The old axiom for creating high-quality negatives is “expose for the shadows and develop for the highlights.” When it comes to printing negatives in the darkroom, this recommendation appropriately changes to “expose for the highlights and control the shadows with contrast.” That is good advice, but as experienced printers know, there often is a small difference between a good and a mediocre print. So, when it comes to fine-tuning exposure and contrast, how concerned do we really need to be about the optimal settings? How much deviation is acceptable and how little is recognizable? What are the smallest increments we need to work with? How do we advance from casual work to fine-tuned images without going completely overboard? Exploring a sample print of the Castle Acre Priory provides some answers.
Castle Acre Priory is located five miles north of Swaffham in Norfolk, England. Its ruins span seven centuries and include an elaborately decorated 12th- century church, a 15th-century gate house and the prior’s former living quarters, which are still fit to live in.
The picture on this page was taken inside of the prior’s chapel in July 1999. I used my Toyo 45AX with a Nikkor-W 135 mm, ƒ/5.6 on a tripod. This metal field camera travels well, and is fast and easy to set up, considering the large 4×5-inch format. The 135mm lens was required because the room is very small and I was not able to step back any farther. I measured the scene with my Pentax Digital Spotmeter and placed the dark interior wall on Zone III in order to keep the option of some detail. The bright vertical wall of the window fell on Zone VII, but due to the bright sunlight, the window sill was clearly on Zone XI. To pull the sill back onto Zone VIII, N-3 development was needed. I changed the EI to 25, which is necessary when dealing with a rather broad subject-brightness range such as this, in order to sufficiently expose Kodak’s TMax-100. This maintains shadow detail when development time is shortened. At ƒ/32, the calculated exposure time was eight seconds, but I extended it to 12 seconds to compensate for this film’s reciprocity behavior.
When printing the image in the darkroom, it became obvious that the N-3 development had pushed the subject Zone III closer to a print Zone II. Actually, the image looks better this way, but I was glad that I had given enough exposure time to get at least good tonality from the shadows, even though most of the detail was lost. With this treatment, the image printed well on grade-2 paper and only required minor burning down of the upper corners. I consider this print to have a full tonal scale from Zone II to VIII, which makes it a prime candidate to discuss optimized print exposure and contrast.
It makes little sense to print highlights lighter or shadows darker than what the human eye is able to discern under normal lighting conditions. Neither does it make sense to worry about exposure differences that are too small to see. With that in mind, some questions need to be answered.
1. What are “normal” lighting conditions?
This question has been answered through extensive research conducted by Henry Dreyfuss in the 1960s and fully documented in his book The Measure of Man. He established lighting conditions for coarse, medium, and fine manual work. These illumination levels were initially meant for working over several hours. I will assume they are also adequate to view photographic prints.
Figure 1 shows Henry Dreyfuss’s findings, but I included the conversion to EVs at ISO 100/21 so that you can quantify illumination levels with your own light meter. Consequently, EV 7 is the minimum illumination at which a print should be displayed, and there appears to be no benefit to illuminate beyond EV 11. This range seems to be reasonable, based on the display lighting conditions in my own home and those found in galleries. When lighting levels drop below EV 7, previously well-detailed shadows get too dark for good separation. At illuminations above EV 11, previously well-detailed highlights tend to bleach out. This is the logic behind the recommendation to print with the display conditions in mind, as advocated in Ansel Adams’ book The Print. A picture to be hung in the dark hallway of the local church must be printed lighter than the same picture exhibited in a well-lit photographic gallery. I recommend printing for normal lighting conditions of EV 7 to EV 11 if the final display conditions are not known.
Figure 1 also reveals that the shadows are more affected by dim light than the highlights are affected by bright light. It would therefore be safer to examine the image at the lower threshold of display illumination while printing. I study my prints in the darkroom on a plastic board next to my sink. It is illuminated to EV 6, and this ensures good shadow detail in the final print. If I can see details in the shadows at EV 6, then I will be able to see them under normal lighting conditions too. If the evaluation light is too bright, there is a danger that the prints will be too dark under normal lighting conditions. As an additional benefit, printing shadow detail for EV 6 also helps to compensate for the dry-down effect.
2. What are the reflection- density limits for tonality?
The Zone System defines the tonality limits as Zone VIII for the highlights and Zone II for the shadows. There is no universal agreement on precise reflection densities for the equivalent print zones, but the existing standard for paper characteristic curves, ISO 6846, can help to define approximate values. Figure 2 shows how this standard concentrates on the textural density range of the characteristic curve by ignoring the low-contrast areas of both the toe and the shoulder because they have little practical value for pictorial photography.
The standard defines the first usable density as being 0.04 above the base density of the paper. Most fine-art printing papers, including Ilford’s Multigrade IV, have a base white of about 0.05 reflection density. Therefore, we will place Zone VIII at about 0.09 reflection density for most papers. Some warmtone papers, or papers with an ivory base, may have a slightly higher value due to the fact that they have a less reflective base white, but they are in the minority.
The current ISO standard defines the last usable density as being 90% of the maximum density, also called Dmax. Another factor to be considered is the sensitivity limit of the human eye to shadow detail. I conducted a field test in normal lighting conditions of around EV 8. Six people were asked to identify the darkest area with still visible detail on 30 different prints. The mean of 180 density readings was 1.88 with a standard deviation of 0.09 density. Today’s glossy or pearl papers have Dmax densities of about 2.10, or higher if toned. The 90% rule of the ISO standard points to a last usable density of 1.89 on these papers.
The almost precise correlation of the two numbers is a coincidence. However, the agreement of these two methods, as well as good corroboration with studies by other authors, including the book Controls in Black-and-White Photography, by Richard Henry, seems to indicate that this value is a good approximation for the last usable density. Consequently, we will place Zone II at about 1.89 reflection density for most papers. There is a minority of matte surface papers, or papers with textile surfaces, which have significantly lower Dmax values, and for these papers the use of the 90% rule is more appropriate to calculate the last usable density.
3. How discriminating is the eye to reflection-density differences?
The answer to this question will determine how concerned we need to be about print exposure differences. A rule of thumb adopted by some printers has been that a 20% change in exposure is significant, a 10% change is modest, and a 5% change is minute. In conventional f/stop timing terms these values closely correlate to 1⁄3, 1⁄6, and 1⁄12 stop, respectively. I conducted another experiment to find the answer.
Two step tablets were exposed and processed. For each, I used a piece of 5×7-inch paper and printed seven, 1-inch-wide bars onto it. One step tablet was printed around the Zone VIII target density of 0.09 and the other was printed around a density of 1.89 to represent Zone II. The bars differed in exposure by 3%, or 1⁄24 stop. The results were presented to a different group of six people. The individuals were all able to see faint differences between the bars in lighting conditions from EV 7 to EV 11, and it seemed to be equally difficult to differentiate highlights and shadows. The test was repeated by cutting the exposure difference to 1.5%, or a 1⁄48 stop. In this test, four individuals had difficulty detecting any bars. I concluded that 1⁄24 stop was about the limit of detecting exposure differences in Zones II and VIII under normal lighting conditions using adjacent gray bars.
The results of the related density measurements are shown in Figure 3. The densitometer revealed that a 1⁄24-stop exposure difference was responsible for a density difference of only 0.003 at Zone VIII, but 0.016 at Zone II. Therefore, we can conclude that the human eye is about five times more sensitive to density differences in the highlights as opposed to the shadows. However, as far as exposure difference is concerned, the discrimination of the eye is about the same between highlights and shadows. Figure 4 can help to explain this fortunate condition.
The contrast at any point on the characteristic curve can be quantified by creating a tangent to the curve at said point. The tangent of the resulting angle is a proportional measure of contrast. As you can see in Figure 4, the tangent at the Zone II density is about five times greater than the tangent at the Zone VIII density. Therefore, the eye’s lack of sensitivity to the density differences around Zone II is entirely compensated for by the increased contrast capability of the material at Zone II. This explains why we need a similar exposure to get the same discrimination between highlights and shadows.
When I repeated the whole test with approximate density values for Zone III through Zone VII, I found that a 1⁄48- stop exposure difference was still detectable at Zone III and Zone VII and not at all difficult to see at Zone V. The increased local contrast in these areas explains these findings, but our print- ing efforts will concentrate on Zones VIII and II to optimize highlight and shadow detail. Nevertheless, the additional data were valuable to complete Figures 1 and 3, and might also be useful for images that don’t include the entire tonal scale.
It must be added at this point, that the entire test was done with adjacent gray bars. My experience shows that our eyes are more discriminating to this condition than comparing two photo- graphs, even if they are identical images and right next to each other. Our ability to compare two identical images in isolation is even further reduced. Therefore, I find an exposure tolerance of 1⁄24 stop to be rather demanding. I have adopted a tolerance of 1⁄12 stop for my own work, which is more practical and sufficient for most prints. However, 1⁄24 stop can be useful with images printed on harder papers, because they have much higher contrast gradients.
In conclusion to our concern of fine- tuning print exposure, we may take a final look at Figure 1 and answer all three questions simultaneously. Normal lighting conditions for display prints should be from EV 7 to EV 11.
The approximate log reflection densities for Zones VIII and II are 0.09 and 1.89, respectively, on most papers. The minimum exposure difference to alter the tonal values of a print appreciably is about 1⁄12 stop, but can be 1⁄24 stop with harder papers. This is a verification of the rule of thumb mentioned earlier. Shadow detail suffers first and rapidly when illumination drops below EV 7, and it is valuable to examine print progress at EV 6 to secure this detail. Highlight detail is not as sensitive to different illumination levels as shadow detail, but it is important to have precise highlight exposure, because the eyes are most sensitive to density variations in the highlights.
The desired shadow detail is typically fine-tuned with paper contrast after the highlight exposure has been set. The recommendation is to start with a soft paper grade estimate, and then slowly move up in contrast until the desired shadow detail has been reached. The trial and error portion of this approach can be minimized if we realize that contrast can also be referred to as the exposure of the shadows.
We can use Figure 5 to determine the required exposure for Zone II. Only the characteristic curves for the paper contrast limits, grade 0 and 5, are shown. I made sure that both papers were exposed so that the highlights of Zone VIII were rendered with the same reflection density. This allows us to measure the relative log exposure difference between the shadows of these two paper grades. In other words, the highlights were placed on top of each other to see how much the shadow exposures differ from each other. The shadows differ by about 1.0 log exposure.
A different method is shown in Figure 6, and it leads to the same conclusion. All standard paper grades have a defined log exposure range to match different negative density ranges. Soft papers have a low grade number and a wider exposure range than hard papers, which have a high grade number. Although all grades have exposure ranges expressed within the shown limits to accommodate manufacturing tolerances, we need only to concern ourselves with the average exposure ranges for this exercise. The difference between the average exposure ranges of grades 0 and 5 is, reading from the table, 1.55 – 0.58 = 0.97 log exposure.
This is a very similar value to the log exposure difference 1.0, which we got from Figure 5 earlier.
Paper grades are often subdivided in 1⁄2-grade increments to provide enough flexibility to fine-tune image contrast. This provides ten increments between grades 0 and 5, and we can assume, within reasonably small error, that the log exposure difference between grades 0 and 5 is linear. Consequently, a log exposure difference of about 1.0, between grades 0 and 5, divided by ten increments, results into a 0.1 log exposure difference between 1⁄2-grade increments. By definition, a log exposure of 0.3 equals one stop of exposure difference, and therefore, on average, a 0.1 log-exposure difference makes for a 1⁄3-stop exposure difference between 1⁄2- grade increments.
Figure 7 shows the relationship between the f-stop exposure differences of the shadows and paper-grade deviations if the highlight exposure is kept constant. This discovery has placed the value of test strips into a completely different light for me. In the past, I looked at test strips purely as a tool to determine accurate highlight exposure. Moreover, I resisted looking at the shadow detail in test strips, because I knew how confusing it can be to determine exposure and contrast at the same time. Only after the highlight exposure was set did I modify the shadow detail by slowly changing paper contrast. I still believe that there is much value in this approach, and I would not recommend anything else to a beginning or practicing printer. However, I’m glad that Paul Butzi pointed out to me that an advanced darkroom practitioner can get valuable information about the desired paper contrast by evaluating the shadow detail of the test strip.
Figure 8 shows three test strips, which differ only by the exposure increments used. They have all been printed at grade 21⁄2 and have the same base expo- sure of 18 seconds in the center. The exposure decreases to the left and increases to the right. Figures 8A–C, from top to bottom, were prepared in 1⁄3, 1⁄6, and 1⁄12-stop exposure increments, respectively. There is enough information here to fine-tune the highlight exposure securely down to 1⁄12 stop. Let’s assume, for this example, an exposure of 18 seconds to be just right. Now we can take a look at the shadow detail on the different test strips. I can- not predict how easy it will be to see a difference in the final reproduction of these test strips. In the originals, how- ever, there are clear differences in the upper left corner shadows of Figure 8A, a modest difference in Figure 8B, and a minute, but still visible, difference in Figure 8C. From Figure 7, we know that 1⁄3, 1⁄6, and 1⁄12-stop exposure differences are equivalent to 1⁄2, 1⁄4, and 1⁄8 grade, respectively.
Therefore, we can look at highlight and shadow detail on different test strips and select one each to our liking. We will then know immediately what exposure time is required to retain highlight detail, and what contrast change is required to achieve that level of shadow detail. As an example, if we like the highlight detail of the center strip in Figure 8A, but we prefer to have the shadow detail of the second strip to the left, then the base exposure would remain at 18 seconds; however, the contrast would have to be reduced by 1 grade. Figure 8B allows contrast selection down to 1⁄4 grade, because the exposure increments are only 1⁄6 stop. Figure 8C allows us to select contrast increments as low as 1⁄8 of a grade.
It should be added here that, depending on equipment and materials used, minute exposure changes might be required to maintain constant high- light exposure when changing paper contrast. I don’t trust any claims of constant highlight exposure and have tested and calibrated all my tools to compensate for the effect. (A detailed working method is found in PT Jan./Feb. 1997, “Saving Time and Paper,” by Howard Bond.)
4. How accurately do we need to select paper contrast?
Filter manufacturers seem to have answered this question for us. All filter sets on the market come in 1⁄2-grade increments, even though variable-contrast filters and dichroic color heads allow for much finer increments. In the previous section on print exposure, we concluded that an exposure increment of 1⁄12 stop is about as fine as we need to go. Theoretically, this statement is true for the highlight and the shadow detail, as proved in Figures 3 and 4 and verified in Figures 7 and 8C. There seems to be a difference, however, between what a viewer of a photograph is able to discriminate and what he or she is willing to discriminate. The eyes are first and foremost attracted to the lighter areas of an image. The shadow areas will eventually get the viewer’s attention, but very dark or empty shadows are not interesting to most viewers.
Nevertheless, an appreciation for contrast changes down to 1/8 of a grade exists, even though such changes are admittedly hard to see. I consider a 1/4-grade increment to be adequate, but find the standard 1/2-grade increment too rough for fine work.
Figures 9A–E show a sequence of an area from the lower left corner of the lead picture. The exposure was adjusted to have a consistent highlight reflection density, but the paper contrast was increased from a grade 2 in Figure 9A to a grade 3 in Figure 9E in 1⁄4-grade increments. Again, the final reproduction capability of the shadow detail is not known to me, but on the originals, one can clearly see the differences without any need for finer increments. Furthermore, we can see in Figure 8B how the desired paper contrast was easily predictable from a simple exposure test. A test strip provides information about both exposure and contrast.
Fine-tuning print exposure and contrast is essential to obtain optimal print tonality.
Granted, finding the most suit- able highlight exposure within 1⁄12 stop and optimizing shadows within 1⁄4 grade takes some effort, because fine-tuning is most sensibly done through the evaluation of traditional test strips. The one-off approach of electronic metering is not suitable for a comparison of boundary conditions. Nevertheless, for the experienced printer, reading exposure and contrast off the same test strip is a welcome shortcut that leads to uncompromising results.