Zone System for Digital?

I am hoping for some comments on the following reasoning about

trying to apply the zone system to digital photography. It seems to

me that "correct" exposure is crucial to digital. If I overexpose,

then I have highlights that cannot be recovered in RAW converter

such as ACR. If I underexpose, then I risk noisy shadow areas. This

suggests that there is an optimal exposure, one that allows

highlight recovery and minimizes the possibility of shadow

posterization. So I PLACE my higlights in Zone VIII for instance

(+3.0 exposure compensation above the meter recommendation for Zone

V), and my shadows FALL in Zone III, I am optimally using the

dynamic range of the sensor. The question is, how to determine where

the highlights should be placed prior to exposure. In trying

different scenarios, it seems that if the highlight region is

colored (e.g. blue sky) then there is a limit to how far one might

push the highlights up. This also seems to depend on the color space

chosen in the RAW converter and the color temperature applied.

Anyone discovered a way to find the optimal placement?

Histogram.

With our old 35MP 4x5 scan back; we do a prescan; check for highlight saturation; and shadow detail. There is no RAW output available; the unit is labeled as obsolete. One must fiddle with the lighting; and fill the shadows more; tweak exposure; reduce local hotspots/highlights.

My simple method is to set D70 to manual, set aperture, spot meter from a dark area in scene that retains detail (& lock), set shutter speed until the bar on the manual scale (in V/F) is at 0. Then I notch the aperture down two stops smaller (eg. from f/5.6 to f/11).

 

I do not like working like this (matrix metering is easier) but it works better than the matrix metering, especially with taming highlights, in 90 percent of the cases I have tried it.

As stated above, the best QA on your exposure decisions is the Histogram. Although it is much more objective than the LCD image, it shouldn't be the only data considered.

 

Not all highlights belong in Zone 8. You need to get a better sense of what belongs where. Clear blue (north) sky is actually Zone 5, along with green grass, medium reds, etc. Zone 8 doesn't have much detail in a digital image, perhaps none. So place highlights you want to keep in Zone 7 or 7.5, but only if they appear as Zone 8 or higher compared to a Zone 5 area of the subject. This is a more reasoned approach to the usual digital advice to "expose for the highlights." Taken literally, highlights would be placed in zone 5, which would cause serious underexposure in most cases.

 

Featureless white sky, as on an overcast day, does not constitute an highlight. You have several choices: 1) Expose for the foreground and let the sky go white, 2) Use a split-ND filter (you would need 3-5 stops), 3) Make two exposures, one for foreground and the other for sky and merge the two in Photoshop or 4) Try to keep the sky out of the picture (or at least minimize it).

 

Trevor's advice is not very good for digital imaging, since he assumes all shadow detail belongs in zone 3, and highlights will take care of themselves. This works as well as anything in a scene with "average" contrast, which explains the "success" ratio, but fails in the tough situations. I agree, however, that matrix metering has its limits. Spot metering is better for many situations.

Of course the zone system works. However, and this in not different than film, you need to know the exposure characteristics - the dynamic range - of the meduim. Moreover, you need to determine what is the most important part of the scene. IOW, sometimes blowing the highlights is what is needed, other times letting the shadows fall dark is what's called for. Optimum for me may not be optimum for you and much depends on the subject and what you want the viewer to see.

 

Back to the question, I, like the others, prefer to use the histogram. It's much more meaningful to me than the lcd.

 

You may also want to learn blending techniques with digital.

(in camera) histogram and the "blown highlight blinking" thing ;o)whatever it's actually called. Then a little thought on whether those blown highlites should be blown out (ie, street light bulbs, yeah probably....white walls, I doubt it....etc) and adjust (or not) accordingly. Then adjust the shadows in the converter to put them on the left edge of the histogram of the converter. Then if needed, use "curves" to make the contrast where you want it. Whether this is a 100% perfect way of doing it, not sure, but it seems to give me the best detail throughout the picture.

as far as noise in the shadows, really haven't had the problem yet. But, I think dark noise would be easier to photoshop fix, than blown out whites

Perhaps I am mistaken, but isn't the zone system related more to the print than anything else? My understanding of it is that the "negative" is to be exposed in such a way as to ensure the "print" will end up in a certain way. The "camera" in this case, being a large view camera.

 

I put all these in italics because these are the names of Adams' three books on photography.

 

I could be entirely mistaken, but from what I gather, the zone system is essentially meaningless within the context of digital photography, due to a) the extremely small sensor on a digital camera compared to a 8x10 view camera, b) the multipattern matrix meters in most digital (and 35mm film) cameras are impossible to override except when one has a spot meter, and even if you do over ride it, so what? the results are pretty much the same, and c) there is no such thing as "the density of the negative," which, as I understand it, was the whole point of the zone system to begin with.

 

My understanding of the subject is that since prints (in digital format) are not made from a negative, and are indeed entirely dependent on some software manipulation for proper printing, the zone system has no meaning within the context of a print made from an image captured by a digital camera.

 

I am sure if I am mistaken in these assumptions of mine, I will be corrected. Discuss.

Just to clarify on my own question, and comment on responses. The histogram can be a bit misleading if you shoot RAW, since the histogram shows what the image looks like after all processing is done to the JPG file (curves, white balance are applied), so the histogram doesn't tell the story of what can happen in post processing with the RAW file. It seems like there is about 2/3 of a stop or so of highlights that still exist even if the histogram is showing blown out. My original question (perhaps not precisely worded enough) is how do I know where to stop in 'exposing to the right'? I see that even if I clip a color channel I can still recover it in RAW processing without any serious problems unless I really go far to the right by increasing exposure. If I go way too far, I notice that even recovered highlights look pretty strange. What I am trying to do is maximize the amount of information in the RAW file for use to build the picture. I have read that half of the brightness levels available are in the first stop of brightness in the histogram. So if I underexpose by 1 stop I am throwing away half of my data. Is there a standard exposure compensation value that seems like it works in most cases? I am finding that +3.0 works on my Kodak DCS SLR/c. If you notice too, this is the point at which the display on all cameras that I have owned stop flashing overexposure; maybe there's a reason. So I'm suggesting that within plus or minus a third of a stop, put highlights in Zone VIII in digital regardless (bracket around this value). I'll continue to test this myself and I was hoping that someone else had tried this technique as well.

you alter film ISOs for the particular film, develper, and camera in the film zone system corrct? you seem to think that your sensor, histogram, and camera underexpose 2/3 stop. Would it not seem logical to add 2/3 stop to your exposure? bring it into your converter and see if the right hand side of the histogram does indeed rest properly?

 

I don't think you can do a one to one comparison to the film zone system..........actually the Black and White zone system. Just like the so called "color zone" systems for film that exist for both negative and slide films, just take the principals of the zone system and apply them to that medium. You would have to apply the principles to the digital world.

 

Color film, both negative and slide, doesnt come to the level of Black and White film as far as "zone" control. Why would one expect digital color to?

Sorry, the zone system is still a bit to "high" for me. The only thing I might know is: RAW postprocessing is the same as pushing or pulling the film. i was to lazy to shoot something on ISO 200 and ISO 3200 and check out how far RAW tweaking the ISO 200 makes a difference, but I can't understand where there could be one. Any use of the zone system should enable you to get perfect TIFs or JPGs out of the camera or it is just bullshit in my eyes. Yes, I shoot RAW, but as mentioned above I still can't handle the zone system.

been thinking on this a bit more. You would need to actually start with a one to one comparison of digital zone system (dzs) to Black & White film zone system (fzs) for starters and get those sorted out. Knowing full well that all items can not be accounted for at the same steps in digital............at this time, anyhow. I will state these following as fact, but if you think I'm wrong, please provide a description as to where I erred.

 

first.......RAW digital files are the negative film equivelant

 

second......the RAW converter is your digital equivelant to the developing stage in film.

 

third.......Photoshop editting to make the image look good on the monitor and the print (assuming your monitor is calibrated properly, and you use appropriate printers........be those printers, actual print machines or labs), anyhow is like your printing stage in film.

 

before RAW, you took a jpeg file, and "printed" it to suit in PS. This would be the equivalent of shooting B&W film, letting John Doe develop it, and then you printing it at home.

 

RAW puts you back in the drivers seat thru the RAW converters........just like in film (B&W, anyhow) during the development stage.

 

four.......all AA's testing and zone system steps revolved around B&W DAYLIGHT film. So, seeing as white balance in the camera changes the histogram even when in RAW, I think the starting point for WB is the Daylight Setting on your cam...........NOT Auto Balance.

 

This gets us with a baseline of thought and in cam settings to get as close to B&W film in your digital as possible. Let me repeat for clarity (btw - just in case, these terms all relate to a Canon 20D): WB=daylight, Quality=RAW, sharpness=0, contrast=0, saturation=0, color tone=0.

 

I think that covers all the settings, but add as needed, just explain why.

 

five........exposure.....you want to place the whitest block of a black to white ten block strip eventually at 254 in photoshop.........I would think.

 

I'll leave it here.......because 1) I gotta go to work, and 2) I gotta think about the best way to do this for a minute or two.

 

anyhow

 

six......would be development........in the RAW converter. And just like development in the "fzs" you would be adjusting the other end of the strip to the equivelant in the dzs.......the left hand side of the histogram or probably 001 in PS. (BTW, I pick 001 and 254 because if you pick 000 and 255, you may error and actually have them at - 100 and 355.......and not even know it....at least this way you can see it change from 000 to 001 and you know you're actually there.)

 

seven.......printing........or photoshop in digital.........and until digital actually gets to the point of B&W in film, this is where everything else that could not be controlled in camera or in RAW converter steps gets done.

"...five........exposure.....you want to place the whitest block of a black to white ten block strip eventually at 254 in photoshop.........I would think.

 

I'll leave it here.......because 1) I gotta go to work, and 2) I gotta think about the best way to do this for a minute or two..."

 

hmmmmmmm.....at work now.....well Charles.......now I see exactly the question you are asking. I guess the first step is to see exactly, without digital noise creeping in, the range of digital actually is. I think I would still start with the white most block of the strip being matched to the right side of histogram in camera, right side of histogram in RAW relationship, and 254 in photoshop being the baseline. Then when I was sure that was correct, I'd call that f-stop shift in my camera my modified ISO (although, you can't change ISO in that manner in digital, so it would have to be a constant exposure +/- compensation value to make it the easiest to work with later).

 

Then when all that was accomplished, I'd use the RAW converter's histogram to "set" the black point to the blackest block of the strip

and Photoshop to 001. Then I would hold any other judgements until I saw exactly where digital noise came in. I think this would be very camera dependent also. CCD, CMOS, sensor size, processor in camera, etc. And depending on that noise point, I think decisions may have to be split, depending on whether your image warrants letting highlites blow out, or getting them dead on.......and vice versa with shadows, blocked up or dead on. Regardless, I'd be amazed if this all results in a digital cam being able to match B&W white to black range. RAW does allow considerable leeway, but.........like you alluded to, where exactly does the noise issue come in. Where ever that is, it is going to have to be handled in post processing, either in PS or 3rd party noise reduction programs.

"...And depending on that noise point, I think decisions may have to be split, depending on whether your image warrants letting highlites blow out, or getting them dead on.......and vice versa with shadows, blocked up or dead on..."

 

in other words, this would be your expansion and contraction decisions. sort of........a sliding scale, rather than actual expansion or contraction. Again, RAW conversion and increasing or decreasing brightness and contrast is very forgiving, but just how much............that is the question and has to be experimented with.

 

BTW - I'm just thinking all this out loud..........so argue with me all you want.

Thanks to everyone, especially you, Thomas. I think you might be getting at some of the points I was hoping to have discussed. I want to respond to some of your comments but I have this to get off my chest first:

 

I've been doing a bit more testing and I guess I should mention that I'm working in Adobe Photoshop CS RAW converter (I call it ACR for short). What I have done is try to establish some sort of calibration method for exposure like AA did for the Zone System. As Minor White and others recommend in the New Zone System Manual, they take a +0.0 exposure B&W photo of an evenly lighted gray card, then after normal development, compare that print to the actual gray card. If they look the same you are OK. Now, what to do in digital?

 

Something similar but provide a standard that makes sense in RAW processing. I've been thinking that a perfect Zone V exposure should be equivalent to a 127 on the scale from 0 to 255 as seen in the RAW converter when you move your cursor over the frame. Since RAW data is linear in brightness levels, I am assuming that 127 (actually 127.5, or the midpoint between 0 and 255) is a perfect middle gray. If I shoot a uniformly lighted grey card in daylight (i.e. color temp of 5500K) and I spot meter on the card and allow the meter to give the same +0.0 exposure, I should have upon opening the RAW file in ACR after setting white balance to 5500K, a reading of 127 should be obtained, or thereabouts. I have noticed that the color space choice matters for this exercise, as well as the shadow slider (should be set to 0). My in camera meter gives me pretty close to 127 every time in various tests I have done, setting the color space to Adobe RGB 1998. Of course, what you really see in ACR is a triplet of RGB values. Only when I desaturate to grey using the saturation slider moved all the way to the left, do I get a (127,127,127) triplet. When the frame is mostly blue, like a frame taken of the sky, the Blue value (the last value in the triplet) is always the highest. Therefore, the key to setting the optimal exposure depends on whether one of the color channels will go to saturation first. I could shoot a perfect blue card for instance and see how far I can push up the highlights on this frame by using the exposure slider, and record that as the parameter to be used in the field for blue sky, for instance. I could also do that for green and red.

 

So now I have a calibrated reference point for middle gray at least in ACR, and I will know how far I can go above this value for highlights. So all I do is determine SBR (subject brightness range), place the highlights in the optimal zone for it's color characteristics, and then I'll know where the shadows fall in terms of zone placement.

 

I also conducted another test. I shot a sequence of frames at different exposures. I selected the one that was 2 stops below the +0.0 frame. This is zone III if my meter is calibrated correctly. I opened that frame in ACR and tried raising exposure on it. I found that if I view at 50% (this is arbitrary, but a decent view of any frame) and I raise the exposure by more than +1.33 (one and a third stops) then I see posterization, or at least some pretty bad noise artifacts. So my shadows are pretty safe in Zone III. So I figure I have 6 stops of dynamic range to play with at a minimum.

 

In summary, I can get an optimal exposure for post-processing if I follow the rule that I use that first stop of brightness in digital photography. If my subject scene has 6 stops or less, I am assured of having a RAW file that contains the maximum amount of information. So basically digital photography is more about gathering data than trying to get the final exposure that you would like to see in the print (or on the monitor screen).

 

Sorry so long winded. I hope this might spur some further discussion.

I now want to respond to some of the other photographers comments. Yes, I am trying indeed to create an analogue to film processing that AA established, but more than that. This analogy does not break down if the sensor is smaller than a large format. Actually, in 35mm roll film the Zone System could not be used 100% since each frame could not be separately developed. In digital, they can be. Now that I'm 100% digitally re-mastered, I want a methodology for digital photography that is as comprehensive and logical as the zone system is/was regardless of what we might call it. When I go into the field I want order not chaos, since I want to concentrate on 'art.'

 

Thomas, you are absolutely correct in saying the RAW file is like the negative in film and using ACR is like film development. In fact, I think the designers of the RAW software packages that are out there want us to use them with similar techniques as with film. Indeed, I can take a long SBR frame (one with a wide range of brightness values) and decrease the contrast in ACR, very much like reducing development (N+1, N-1, etc, was the terminology in film). But I am moving a slider and not bathing in chemicals.

 

Yes, 254 would be the closest value to 255 and not theoretically be 'blown out.' Then the question is how much exposure gives a consistent 254 (or a bit lower, since we are working in one third of a stop steps on most cameras) is the root of my original question. I have done some testing towards this and it's complicated by several factors. Your white balance has to be spot on in ACR for the shot. If I move the white balance slider, I can blow the highlights or recover them as well. I also can recover highlights if I move to a larger color space such as Pro Photo RGB. I'm finding that I want to move the white balance away from what the camera shot to give a warmer color balance. If I do so, I'm risking blowing out the red channel. So I've got to take that into consideration when I expose in the field.

 

Your other comment about letting some highlights blow out to avoid blocked shadows is also correct IMO. You can sacrifice H (highlights) to allow S (shadows) to be increased in exposure, if H is small and/or unimportant and S is large and important. Alternatively, you can let S go to black if S is small and/or unimportant and H is large and important. These are rules I try to follow when the SBR is larger than 6 stops. Or, I reach for a grad filter if appropriate. Someone mentioned that solution, as well as digital blending. These are useful of course, but my purpose here is to find the optimal point before any remedies such as these are employed.

 

I'd say the print phase is what you do in Photoshop proper, after the RAW data is converted and the white balance is set, whether the output is a TIFF, JPG or whatever. At this stage you might do sharpening or play with saturation. But the 'negative' has been translated into values you can see like a photographic print. It is no longer just data that your sensor collected. So if you shoot JPG, the camera throws away the negative and you are left with 'the print' whether you like it or not. Perhaps not a perfect analogy but close.

 

The bottom line here is that testing is needed if you want to find out where the optimal points are for your camera. My sensor may not behave like yours. I do think that exposing a middle toned subject as middle toned can be used if you are staying with JPG. If you use RAW, then I think that that methodology isn't optimal.

"So basically digital photography is more about gathering data than trying to get the final exposure that you would like to see in the print (or on the monitor screen)."

 

after playing around......no detailed testing like yours, mind you...but just experimenting arbitrarily, and seeing how much I could alter things in the RAW converter (Canon's DPP in my case...although PSCS is going to be eventually bought for this), from different "in camera" exposures of the same scene, I was beginning to come to that conclusion also.

 

Glad to see someone else is leaning in that direction also. There might be something to further pursue there.