Doug Kerr
Well-known member
The "Zone System" is a famous discipline, developed by famed photographer Ansel Adams and others, intended to guide exposure strategy. To vastly oversimplify it, this discipline (in its "classical" usage) leads to an exposure result where, to use the familiar metaphor, the image of a light gray cat on a snowdrift comes out light gray on white on the delivered print. By contrast, a metered exposure using a frame-average metering system would typically come out as a dark gray cat on a medium gray background.
In the March, 2010 issue of Popular Photography, Michael Frye's article "The Digital Zone System" discusses how we might apply this scheme to modern digital photography.
I thought it would be interesting to examine how the discipline suggested by Frye fits in with our understanding of exposure metering in modern digital cameras.
First, let me postulate a digital camera with the following properties of its automatic exposure control system:
• The "metering" portion is calibrated as prescribed by ISO 2721.
• Then reckoning of "exposure index" for the various "ISO" settings of the camera is in accordance with the ISO SOS ("standard output sensitivity") as prescribed by ISO 12232:2006. (Note that this is "1/2 stop less" than the ISO speed rating, as defined by that same standard.)
The joint result of this is that such a camera, seeing a uniform luminance scene, would use an exposure such that the photometric exposure on the sensor was about 18% of the saturation photometric exposure. (Note that, for several reasons, this is not appropriately described as "the exposure meter is calibrated to 18% gray"; I won't belabor that here.)
This is very nearly the behavior of typical modern Canon dSLR cameras.
Note that relating to the traditional interpretation of the Zone System, this says that such a uniform luminance scene would be planted, by the automatic exposure control system, in "Zone 5" of the tonal scale as envisioned under the Zone System.
Now, to Frye's suggested practice.
Frye first takes the case of an identified highlight area of the scene - an area he describes as "the brightest significant part of the scene that needs to have detail and texture." But note that he really assumes this to be an area of fairly high reflectance. He doesn't mean the lightest part of a "dark gray on darker gray" building façade.
He suggests that we "spot meter" on that scene feature, and then take the exposure "suggested" by the meter and bump it up by one or two stops. He says that this would "plant" that area (on the average) in Zone 6 or Zone 7 of the gray scale of the Zone System. In particular, he recommends the "two-stop" bump (to "zone 7") for scene regions of very high reflectance, such as snow. Let's follow that specific case.
In terms of our "18%" camera, that means that the average photometric exposure for this region of the scene would be 72% of the saturation photometric exposure. That is (on the average) about 1/2 stop short of saturation.
The reason I speak of the "average" for the highlight region is that the luminance of this "highlight" region is not uniform. If it were, the region would contain no detail which we had to be concerned with preserving.
Thus it may well be that that the "1/2-top cushion" in this situation is appropriate to "contain" the range of luminance across our "highlight" region, plus a little margin of safety.
Frye further suggests that, for a situation of "sunrise or sunset light" on mountains, a "one-stop bump" should be used (essentially planting the mountains in Zone 6). For our assumed camera, this would place the average photometric exposure for the mountains at 36% of saturation.
Under this outlook, a scene area that was "planted" at Zone 8 (via a 3-stop "bump") would be "fully saturated on average", and detail across it would be at substantially (if not entirely) suppressed.
This is consistent with a traditional interpretation of the Zone System, in which zones 3-7 are considered desirable habitats for scene areas of any importance (essentially postulating only a 4-stop usable dynamic range).
In the March, 2010 issue of Popular Photography, Michael Frye's article "The Digital Zone System" discusses how we might apply this scheme to modern digital photography.
I thought it would be interesting to examine how the discipline suggested by Frye fits in with our understanding of exposure metering in modern digital cameras.
First, let me postulate a digital camera with the following properties of its automatic exposure control system:
• The "metering" portion is calibrated as prescribed by ISO 2721.
• Then reckoning of "exposure index" for the various "ISO" settings of the camera is in accordance with the ISO SOS ("standard output sensitivity") as prescribed by ISO 12232:2006. (Note that this is "1/2 stop less" than the ISO speed rating, as defined by that same standard.)
The joint result of this is that such a camera, seeing a uniform luminance scene, would use an exposure such that the photometric exposure on the sensor was about 18% of the saturation photometric exposure. (Note that, for several reasons, this is not appropriately described as "the exposure meter is calibrated to 18% gray"; I won't belabor that here.)
This is very nearly the behavior of typical modern Canon dSLR cameras.
Note that relating to the traditional interpretation of the Zone System, this says that such a uniform luminance scene would be planted, by the automatic exposure control system, in "Zone 5" of the tonal scale as envisioned under the Zone System.
Now, to Frye's suggested practice.
Frye first takes the case of an identified highlight area of the scene - an area he describes as "the brightest significant part of the scene that needs to have detail and texture." But note that he really assumes this to be an area of fairly high reflectance. He doesn't mean the lightest part of a "dark gray on darker gray" building façade.
He suggests that we "spot meter" on that scene feature, and then take the exposure "suggested" by the meter and bump it up by one or two stops. He says that this would "plant" that area (on the average) in Zone 6 or Zone 7 of the gray scale of the Zone System. In particular, he recommends the "two-stop" bump (to "zone 7") for scene regions of very high reflectance, such as snow. Let's follow that specific case.
In terms of our "18%" camera, that means that the average photometric exposure for this region of the scene would be 72% of the saturation photometric exposure. That is (on the average) about 1/2 stop short of saturation.
The reason I speak of the "average" for the highlight region is that the luminance of this "highlight" region is not uniform. If it were, the region would contain no detail which we had to be concerned with preserving.
Thus it may well be that that the "1/2-top cushion" in this situation is appropriate to "contain" the range of luminance across our "highlight" region, plus a little margin of safety.
Frye further suggests that, for a situation of "sunrise or sunset light" on mountains, a "one-stop bump" should be used (essentially planting the mountains in Zone 6). For our assumed camera, this would place the average photometric exposure for the mountains at 36% of saturation.
Under this outlook, a scene area that was "planted" at Zone 8 (via a 3-stop "bump") would be "fully saturated on average", and detail across it would be at substantially (if not entirely) suppressed.
This is consistent with a traditional interpretation of the Zone System, in which zones 3-7 are considered desirable habitats for scene areas of any importance (essentially postulating only a 4-stop usable dynamic range).
Last edited by a moderator:
