Digital Photography and the Zone System

[girod]

Each step in this lesson is progressively exhilarating for me. Thanks a lot Keith.

I opened a NEF file (with in-D700 custom linear tonal curve) in CNX2, retained the linear curve in CNX2; applied shadow and highlight recovery optimally, applied a conservative S-shaped tone curve and then slid the midtone lever very slightly to the left of the histogram upto a value around 118 - I am very pleased with the result. Is this right, keeping the midtone value at around 118 is most important?

girod

[keithsnell]

I opened a NEF file (with in-D700 custom linear tonal curve) in CNX2, retained the linear curve in CNX2; applied shadow and highlight recovery optimally, applied a conservative S-shaped tone curve and then slid the midtone lever very slightly to the left of the histogram upto a value around 118 - I am very pleased with the result. Is this right, keeping the midtone value at around 118 is most important?

girod

Hi Girod,

Keeping the mid-tone value at around 118 is important if those tonal values are an important part of the image.  If you were using one of the tone curves that shipped with your D3 and were not aware that NX2 shifts the mid-tones to a brighter position on these tone curves by default, then you might incorrectly assume that your images are "over exposed."  As I mentioned, this is especially evident in portraits, where the skin tones might look too "hot" or over exposed.  It is important to recognize that these too-bright mid-tones are mostly due to the "tonal shift" and not because of over exposure.  That knowledge, along with the knowledge of how the default tone curves (aggressive S curves) and shadow and highlight protection functions are affecting your image will enable you to optimize the rendering of your captured image.  

(As an aside, the optional D2X tone curves do not apply this tonal shift, which is why many portrait shooters prefer the D2X tone curves over the ones that shipped with the D3.  Nikon recommends setting the "brightness" to -1 when using the D2X tone curves in order to match the "out of camera" rendering from the D2X.)  

Notice that I said the perceived over-exposure is "mostly" due to the tonal shift.  NX2 (not the camera) also applies what some people have called a "hidden" exposure compensation of about +0.5 EV to the image.  Therefore, using a linear tone curve corrects for the "tonal shift" but the mid-tone will still be rendered slightly brighter than you "exposed" it.  If your goal is to correctly render the mid-tone at 118 in RGB, you will need adjust the midpoint in post processing, or expose with -0.5 exposure compensation.  (The -0.5 EV exposure compensation correctly places the mid-tone, but still implements the "tonal shift" of mid-tones in relation to other tones in the image.)  In general, I find that the -0.5 EV exposure compensation often results in the most "pleasing" rendition of the image.

In the "old days" of digital processing, the built-in tone curves in Nikon cameras were much "flatter" (like the linear tone curve) and much more "accurate," without the tonal shifts, brightness adjustments and aggressive S curves that you are seeing with the "default" curves in the D3 and D700.  The reality is that these "flat" tone curves didn't look pleasing right out of the camera, which made post-processing necessary for almost every image in order to make it look pleasing to the eye.  Nikon has incorporated years of experience into the "new" tone curves, automatically applying many of the "corrections" that used to be required in order to produce a pleasing image from a raw file.  I used to have to "brighten the mid-tone" on many of my images to make them look appealing.  Now this is applied automatically by what some call "hidden exposure compensation" and the tonal shifts in the "in camera" tone curves.  The end result is that the majority of images do in fact look better "out of camera" than they would have in the "old days."  However, if you truly want to be in control and "optimize" the output, then you need to understand what is going on "behind the scenes."  

Note that much of this discussion is predicated on the belief that "proper" placement of mid-tones is 118 in a gamma 2.2 RGB space (sRGB and AdobeRGB).  This is based on Ansel Adams' placement of "mid-tone" at 18% gray, which equates to a luminosity of 50 in Lab space, which equates to 118 in RGB gamma 2.2 space.  I'm not 100% convinced that "mid-tone" should be 118 in a gamma 2.2 RGB space.  Notice that the "mid-point" in your curves adjustment in your raw processing software is 128, not 118.  This is true for NX2, Photoshop, and every other image processor that I am aware of.  I believe that if Adams were alive today, he would be an avid digital photographer, and would choose to define 128 as "mid-tone" in a gamma 2.2 RGB space.  Defining "mid-tone" as 118 complicates matters unnecessarily, and causes us to perform mental gymnastics that are not necessary if we accept 128 (just +0.5 EV brighter on the scale) as the "mid-tone," as Nikon seems to have done in their latest generation of digital cameras.  For now, I will defer to the popular opinion that "mid-tone" is 118, but recognize that many image processors treat 128 as the mid-tone when applying curves adjustments, etc.

The "most important" thing we have talked about in this discussion is "proper" exposure of the initial image capture in order to optimize the quality of the raw data.  That said, you also need to understand the default "behavior" of your raw processor, and how to adjust this "behavior" to optimize the image rendering to match the way you intended to render the image.

I hope this helps.  I know there are many "variables" to consider during the image capture and post-processing; however, understanding the behavior of our cameras and software will help us optimize our image capture and rendering.

Keith

[girod]

Hello Keith,

This whole lesson on knowing the principle behind proper exposure is a revelation for me. As a hobbyist beginner, it lifted a tremendous barrier in my understanding and practice of photography.

All I have to do now is shoot as much as I can so I will know my D700 (and lenses) better.

You are a Great Teacher Keith, thank you very much.

girod

[keithsnell]

Hi Girod,

Thank you for sticking with me throughout the discussion.  The discussion was helpful for me as well, since it helped to reinforce my understanding of how to optimize the exposure and post-processing for images shot with the D3.

Happy shooting,

Keith

[Shane]

I realize this is an old topic but I think my response/question is pertinent to the original post. If not feel free to initiate a new posting.

I have been trying to get organized with my different DSLR bodies with respect to exposure "calibration" and highlight clipping. I have been using UniWB and Linear Tone Curve since 2007 and developed a "gut" feeling for exposure settings and clipping based on experience and the histogram. However, this worked well for the initial few years when I only maintained two different bodies but now with several bodies and different wavelength conversions it was time to get organized and perform some "calibration". This is when I dug this very useful post up from my notes and decided to try a structured approach to the problem.

First some background. I am working with a D700, D300S, D200, D200UVIR and D200IR 830nm. I currently run linear tone curve and UniWB on all of these. Some of the UniWB files I have created from scratch and (due to laziness) some have been downloaded. Typically I use Rawnalyze, Histogrammar and exiftool to help with evaluating RAW data. Being newer (and less familiar) I decided to start with the D700 and D300S. This where the first "issue" arose.

Test was performed outside on a "medium" bright but solid overcast day. I used a Sekonic L-558 to meter the Kodak grey card before during and after the test to verify that lighting was not changing.

Based on evaluating exposure and clipping using Rawnalyze, with the D300S it appears that clipping starts to occur at +3.7EV over the metered grey card. I also ran the D700 and got a similar result of around 3.7EV. However, I need to rerun the D700 as the light started to change slightly during the test.

Based on Rawnalyze, the 0EV exposure is around 3.7EV below the G clip point which is also consistent with my final result where clipping actually occurred at just above +3.7EV .

This was in contrast to the +2.7/3.0EV reported clipping point in your D700 test. Although I will rerun the D700 when lighting is more consistent I don’t expect to see much of a change. Any comments would be welcome.

Shown are D300S at 0EV, +2.7EV and +3.7EV. However, preliminary tests for the D700 seem to be quite similar.

[keithsnell]

Hi Shane,

How were you metering?  Spot meter with the camera's meter?  Spot in the center of the frame?  Was the camera analog exposure scale showing 0 for the 0 EV shot?  You verified no exposure compensation was set?

It's always tough to determine why there might be differences when I can't observe your test procedures.  It doesn't seem like your results are consistent with what others have reported (and what I have observed), but at this point I'm not sure why. 

Keith

[Shane]

Hi Keith

ISO 200, f/8, Spot metered (center spot), zero on the analog meter for the 0EV shot (same meter reading from the Sekonic), Exp Comp. zero, 70mm lens, slightly defocused.

The only difference I can see with regards to your test is that I am using a UWB but this shouldn't result in that kind of difference*. You might notice that even our exposure times are quite similar and by coincidence I realised that our focal length was identical.

* Unless perhaps the lighting environment is extremely blue or red rich (afterthought - I am thinking of perceived exposure not analog metering). You can see from my Rawnalyze that this is not the case.

The 0EV shot as shown in Rawnalyze is very important. As you can see it sits about 3.7EV below the saturation point. It would be helpful if you could show your 0EV shot for the D3, based on your results it should be 2.7EV below saturation.  I also noticed that you specified to use a Daylight WB and to avoid Auto WB however your initial Rawnalyze images show that an Auto WB was used, later images indicate Manual WB, was this an oversight?.

I found your overall approach to "calibration" to the Zone system quite interesting and will be continuing with that but would like to resolve this issue first..

[keithsnell]

Hi Shane,

I'll see if I can dig up the 0 EV shot.  I'm not sure if I still have that series on the computer, but I'll look.

With respect to the white balance setting, it definitely has an effect if you are using an external raw editor (for the white towel shots, etc); however, since Rawnalyze enables us to look at the files without WB multipliers applied, I don't think it's much of an issue when examining the files in Rawnalyze.

I'll see if I can dig up my old files from this test.

Keith

[girod]

Hello Keith and Shane,

Iliah Borg have just published their new program to analyze raw file and meter calibration: rawdigger.com or here: http://www.luminous-landscape.com/forum/index.php?topic=62739.0

With my D700 and D7000 + RawTherapee (v4.0.7.1, Windows 7, 64bit) raw developer, I too have been using +3.7EV for Zone VII; AWB and I don't rely anymore on the in-camera Histogram.

[Shane]

Girod
thanks for responding. I did notice the release of the new rawdigger but haven't downloaded it yet. I prefer to wait a little usually until the bugs have been sorted. How do you find it compared to Rawnalyze?

[girod]

Shane,

I have not tried rawdigger yet.

[keithsnell]

Thanks for the heads up on rawdigger.  I'll have to give it a try.

The objective is to get consistent results in a way you are comfortable with, so if +3.7EV for Zone VII works for you then go for it.  Personally I find that +3.7 EV results in too many blown highlights in "real world" shooting, but you should use whatever method gives you consistent results that meet your requirements.

Shane, I looked for the old D3 files on two computers and didn't find them.  I'll try to find time to dig out my old laptop and look for them there.

[keithsnell]

Hi Jaime and Shane,

It's worth noting that in the link that Jaime (Girod) provided, Bill Janes concludes "So one can spot meter the area in the scene where one wants to retain highlight detail and increase exposure by 3 stops, placing the metered area at clipping."  This is in reference to the D3, and is consistent with my own testing (and Iliah's).  I'm positive that Iliah's initial testing of the D3 pegged the raw clipping point (for the green channel) at +3 EV.

Still, if +3.7 EV works for you, then that's what matters.

Keith

[Shane]

Hi Keith

I haven't got to the point in this test sequence of determining how far I can push the exposure and still retain detail in VII, so in that respect I don't know if +3.7EV will be the final result. At this stage I am trying to determine why the test indicates the RAW clip point to be +1EV higher than your D3 RAW clip point.

Thanks for looking for the old file but perhaps if you have time, it would be quicker to just reshoot a Grey card at 0EV and check in Rawnalyze to see where the G channel falls.