Fact 1: It is correct to say that the 18% grey tone is 'at the center ('median', not 'mean') of the range of brightness found in the average scene.

Here we see a 26 zone range of brightnesses, from White to Black

In the Zone System practiced by Ansel Adams and others, we had a 10 zone range of brightnesses, from White to Black.

The 'center' in the 26 zones is 'M', the 'center' of the Zone System is known as Zone V (roman numeral for '5'). The name (label) is not important, the fact that it is a 'middle tone' is the important concept.

Fact 2: If one compares the ISO calibration equation for an Incident Light meter, and the ISO calibration standard for a Reflect Light meter, the two have equivalent exposures if you meter a 12% grey card with the reflected light meter!

Here is a pair of exposures of an 18% grey target (center zone of target with Black, 18% Grey, and White zones),

the first shot 18% target metered by the Canon dSLR, to reproduce an 18% grey mid-tone in the center of the three-toned target ...

Now here is the result of metering a 12% grey card, to reproduce the 18% grey mid-tone in the center of the three-toned target

Myth #2: 18% grey is the middle of the tonal scale, and the exposure recommended by the meter (when reading the 12% grey tonality which represents the calibration density used by the ISO standard) puts 18% tonality in the center. (see Myth #1, explored above)

Fact, shot A: First is a shot composed of a pattern of alternating black squares and white squares printed on glossy photo paper, representing the extremes of the tonal range, so 18% tonality is reproduced to be 'in the middle' of that range. Illuminated by full overcast sky, the Canon was put into full frame Averaging mode for this shot, and exposed as recommended with no exposure compensation adjustment (ISO 400, 1/250 f/5.6:

Just to prove that the 18% grey card is centered between 'black' and 'white' of my target pattern, I decided to include the grey card in the photo so three spikes can be seen in the histogram, when exposed per my Minolta Autometer Vf incident meter.

images 2 and 3 in first post aren't of a grey card!?

One issue that can trip things up is what physical method is being used to meter the light, what the nature of that light is, and the nature of the material reflecting said light. This can be further compounded by the cell response to a given wave length of light, whether the biological cells in our eyes or the individual photo-sites on a digital sensor.

Light doesn't actually have a colour, it has various waves of energy across a continuous range of wavelengths, and it is not always an easy thing to have a perfect response curve to all of those various possibilities. The really fun thing is that those waves also aren't confined to just the wavelengths we can see. The method used to measure the light in one device may differ from the method used in another, and one device may respond just a little more to red vs blue, even if they're effectively very similar levels of light overall for example.

Depending on the tools you're using to measure the levels of light you can have your readings skewed a bit if there is an unusual amount of infrared or ultra violet light hitting the sensor, and two different sensors may read those values differently.

I know that my hand held light meter tends to read a little hot based on UV light compared to what both of my 7Ds read off a general grey card. How light reflects and the full spectrum involved can also mean that if you meter off one source, like a known grey card, and then try to meter off a 'neutral scene' in the same light, you may actually get a slightly different reading simply based on differences in UV/IR and general response curves of everything between them.


Of course then there is also the issue of calibration itself. Is the device you're using actually working properly? What factors are muddying the measurements, such as temperature of the device and level of a reference voltage? Has it drifted one way or the other since it was actually calibrated to some kind of reference, and how precise was that reference?

So the long story short: Practice with your tools and learn how the ones you have in your hands actually respond and work. They may not match up with what documents and data sheets claim they're supposed to.

Martin Dixon wrote in post #17614392
images 2 and 3 in first post aren't of a grey card!?

The CENTER AREA of the target (PhotoVision) is indeed 18% grey. That center area was used for Spot mode metering to shoot the first shot.

The 18% target reading was ADJUSTED by +0.33EV, the difference in reflectivity between 12% vs. 18% grey target, to make the second shot as I did not have a 12% target.
Lastolite makes an 12% target as well as an 18% target, but oddly it is impossible to buy the 12% target in the US, as their distributor in the US does not market them.

The images have changed now! Probably a blip with photobucket

I won't pretend to understand everything in this thread, but I have noted that there seems to be a lack of threads discussing metering.

I believe it's been demonstrated (and pointed out in Canon literature) that Canon meters and exposures don't "fixate" on 18% grey but rather on 12% grey, although I'm not as adept at the technical issues as Kirt is.

However, in practice, I'm aware that Auto Exposure can fall short, and so as a rule I'm ready to adjust, either via Manual exposure or adjusting the Exposure Compensation. This also a good argument for shooting RAW and exercising the Expose To The Right method, since if you blindly follow the meter, you may not get the desired results!

tonylong wrote in post #18104266
[..] if you blindly follow the meter, you may not get the desired results!

Actually, yes you will, if you follow an incident meter. If, on the contrary, you let the camera make the decisions ('full Auto, man! The circuitry in this thing is more modern and smart than those archaic hand-held relics!'), you will not get the desired results a lot of the time.

As Wilt's experiments shewed, the incident metering was spot on (pun intended, you'll see why below) every time. It seemed to be off once, when metering the chequered pattern printed on glossy paper, but it wasn't really: you don't expect solid blacks from a glossy surface that is reflecting light, however diffused.

Here's a better explanation than I can provide: Black is light, White is dark (!!).

Spot on. Spot on. I mean, if you don't have a hand-held meter, then get your camera out of those silly übersophisticated and unreliable modes and set it to Spot Metering, then meter off a surface that is a mid tone, then use the readings the camera suggests.

Here, help yourselves to a tonal value cheat sheet.

Alveric wrote in post #18104658
The fabled ETTR method is another thing I find disingenuous. Whilst its theoretical foundation is sound (namely, that sensors record most of the data to the right of centre, ergo 'expose to the right' in order to keep as much detail as possible), pushing the exposure yields you an image that's not a correct representation of the actual image your eyes saw, but it's overexposed, even if it's just a mite so. Now, if you are able to remember the actual view for every image (such a photographic memory indeed) and pull back the exposure in post-processing, all's well, methinks. Otherwise, the photographs are not particularly accurate.

^^^ my sentiments exactly

Alveric wrote in post #18104658
The fabled ETTR method is another thing I find disingenuous. Whilst its theoretical foundation is sound (namely, that sensors record most of the data to the right of centre, ergo 'expose to the right' in order to keep as much detail as possible), pushing the exposure yields you a picture that's not a correct representation of the actual image your eyes saw, but it's overexposed, even if it's just a mite so. Now, if you are able to remember the actual view for every image (such a photographic memory indeed) and pull back the exposure in post-processing, all's well, methinks. Otherwise, the photographs are not particularly accurate.

Agreed, but I think there might be a difference between the result you get if you:
1) Evaluate the metering and decide you need +1 stop to get the exposure right AND overexpose +1 stop to ETTR and then in post pulls exposure back by -1 stop.
2) Evaluate the metering and decide you need +1 stop to get the exposure right.

I don't usually use ETTR myself, but my understanding of the concept is that #1 should reveal more detail in the shadows than #2.
I usually try to get exposure right in the camera and can't really tell if ETTR really does make a difference, but guys with far more experience says it does, so I am inclined to think they are right.

Larry Johnson wrote in post #18104117
I won't pretend to understand everything in this thread, but I have noted that there seems to be a lack of threads discussing metering.

I don't understand either, but this thread is much more effective than any sleeping pill on me!
As for metering, I just use S16 or take test shot to make sure what exposure is right.
But I really appreciate this thread and want to see more scientific research reports on this issue here. I have it bookmarked for tonight!

I also use Sunny 16 or Moony 11, especially when shooting film and there's no hand-held meter. Consistent exposures all the time.

D-Noc wrote in post #18104803
Agreed, but I think there might be a difference between the result you get if you:
1) Evaluate the metering and decide you need +1 stop to get the exposure right AND overexpose +1 stop to ETTR and then in post pulls exposure back by -1 stop.
2) Evaluate the metering and decide you need +1 stop to get the exposure right.

I don't usually use ETTR myself, but my understanding of the concept is that #1 should reveal more detail in the shadows than #2.
I usually try to get exposure right in the camera and can't really tell if ETTR really does make a difference, but guys with far more experience says it does, so I am inclined to think they are right.

Ahem, what you describe is essentially the methodology we should follow when using reflective meters (either the camera's or the hand-held meter's spot). Your 'evaluate the metering' strikes me as basically finding the mid tone and reading off it to set the exposure. Mid tone is mid tone, and should be rendered as mid tone, otherwise you no longer have an accurate photograph. It also strikes me as compensating for memory tones, namely, white snow should render as white, so if you meter off of it then you need to up the values by ~2 stops to render the snow accurately (i.e. white, not gray [assuming it's sunlit, for there is indeed such thing as grey snow, like when the sky is overcast and it's late in the day]).

No, no, the whole ETTR paradigm, which is gospel-like dictum for those obssessed in having every bit of detail shewing –however inane and irrelevant it is– is a bit of a different animal. It's right up there with the whole HDRI craze and its flat photos. And yes, it's been even featured, acclaimed, and recommended in reputable photo magazines (even books I'd wager), which gives it more credibility. Alas, even very experienced photographers can be, well, wrong.

Take an image like this one:

IMAGE LINK: http://www.ipernity.co​m/doc/diamantstudios/4​2852536