[John Zeissig]

Marvin, I'm not just cooking up some data to support a position. The logarithnic relationship between intensity and perceived brightness goes back to the middle of the 19th century, as summarized by Fechner's Law, Delta I/I = klogI. Ernst Mach first described the perceptual effects of lateral inhibition at the end of the19th century. Hartline and colleagues exhaustively investigated lateral inhibition for several decades, starting in the 1930's. These phenomena are ubiquitous in both vertebrate and invertebrate visual systems. The intensity-response function data goes back at least to the early 70's, principally by Werblin and his colleagues. This is not to mention the thousands of experiments relating these physiological mechanisms to visual perception. There's a vast literature on the subject.

But I think your example, particularly the reference to using a photometer, misses the point of the graph above and emphasizes your seeming disagreement with Tim, which I see as mainly a disagreement based on ambiguity. I'd like to try to resolve it.

Steven and I were discussing events at the upper end of the intensity-response function, but since shadows are important, let's look at the lower, or decrement end. Imagine that you're standing in the noonday sun and looking at the facade of an adobe cantina. The open entrance appears to be a nearly uniform black rectangle. Within the entrance you cannot discern any differentiation of form. If, however, you take measurements with a spot meter, you discover that there are significant variations in light intensity within the entrance. Clearly the meter is responding to something to which your eye is not.

Let's relate the example to the graph below. Here the intensity difference between the brilliantly illuminated adobe walls and the interior illumination (shadow) of the cantina coming through the doorway is so great that the receptors under the image of the doorway are saturated on the low end of their response range. They are receiving relatively little illumination from the interior to excite them, but are powerfully inhibited by their neighbors that are strongly excited by the intense illumination from the adobe walls. The combination of little excitation and massive inhibition sums algebraically to a near total shutdown of response in the receptors within the image of the doorway: the heart of darkness, the blackest of blacks.

First, look at the range of intensities reflected from the adobe walls, indicated by the red lines in the graph below. The slope of the function is steep in this region. A small change in intensity results in a relatively large change in response, so we would expect good discrimination of half-tones.

Now look at the range of intensities reflected from the interior of the cantina, indicated by the black lines. Here an equal change of intensity results in almost no response. Everything in this part of the image looks black and we can detect no half-tones.

Now let's take our best black paint and paint a trompe l'oeil doorway on a portion of the cantina wall. The range of intensities in our painted doorway is shown by the blue lines. It's not as dark as the "real" doorway photometrically, by any means, but on the response axis it hardly differs from it (exaggerated here). It's in the region of shallow slope and appears visually to be a uniform black. It is in this sense that I intended the title of this thread. To our visual systems the trompe l'oile doorway represents the "real" doorway fairly accurately.

Taken in this sense, I find Tim's statement, "If you can see it, you can paint it.", to be unexceptionable. Even if we had a superior paint that could somehow reproduce what the photometer "sees", it wouldn't help very much. The physiology of vision precludes good discrimination at the extremes of the response function. In this context, his post of a painting that convincingly captures the brilliance of the lightbulb, as well as the darkest shadows and everything in between, seems entirely to the point (whatever else one might think about the painting or the man who painted it).

On the other hand, Marvin points out that some natural conditions result in intensity differences that are greater than those reflected by paints under uniform illumination. Well. I wouldn't argue with that, since that describes my cantina example. Painting can't capture what the photometer or other instruments can record under all circumstances. So, in that sense, thinking you can accurately copy nature with paint is illogical.

What I hope is coming across is that our vision doesn't give us a complete copy of nature, nor does painting, nor does a photometer. All we need to do to demonstrate this is to walk into the cantina, and everything changes dramatically. The intensity-response function slides to the left on our first graph, to center on the new illumination level of the interior. What was formerly a black hole is now perceived as a reasonably well-lit room where people are playing cards and reading newspapers. We might see a painting on the wall depicting the front of the cantina in the noonday sun, and be struck by what a good job the artist did in "realistically" painting that scene. He did it by painting what he saw. He didn't have to do it that way, but that's what he did. "Real" to our vision? Yes, pretty close. But Reality (without the scare quotes)? Never!

Steven, yes, the description of halation is exactly what I would expect, and, as a matter of observation, what I see. I'm just not sure how to paint it, so I've got to get those books.