Practical application of metering basics

[Colin Knight]

Help me out here please. What I think I know in theory doesn't seem to work in real life. I would like to use two scenarios, one with wildlife and one with landscape photography, and you tell me where I'm wrong. Both scenarios have to do with spot metering, how the camera renders the image in middle gray, and the appropriate correction required by the photographer. Both cases assume manual exposure control, and spot metering.

Landscape scenario: a given composition includes a waterfall, the associated cliff face, trees, and blue sky. If I wanted to apply the zone system to capture the image, I would spot meter on the white of the waterfall. With no correction from me, the camera would underexpose, rendering the waterfall middle gray, about a zone 5. But if I want the waterfall to be a zone 8, I would open up 3 stops, from the initial exposure that the spot metering read. This would result in the highlights of the waterfall being properly exposed and preserved. Is this correct?

Wildlife Scenario: Just say I'm photographing a Raven, or anything that's darker than middle gray. If I spot meter on the subject, the resulting reading will overexpose. I will have to stop down to some degree to get a properly exposed image. Is this correct?

The overall concept I have is that if something is brighter than middle gray and spot metered, the camera will underexpose, because it's trying to make the scene middle gray. And the reverse for subjects darker than middle gray. My examples above just restate this in more words, but the practical application may be where I'm off.

If my understanding is correct, then my metering is the problem. If not, please tell me where I went wrong. :)

[Roger Clark]

The overall concept I have is that if something is brighter than middle gray and spot metered, the camera will underexpose, because it's trying to make the scene middle gray. And the reverse for subjects darker than middle gray. My examples above just restate this in more words, but the practical application may be where I'm off.


Colin,
You have it correct. Although the 3 stops correction on the waterfall may not be quite right (it would depend on the waterfall). But fortunately in the digital age you can check the histogram and for any "blinky" overexposure pixels to be sure the exposure is correct. Knowing metering and possible corrections gets you to the optimum exposure faster, and sometimes even on the first try.

I'm not sure which camera you are using, but if it has separate red, green and blue histograms, turn that mode on and use all three histograms. Sometimes on strongly-lit subjects, one color can be saturated when the luminance histogram shows no saturation. This is important in post processing too: check full color histograms.

Roger

[Alfred Forns]

Hi Colin

If you were working with Black and White film you would be more or less right. A zone 8 would be white but with little or no texture, the waterfall would best be rendered Zone 7 ... with the equipment you are using opening up on whites three stops will overexpose the image by one and one half. The proper exposure would be one and one half.

Some people (not many) will recommend spot metering. I would stay away from it, use the camera multi matrix. For using spot you need to know what you are doing and it is not as fast. Also the spot meter in your camera is not a real spot meter ... would consider a one degree Digital Pentax Spot Meter the real thing .. you get one degree !!

With the matrix meter it would be the same, if the scene is middle gray shoot at meter reading but if its brighter open up or close down for darker. With time you will be able to judge the precise exposure, does not take long since you have the histogram to guide you. You do want to learn getting it right the first time since situations will arise with no second chances..... also consider the size of the subject in relation to the bg.

Might want to take a look at Arties book, can't imagine a better tool for learning exposure other than attending a workshop.

[Colin Knight]

Thank you both!

Roger, I do have the RGB histogram set to display when previewing images. I use it every time. I've heard the luminance histogram on certain cameras may only be reading one channel. And you're right, this is the ultimate guide when determining the correct exposure! Maybe I'm over-thinking it.

Alfred, thanks for the tips and recommendations! Greatly appreciated.

[Desmond Chan]

Landscape scenario: a given composition includes a waterfall, the associated cliff face, trees, and blue sky. If I wanted to apply the zone system to capture the image, I would spot meter on the white of the waterfall.


I think if you really, really apply zone system, you should meter more than one spots to find out out the dynamic range of the scene. Then you decide where you put the zones (the waterfall may not end up in Zone VII) and how you process the "film" afterwards to get an image the same as the one you envisioned at the scene.


But if I want the waterfall to be a zone 8, I would open up 3 stops I agree with Al on opening up only two stops to put the water on Zone VII to be on the safe side. I personally could add a bit more than two depending on the scene.


Wildlife Scenario: Just say I'm photographing a Raven, or anything that's darker than middle gray. If I spot meter on the subject, the resulting reading will overexpose. I will have to stop down to some degree to get a properly exposed image. Is this correct? Correct as indicated by others. I would close down one or two stops in this case.

[Alfred Forns]

Hi Colin

Forgot to mention .... Ansel did come up with a zone system for color but never was popular. The range was much smaller, just over four stops and not much practical.

One thing that works fine is metering on snow and opening one stop and a half .... don't get to do that much around here :)

[John Chardine]

Also, Al, could it said that the zone system was designed for film? In the digital age if you shoot RAW your optimally exposed image out of the camera should look washed out with no blinky highlights, which you can then adjust by sliding back the exposure, in post processing. Ultimately then, the optimum exposure setting for a digital RAW image would differ from what the zone system would tell you if you were shooting film. Does this make sense?

[Alfred Forns]

Yes John it was for film and Black and White at that !!!

Some people like to have the mid tones at the middle of the histogram while others like to push them to the right. Reasoning there is more information on the last box to right in the histogram than the all the previous.

I've tried it both ways and both seem to work well, no big difference for me. Maybe Roger can shed some light on this mater. I'm exposing to the right without going over.

[Colin Knight]

John and Alfred, since the camera's meter isn't reading color, why would the zone system be different from color to black and white?

[Alfred Forns]

Because the the sensor can not record the same range of tones as black and white film !!!

The range for color is around four stops only !!

.. you are reading tonal ranges !!!

[Roger Clark]

Yes John it was for film and Black and White at that !!!

Some people like to have the mid tones at the middle of the histogram while others like to push them to the right. Reasoning there is more information on the last box to right in the histogram than the all the previous.

I've tried it both ways and both seem to work well, no big difference for me. Maybe Roger can shed some light on this mater. I'm exposing to the right without going over.

Hi Alfred,
Because digital camera sensors respond linearly to light, and with DSLRs with good sized pixels which collect a lot of light, being off a little in exposure only changes the noise a little. So I would expect no big difference, especially in raw converted data. In the raw converter, adjusting exposure is simply a multiply of the data and moves the measured signal up or down that linear response. When the data are converted to an image on screen and then a file, a variable-gamma tone curve is applied in software that to first order mimics the response of print film. Finally, our noise perception is dependent on both the tonality and fine detail; e.g. noise is more apparent in some skies but harder to see in a tree.

So when I do my exposure, I do try and expose to the right to increase the signal-to-noise ratio, but I'm not obsessive about it. I would rather be a little under exposed than saturate something. If you underexpose by one stop, signal-to-noise ratio only drops by root 2, or 1/2 stop, and the exposure error is easily corrected in the raw converter.

To make it a little shorter, your experience is well described by the data and the statistics of noise.

Roger

[Roger Clark]

John and Alfred, since the camera's meter isn't reading color, why would the zone system be different from color to black and white?


Because the the sensor can not record the same range of tones as black and white film !!!

The range for color is around four stops only !!

.. you are reading tonal ranges !!!

Alfred,
I'm not sure this is correct, and the reason is a little obscure. The noise and characteristic curves of film are measured with a 48-micron diameter spot (1810 square microns). That would be like comparing a 1D Mark IV (5.7 micron pixels; 32.5 square microns) by averaging 55.7 pixels. Averaging that many pixels would increase the dynamic range 2.9 stops, adding to the 11.3 stops/pixel (at ISO 100) would give 14.2 stops. I don't think most black and white film has that high of a dynamic range.

Were you referring to the color four stops for color film? Digital has the same dynamic range in each color, so I would think the zone system for black and white might be reasonable for color digital. But even if so, do we really need the zone system any more? We have histograms and much finer control in processing and can choose any shape to the tone curve we want. That can take an image way beyond what could be done with film and precision development.

Roger

[Desmond Chan]

do we really need the zone system any more? We have histograms and much finer control in processing and can choose any shape to the tone curve we want. That can take an image way beyond what could be done with film and precision development.



If I may jump in, I think the answer to the question: "do we really need the zone system any more?" has always been: "No." Even in the film days, many people did not use or care about zone system and still produced good black & white images.

To me, zone system is a systematic way to produce an image so that you know what you'll get, what your print will look like right from the moment you press the shutter button. In my opinion, this system worked then and it should still work now with digital the digital cameras and computer.

I know many of you test the AF of your camera and lenses. I wonder how many also test its color profile and exposure to see how accurate the camera can produce colors and how accurate its exposure meter is.

[Alfred Forns]

Hi Roger

Thanks for the explanation, you hit the nail on the head with this statement "I do try and expose to the right to increase the signal-to-noise ratio, but I'm not obsessive about it. " Best way to go !!!!


Regarding the total range captured by a digital camera, I would try measuring by using a light meter? With a one degree Pentax meter for example ... using a black and white duck, meter the blacks then whites, not sure how many stops, at mid day guessing five or so.... if the MKIV had 14.2 stops it should be able to render both? Maybe is just the way it works out in theory

Either way we are spoiled by being able to change the ISO and having feedback from the histogram !! ...tough to go back to film :)

Desmond we do use the basics of the Zone System today, the key to the system was pre visualiztion and we still do !!! We have particular look we want and after converting and processing in PS we get it !!!

[Roger Clark]

Regarding the total range captured by a digital camera, I would try measuring by using a light meter? With a one degree Pentax meter for example ... using a black and white duck, meter the blacks then whites, not sure how many stops, at mid day guessing five or so.... if the MKIV had 14.2 stops it should be able to render both? Maybe is just the way it works out in theory


Alfred,
Yes, and it would be interesting to do the same image with B&W film. I have shot a little B&W 35mm film (did do a lot of 4x5 B&W), mostly color slide where the dynamic range is really low and we all got great images from such a small dynamic range (5-7 stops) so it seems strange that people complain about 11+ stops per pixel of DSLRs.
Here is a page on exposure latitude of a 1DII along with some print film data for comparison. I will try it again when I get a 1DIV (and then have some time).
http://www.clarkvision.com/articles/exposure_latitude-1/

Also, if one is converting dslr images to B&W, usually the red and green channel signals are not as high, so saturate well after the green channel. It is that property that allows highlight recovery in raw converters, at the loss of color info. But in converting to B&W one can use that highlight recovery to extend dynamic range more, probably about 3 stops. Though it would be hard to exploit all that range with the exact exposure in the field.

In my opinion, if the light is good, dslrs have adequate dynamic range, and even some to spare, and I think that is proven by all the really spectacular images on BPN. We rarely hear complaints of inadequate dynamic range in image critiques. And if the light is bad, more dynamic range is not going to make it a great image.

Roger

[Kaustubh Deshpande]

Great stuff in here from some real gurus. I will have to spend more time to digest it. Thx folks for sharing the wealth.

Colin, I used to do those kinds of things when shooting velvia slides. Landscapes mostly. Getting the exposure right was so critical then compared to today. And with landscapes, where you can bracket and take as many shots as you want, things are much easier now. With wildlife and birds, yes, its still very much important to get the right expsoure ASAP.

I find evaluative metering much more practical these days. If you know your camera well and know in which conditions to dial in extra exp. comp to expose to the right, things do turn out pretty well. I usually start off making some test shots of the subject( with evaulative metering) just to figure out expsoure. Look at the histogram...all three channels....and then make the exp. comp. adjustments. Many times, when exposure to right is not possible( say, handheld camera, fast SS to freeze motion, iso limitations), I'll make the sure I have SS needed and the acceptable ISO....and let it underexpose a bit. Better to have an underexposed shot than one with camera shake. my 2 cents.

[Colin Knight]

Very good discussion, very helpful! Thanks for chiming in.

Kaustubh, thanks for putting it into a modern perspective too.

Colin

[Alfred Forns]

Thanks Roger !!! ... btw need to give the link a second read, .... good !!!!

[Charles Glatzer]

Different strokes for different folks.

Joe McDonald, John Shaw, myself and many other pros prefer to use Spot metering. Your in camera spot pattern size may be larger than 1 degree, but it is a spot pattern nonetheless. Spot is actually easier in many circumstances than Eval, especially when the subject varies in size and tonality from the background. I recommend learning all the metering methods available (Spot, Eval/Matrix, Incident, etc), as no one method is best in all conditions. Additionally, learning how to take and use reference values will greatly speed up your exposure process, and go along way in improving your photography. Eval's job is to render the scene as viewed by applying varying degrees of compensation based on pre-determined scenarios, not to render the scene as a mid-tone, although it may bias the exposure as such.

The more crayons in the box the better ;)

What matters most is that the photographer understands the ramifications of using each pattern and method, exploiting benefits and recognizing detriments as quickly as possible to obtain the desired result.

Best,

Chas

[Charles Glatzer]

Better to have an underexposed shot than one with camera shake. my 2 cents.

I think it is most often better to have a bit of noise increasing the ISO than underexposing the image.

Note-Canon does over/under expose and process in camera when using intermediate ISO in .3 stops.

Best,

Chas

[Reza Gorji]

Great discussion. Trying to digest all this stuff. Not to get sidetracked but can one of you explain to me why it's easier to recover parts of an image that are underexposed when compare to overexposure or the highlight end of the histogram?

Thank you