Ever since I first heard about Robert Green’s project to create a digital version of the Doves type, I’d wondered if anyone (besides Green, presumably) had used it to recreate an original Doves letterpress page, to see how the two compare. If anyone has, I don’t know about it. So I finally decided to give it a try, partly as a limbering-up exercise before starting to print the next HM project (I haven’t done any extended printing since the end of last year).
The first question was, how to conduct the experiment. Ideally my facsimile setting could be directly compared to the original. Since part of the experiment is for people to see how the digital compares to the metal original, it seemed appropriate to pick a work that was reasonably well known. I thought about Credo (1908), but all those I believes get tiresome. So I settled on The Ideal Book. It’s just 10 pages, so three sheets.
But the idea of being able to compare my facsimile directly to the original lingered. Some years ago I acquired a few dozen copies of Cobden-Sanderson’s “Note on a Passage in Anthony and Cleopatra” (1913) This is one of the ephemeral parerga C-S issued over the years. The copies have been lost & rediscovered on my shelves a few times, and I finally had an idea for putting them to use: I could set the first page in facsimile, and present it opposite a copy of the original. I have 25 copies of the “Note...” so it will be included in that many copies of my Ideal Book facsimile.
I grabbed the text for Ideal Book from a digitized copy online, then went through and found all the scanning errors. The digital text conveniently had breaks at the end of each printed line, so that work was already done for me. I pulled out my copy of the original, and measured the page, the margins, the line measure, the type size, and the leading.
In using these measurements to create a digital version, I had to take into consideration the shrinkage of the paper after printing. Like all good handpress printers, Cobden-Sanderson’s paper was dampened for printing, which causes it to expand. When dry it contracts, thereby shrinking whatever was printed ever so slightly. In general I have found paper shrinks 1 – 1.5%. The text line in the book measured exactly 4 inches, so I made the measure in my file 4-1/16 inches to allow for the shrinkage.
The photos above illustrate the shrinkage. For each the ruler’s 4-inch mark is centered under the period at the end of the line (despite what the first image suggests). The second image shows the line on a sheet of proofing paper (i.e. printed dry), measuring just past the 8-inch mark (i.e. a little over 4 inches long). The third image is a sheet of the Saunders paper, printed damp and then dried; it shows the same line ending just under the 8 inch mark, a difference of about 1/16 inch.The fourth is the Barcham Green paper, likewise just under 4 inches. The last picture is the Doves original.
I use Affinity Publisher for doing layouts, & it’s a poor substitute for InDesign. I created a document with page size 6.5 x 9.25 inches. By my estimate, the printed (metal) type equated to 16.1 pt in digital; the leading was 15.4 pt. Kerning was off (I never use auto or font kerning), just as it would be when setting metal.
I dumped the digital text into the text boxes, and was surprised at how well it all fit right off the bat. When it didn’t it was because the word spacing resisted (there are some very tightly-set lines), so I had to make those adjustments as required.
While all of the lines fell into the measure easily enough, what I noticed was very slight variations in where a given letter in the line fell with relation to the ones above and below. Sometimes a hair or two ahead, sometimes behind. Much of this I put down to discrepancies in the word spacing, and I wasn’t going to attempt a facsimile setting to that degree of detail.
The italic words in “Note...” posed a problem: As everyone knows, there is no italic companion to the Doves roman. When an italic was required, C-S used one provided by the Miller & Richard foundry. The closest approximation of it I could find (digitally) was Garamond; without the actual face, my setting can’t really be called a facsimile.
I also had to come up with a pilcrow. Green’s font doesn’t include one, so I simply scanned one from “Note...” and dropped it in.
When output to my cheap Epson inkjet printer, the text looked a little heavy, but I put that down to the printer.
As with all my polymer printing, the plates are KF 95 plastic backed. The slim plates (vs the thicker 152 used by most letterpress shops) allow for sharper details (e.g. serifs) and finer lines without risk of breaking. I have my plates made by a nearby commercial letterpress shop that I’m not naming because they do it as a favor to me and don’t want to be in the plate-making business.
I pulled plate proofs on newsprint with no makeready or particular care, I’m just looking to see if there are any flaws. The type in the proofs looked noticeably heavier than the metal original. I hoped this was due to my offhand proofing, but knew it could also be a result of decisions Green made.
To make the comparison as accurate as possible, the digital version needed to be printed the same way the original was – inked by hand, and printed on a handpress – on paper as close as possible to the original used. The image above shows (l to r) the roller, ink slab, and pile of damp paper under a towel waiting to be printed.
The Doves books were printed on a laid sheet made by Batchelor & Son. I couldn’t get any of that. I did have paper that the (old) wrapping in which it was found claimed to be Saunders c.1950s, but there is no watermark. It’s more cream than the Doves, which might be described as off-white, but it has a similar weight and finish. It was dampened in the usual manner, with blotters, five printing sheets per blotter. I also had a few sheets of Barcham Green Bodleain, also more cream than white, laid, and with a little more tooth than the Doves or Saunders sheets. I printed five copies on this paper.
I started printing with the facsimile page from “Note...” Because of the potential thickening issue, I’d decided to print with as little ink as possible, which thus might require more impression than normal. I rolled out a minimal amount of ink – less than I have ever used to print – and pulled an impression. It was essentially perfect. This never happens for me; there are always agonies and torments getting a form to print well. The impression was good – about the same as on the Doves original – and the ink coverage was solid. And the type still looked thicker than the original.
I checked with the person who made the plates, in case something in how the plate was made (i.e. exposed) would noticeably affect a type’s weight. She compared what I’d printed with a digital print from their good printer, and the two were the same – the type was printing as it was designed to look. So Green had thickened up the strokes, which is not an unusual decision when adapting a metal face for use in offset, digital or Web use.
The image above shows a form being printed. The four red squares at the corners highlight the platen bearers, the two rectanular squares show the roller bearers with tape added in places to adjust the inking in those areas.
Printing a sheet at HM typically take two days, one for each side. It takes two to three hours in the morning to get a form set up and the makeready settled (sample below), then I average about 15 impressions an hour. I have printed up to 120 impressions in a day, but am not good for much the next day. After backing up, the sheets are (gently) pressed between blotting boards overnight, to dry. They usually require drying in a second set of boards for another day. (Note that it’s not the ink that’s drying – it’s dry almost as soon as it’s printed – it’s a question of extracting the moisture that was added to the fibers for printing.)
The 25 copies with the “Note...” facsimile and leaf (above) will be cased in printed blue paper over thin boards. I ended up with another 13 copies, which have been sewn into a painted handmade paper wrap (see top for both versions).
So that’s the experiment. My conclusion is that, if you wanted to use the digital Doves font to make a truly accurate facsimile, you’d have to make some alterations to the font to decrease the stroke weight. And I can’t really think of a legitimate use for the font outside a facsimile. Cobden-Sanderson went to such lengths to prevent its use by others (unscrupulously, given the agreement he had made, in my opinion), that appropriating it for some other use would be jarring. I considered using it for the Pollard project, for about a minute, before realizing it was just a bad idea.
Attempting to replicate someone else’s setting of type ended up being an interesting exercise. It resulted in a deeper understanding of C-S’s aesthetic through all the small decisions involved. Several of his lines are set so tight the words have to be picked apart, while others employ an ampersand when there was plenty of room for the word and. I imagine it’s similar to young writers typing out their favorite novel, to gain some insight to the author’s process.
I feel limbered up and am looking forward to the new project, which will entail printing more colors than any five previous HM books combined!
AND ANOTHER THING!
I’m sorry for all of you not in Vancouver tonight because you’ll miss the loscil/RafaelAnton Irisarri double bill. An ambient mega-bash.