The Zone System and digital cameras

Warning: this post assumes at least a nodding familiarity with the Zone System, as described in Ansel Adams’ The Negative (do yourself a favor and get at least the 1981 edition), and in many other places.

I recently participated in a discussion about whether Ansel Adam’s Zone System applied to digital cameras. I said that, in AA’s formulation, it did not. As is often the case in such discussions, I came away with some analysis that I wanted to do.

The first — dubious, in my mind — claim to which I turned my attention was that one way to get the correct raw exposure in a digital camera was to meter an 18% gray object in the scene and place it on Zone V. Assuming that the meter is calibrated (probably using UniWB) so that an 18% reading produces a maximum raw value (with most lighting, in the two green channels) of 18% of full scale, here’s how the zones would map:

You can see that Zone VIII is well and truly blown.

I have read that modern exposure meters are adjusted to 12.5% reflectance, not 18%. This would be a departure from the Zone System already and would be calibrated out in the initial setup, but let’s see what happens in that case:

Well, we got Zone VIII back, but everything above that is still blown.

Why are Zones IX and X MIA?  AA was counting on the “shoulder” portion of the H&D curve reducing contrast in the top two Zones (and figuring that the shoulder shape on all films was pretty much the same, which is a little iffy). There is no shoulder in a digital sensor; its response is virtually linear until clipping.

If we want Zone IX back, we’ve got to reduce exposure, and the chart looks like this:

There are two concerns. First, Zone V has dropped to 6.3% reflectance, which is not what we photographers think of as middle gray. The second is that Zone I is only marginally printable — look at the density! It would not have been printable at all with the photographic papers of AA’s time. A digital camera, being linear, is more contrasty, especially in film’s toe and shoulder regions, than the photographic printing process envisioned by Adams.

If we want Zone X, and often we do, the table approaches ridiculousness.

Zone V has sunk to 3% reflectance, and Zone I is unprintable. It is clear that we need to boost the shadows and introduce the shoulder in the development process. It turns out that Lightroom and ACR are very good at doing both those things. First, most (all?) of the Lr camera profiles include both a toe and a shoulder nonlinearity. Second, the positive settings of the Exposure control in Lr produce a shoulder-like nonlinearity that becomes more extreme as you increase the boost. Thus, a way to get Zone X in an image with a camera whose light meter is calibrated to 12.5% reflectance is to underexpose by two stops and push two stops using the Exposure control in Lr. If Zone IX were the highest value we cared about, we’d underexpose by one stop and push one stop in postproduction. Alternativly, we could place a 12.5% gray card on Zone V, expose normally, and get something like what would be in the Zone System N+2 development.

If your camera is calibrated to an 18% reflectance, then the best thing to do is set the Exposure Compensation to +1/3 stop and proceed as above.

If you are only moderately familiar with the Zone System, you may be asking yourself about the utility of recording light brighter than 100% reflectance. The 100% reflectance calibration assumes a matte subject with perfectly Lambertian reflectance. The higher light levels occur when the subject has some specularity or gloss to it. Water is a classic, and photographically-important example. Metals, glass, plastics, skin, stone and many paints are also often far from Lambertian. So we often need the headroom to accommodate those materials.

All of the above assumes that the camera is calibrated using UniWB, and that the UniWB hack works perfectly to predict raw file values. However, UniWB never achieved much penetration among digital camera users, and I think the number of users is waning today as cameras get more and more dynamic range, so it’s worthwhile looking at what happens with widely-used camera setups.

I took a Sony a7III, set the color space to Adobe RGB, the creative style to Standard, the white balance to AWB, the meter to spot, attached a Batis 135/2.8 lens, set the ISO to 100 and the aperture to f/16, and made exposures of a Wescott 1×2 foot LED panel set to full brightness behind a Wescott softbox.  I used correlated color temperatures of 5500 and 3000 degrees Kelvin. The camera light meter gave the 5500-degree target 1/3 stop more exposure than the 3000-degree one.

Here’s what the raw averages of a 600×600 central region of interest look like with various placements of the target:

The above numbers are normalized to full scale, and the black point (512) has been subtracted.

With the warmer, 3000-degree setting, the Zone VIII placement is not clipped, but Zone IX will certainly be. In the 5500-degree case, clipping starts before Zone VIII placement.

Photon noise and photon response nonuniformity mean that we need to consider the tails of the probability density curves associated with each raw channel. Here’s what happens when we consider average (aka mu) plus three times the standard deviation (sigma).

Clipping occurs slightly sooner, since we changed its definition.

Clearly, the sensor in the camera is not behaving at all like the film upon which the Zone System is based, and the Zone System rules are, in situations with bright important highlights, going to result in some really ugly files that will force Lightroom into digital guesswork in the form of highlight recovery.

It gets worse, but that will have to wait for another post.