newsbegin page 86 |
Reproducing The Characters of Spenser’s Faerie Queene
Late in 1973 the Greater UNM (University of New Mexico) Fund gave the Blake Newsletter enough money to get started on a project we had been thinking about for a long time—reproducing Blake’s Spenser painting in color and on a scale large enough to give some idea of the effect of the original. Without the help of the Fund, and the cooperation of the National Trust, which kindly gave us permission to reproduce the painting, we would still be dreaming. As it is, we can dream about other pictures and projects.
We knew before we began that we would never be able to reproduce the Spenser picture perfectly. Not only is that more than modern printing technology can manage, but also more than our money would ever be able to buy. So we went the middle way, knowing all the time that we were compromising at nearly every step, hoping all the time that we wouldn’t lose our nerve, and that good luck would bend the compromises in our favor.
We didn’t, and it did for the most part, and we are thankful, but compromises are compromises no matter how much lucky bending there is, and we believe that the readers of the Newsletter deserve at least an account of what was done to get the reproductions that appear in this issue. This is only for readers who want it, of course, and it will be a lot more than most readers will care to have. But the few who may need to make some scholarly use of the reproductions now or later will want to know the limits of their usefulness. An account of this sort is not the custom in fine-art reproduction, but we think it should be.
The photography was done by G. H. Roberton of A. C. Cooper Ltd., fine-art photographers, London. Roberton photographed Blake’s painting in June 1974 at Petworth House in Sussex. Because of its fragility, the painting had to remain behind glass and in its frame. Roberton took a color transparency on Kodak Ektachrome sheet film measuring 6″ × 8″. We wanted a large transparency because we had decided to publish the reproduction as a four-page foldout from the center of an issue of the Newsletter. We had 10 ½″ × 32″ of space to use, and we wanted to use all of it. That meant getting a large transparency. The usual projection slide of 2″ × 2″ (including mount—real picture size much less) would not enlarge well enough. The painting is so wide in proportion to its depth that the actual picture area would have been very small—no larger than about ⅖″ × 1 ¼″. The size of the picture area in the sheet-film transparency is about 2 ⅖″ × 7 ⅕″. We enlarged that approximately 440% to our printed size of 10 ½″ × 32″. The detail in the photograph is fine enough to stand that percentage of enlargement.
The accuracy of the color is assured by testing each batch of film with color and contrast scales, then correcting with appropriate filters. The camera was a whole-plate (6 ½″ × 8 ½″) Gandolfi, with illumination by two 500-watt lamps.
On the same day Roberton also shot for us four black-and-white views of the painting, one of each quarter from left to right. The negatives, from Ilford FP4 film, are large, each 6″ × 8″. Two sets of prints were made from these. Contact prints—directly from the negatives, using no enlarger—were made in London by Cooper. An industrial photographer in Albuquerque, New Mexico, made a set of 16″ × 20″ prints with an 8″ × 10″ enlarger (one that will accept negatives as large as 8″ × 10″). The contact prints are the basis of the black-and-white reproductions of the Spenser painting quarter by quarter (following the introduction of Grant and Brown’s essay). The 16″ × 20″ prints are the basis of the other black-and-white details reproduced in the essay. They are all reproduced with 150-line halftone screens—screens of medium fineness, coarse enough to be easily and economically handled by the printer, and yet fine enough to reproduce detail acceptably well.
The quality of the photography, both black-and-white and color, is very high. Color and contrast scales on the color transparency show that color and contrast are reproduced accurately. The large black-and-white negatives register fine details in the painting extraordinarily well.
Everyone knows that color photographs, even good ones, are seldom strictly true to the objects they reproduce. But as a medium for reproduction, color photography is much more accurate than color printing. So here for the first time we butt our heads against the limits of the process we are using, and the process wins.
The color picture in this issue is printed in what is called “process” color, meaning dyes and inks of four (in this case four—sometimes, in more expensive versions, more) standardized shades are used to make the color “separations” and then to print the picture. Since four colors combine to make all the colors found in the printed picture, “separations” must be made to divide all the colors in the photographic transparency into combinations of the four process colors, which are shades of black, blue, yellow, begin page 87 | and red. Each of the four colors calls for a separate negative, and each negative is used to make a plate that is in turn used to print one of the four colors. The colors are printed one on top of another (actually one beside another, in little groups of four dots) until they combine to reproduce all the shades of color and contrast in the original.
Obviously there is far too much room for loss—of detail, color, and contrast. And much is lost. When the separation negatives are excellent, the plates made from them excellent, the registration of one color on another excellent, the paper appropriate, and the colors well controlled, the results are only fairly close to the original—and every stage can never be quite so excellent as that.
Harper House, Dallas, Texas, made separations for us in June 1974. The color was checked for accuracy with a partial Geva proof and corrected. A Geva proof is made photographically; it is far superior to no proof at all (unproofed color sometimes called “pleasing” color), but inferior to a press proof, which was too expensive. The color as corrected and shown on the proof was fairly accurate, though a little heavier than the color in the transparency. A 150-line halftone screen was used; 150 lines is fairly standard for color reproduction. More lines to the inch would give more detail and more accurate range of contrast; fewer would give less detail and cruder contrast.
We chose Starline Creative Printing of Albuquerque, New Mexico, to make plates from the separations and to do the printing in July 1974. They have won national awards for fine color printing, and their location allowed me to be on hand for the making of a press proof.
Cropping is very conservative. It was done freehand, dot by dot, so as to lose as little of the picture as possible. Most of the loss is on the left edge, and is due to the trimming of uneven separation negatives. (The full painting—at least from the edge of one mat to the edge of the other—is visible in the black-and-white photographs.)
The printing was done on a four-color press with a large bed, which made it possible to print the picture with the full width across the bed for superior registration. A four-color reproduction can be printed on any press of the necessary size, but the ordinary press can print only one color at a time. Running one piece of paper through a press four times, once for each color, creates problems that it is nice to be able to avoid. The most obvious and serious problem on a single-color press is the unpredictability of the colors in the final version. The color printed first cannot be changed—except by starting over—after the second, third, and fourth colors have been printed, and the same is true for the other colors. Thus no final proof is possible until the fourth color is printed—too late to do anything except wish for better. A four-color press prints all four colors in sequence on one press in a single run, making it possible to adjust the colors and establish a press proof at the beginning.
We printed on 80-pound white enamel paper of a size 25″ × 38″—heavy, big, and glossy for efficient operation of the press and good reflection.
With the plates on the press and the crew ready to make a proof and finish the job, we were all too able to see our compromises shining back at us from heavy enamel paper. There was too little black on the black plate to give us the very best kind of delicate definition without heaviness. Fine details in some parts of the picture and intensest reds in other parts did not want to appear until the red was a bit heavy in yet other parts. And of course changing any one color changes many of the others. A half-hour of adjustment and 500 or so copies of the reproduction brought us what we decided was the best possible combination of compromises. We chose to get some detail, to lose some, to take more red than we wanted but not as much as we would have had to take to get all the detail, and to settle for less black than we would have liked. The color that results is not quite true. The original at Petworth is more muted in tone, and there is more gray and faded yellow there than here. But we had to choose. We hung a proof on the press for reference, and the machine began to chug to hit its stride.