The Haversian Canal

« You heard it here first | Main | Narcissism »

April 6, 2006

Shameless Republication

Republication. Hah! I kill me. . . Too many ways to have read that. . . Anyway:

I often comment on Edward Tufte's forum. My comments, like those of many others, are often deleted. I think he even deletes some of his own comments. Some comments I want to keep around though. Here's an edited version of some of my recent comments on his thread about formalizing photographic aesthetics.

Here's the first article google searching for dynamic range ccd velvia: Photoshop for Astrophotographers:

Different recording media can accommodate a variety of dynamic ranges. Early model consumer digital cameras could capture detail in a brightness range of only perhaps 3 to 4 F/stops and the latest models can record about 6 F/stops. High contrast transparency films, such as Kodachrome and Velvia, can capture about 5 - 6 F/stops. Color negative film can capture 8 or more F/stops of usable detail if correctly exposed.

Note that if we measure the actual optical density of transparency film, there will be more optical dynamic range from the absolute blackest black (the D-Max)to the absolute whitest white (totally transparent) than recording dynamic range. This optical density range is greater than the range that the film can actually record detail in, so the effective useful recording dynamic range is less than the optical density would indicate. For instance, Kodak's Kodachrome has a D-Max of about 3.7, which is almost twelve stops, but its usable dynamic range is only about six stops.

Astronomical CCD cameras have a dynamic range of about 10 to 11 F/stops.

Thom Hogan reports about 7.5 stops dynamic range with the D100 or S2. If Fuji's claims are correct, then they're realizing about a 3 stop (400%) improvement. However, it's not at all clear to me that those 3 stops can be realized in prints, even from an Epson printer. I'm even more pessimistic about CMYK.

There are also certain visual deceits the visual system plays on the rest of the brain in real time. The eye very quickly changes aperture, by dilating the pupil, based on what part of a scene falls on the fovea, the very central high resolution optic disk, which is only a couple millimeters across. The subconscious also considers the ambient light of the entire scene. There are other control systems for enhancing visual acuity, inhibitory surround being the classic example. It increases contrast through neuromodulation.

Some additional articles by Roger N Clark and others have some interesting tidbits. I found them with the search dynamic range human eye stops.

One last one, and I'm done. From the Wikipedia:

At any given instant, the retina can resolve a contrast ratio of around 100:1 (about 6 1/2 stops). As soon as your eye moves (saccades) it re-adjusts its exposure both chemically and by adjusting the iris. Initial dark adaptation takes place in approximately four seconds of profound, uninterrupted darkness; full adaptation through adjustments in retinal chemistry (the Purkinje effect) are mostly complete in thirty minutes. Hence, over time, a contrast ratio of about 1,000,000:1 (about 20 stops) can be resolved. The process is nonlinear and multifaceted, so an interruption by light nearly starts the adaptation process over again. Full adaptation is dependent on good blood flow; thus dark adaptation may be hampered by poor circulation, and vasoconstrictors like alcohol or tobacco.

There is also the phenomenon of bleaching: to much light can overwhelm retinal pigments. Thus the spot after staring at a bright light or a white paper looking faintly red after staring and a green sheet of paper.

20 stops of resolution seems enormous to me. Racing mountain bikes in forested areas, one experiences some tremendous variations in light, ranging from the darkness under the edges of roots to full sunlight in the same "frame", all at 3 to 30 miles an hour and looking with a central axis of the visual field regularly, forcibly, ranging between 45 degrees above the horizon to 45 degrees below the horizon. Imagine a wooden rollercoaster inside a forest but with a terrible need to understand what one is seeing because there are no rails. The decision
points are coming very fast and have serious consequences. Trees hurt, but not as much as the rocks at the bottom of a ravine. The sensations of blow-out and blur are quite familiar to this crowd. I suspect the five to eight stops of dynamic range that engineers seem to have been targetting in most production films and DSLRs may hint at the working range of the eye most of the time.

Posted by Niels Olson at April 6, 2006 5:22 PM

Comments

Post a comment