> maybe

Even if I don't necessarily believe that Dwiggins did that on purpose, I can still ask: How far would you be willing to take this? Would you incorporate such "misalignment" into your own work? On what grounds, by what degree(s), and with what techniques?

hhp

"What I'm talking about is very much fuzziness (albeit at the subvisible level), and you can't do that with [current] software."

What? You mean you're not aware of the "Subvisible Fuzzy Linotype Impression" filter in Photoshop?!!

If I were noodling with Electra to make it work better, i.e. making it look the way I would want it to look at particular sizes under particular print conditions, then I may indeed fiddle with the alignment, but I wouldn't do this simply because Dwiggins appears to have done it. If I were going as far as editing a font to make it more useful to me, I would make whatever alterations I thought beneficial. My point is that this approach is quite different from conducting a 'revival', which obliges one to refer to the original and to conduct oneself with some respect. What I'm talking about is a kind of pragmatic sacrilege.

Hi all

Sorry I haven't replied to the latest. My connection went down early this afternoon and I just got back from three hours with Hrant. If you can imagine what that's like.

Some answers to belated quiries. Yes overinking is the norm. I'd say you could almost define fine printing as a noble attempt to prevent this.

Fuzziness in letterpress is usually the result of paper dust accumulating on ink residue on the beard of the face. Constant wiping of the form will eliminate this but in long edition runs it will be more common. Tends to be more of a problem with photopolymer because of the shallow relief surface.

I notice in the Linotype specimen book, yes, printed on calendered stock, that almost all the very small text sizes exhibit baseline shifting of certain characters. This is also a characteristic of Monotype. I'd say this may be deliberate, for legibility, but I can't recall that I've ever noticed it in foundry castings. It always looked wrong to me. When working with Monotype, in which the matrix position is adjusted by the operator, I always suspected incorrect alignment. I'd say the same with Linotype but the specimen was produced by the firm so I doubt that it was a mechanical aberration.

Tiffany, its an old saying. From an instructor. Never put words in the mouths of those from different times and different places was the explanation. I think that is good advice.

Hrant

I misunderstood your fuzziness question. Impression is a technique to facilitate correct inking. Given all the various factors that come into play during presswork there is only a variance of about a thousands of an inch that will provide you with a consistent pattern of inking based on impression. Too little impression and you will be consistently overinking to compensate. Too much impression and you are encouraging inkspread no matter what the pattern of inking. The feathering you are seeing is ink growing (forced) into the paper fibers. Traditional oil-based letterpress inks dry by absorption, so there is a tendency for them to want to travel.

So, to answer Tiffancy, IMHO overinking and underinking is not a matter of taste, impression is a technical consideration. If you want to have clarity of letterform, inking and impression have to be coordinated.

On difficult stock, many fine printers would deliberately print "gray" with a "kiss" impression just to preserve this. Since a lot of graphic designers today want the crunch and splat look, a lot of contemporary letterpress folk will accomodate them. In that regard, taste is everything. You wouldn't want to see dear Electra that way though.

>Even if I don't necessarily believe that Dwiggins did that on purpose,
>I can still ask: How far would you be willing to take this? Would you
>incorporate such "misalignment" into your own work? On what grounds,
>by what degree(s), and with what techniques?

I have contemplated this very thing. An early version of Whitman had a fairly large overshoot on some characters and looking back at early proofs, I think there was a certain vitality that this may have lent to the line in text sizes and I am faintly sorry that I was so quick to regularize it out. At the least, I think it bears some more exploration and I will likely revisit the ideas in my next design.

When I am able, I hope to ascertain whether or not Dwiggins deliberately designed these apparent inconsistencies in alignment into his types and to what extent. I have no concrete proof now, but it wouldn't surprise me if he did. I say this based on other elements that he and Griff consciously allowed to be inconsistent. In a letter from February 1931, regarding early proofs of Electra, Dwig wrote to Griff: "There is a variety in the weights of stems and in serif details. I think this is not bad -- better than if they all measured exactly the same. Might it not be good to keep these varieties -- differences in serifs, in weight and shape of stems (as between 'i r j' for example; and the long stems of 'b d k h l') rather than to adopt one particular stem or serif as standard and bring them all to that one?" Following this, in his typescript dossier on the project, Griffith appends this annotation: "[This is the essence of good design, little understood or slothfully disregarded by the best of designers. A notable example of the effect of such variety of detail on the movement of letters is to be seen in the cutting of Linotype Monticello.]"

From this I am lead to believe that in examining Dwiggins's types, idiosyncrasies and irregularities are not to be automatically discounted as imperfections of the mechanical processes. That is not to say that every slant and wiggle is intentional and should be slavishly followed, just that they must be considered carefully.

For one thing, a digital font, like it or not, will be used at any number of sizes. As a type gets larger these inconsistencies become more visible and no longer contribute the desired effect. In a note on the uncompleted project for a 24-point Electra, Dwiggins observed: "The details of serif, etc., begin to be seen in 24 pt., and need to be more uniform than in 12." And also "The slight concavity of the stems needs to be uniform throughout for the above reason." So even if inconsistencies were designed into the 12 point master, it might in fact not be feasible to incorporate these into a contemporary font, given today's practices and aesthetics. I have no conclusions yet, just an interest in delving deeper into the matter.

-- Kent.

Gerald, as there is a discussion of inking, perhaps this won't be too off topic. The classic printing red is crimson, I believe, but now I see darker reds generally used instead. Is this related to the difference in inks on letterpress and offset, and the interaction with the color of the paper showing through the non-black inks? Or is this just a change in taste?

> in examining Dwiggins's types, idiosyncrasies and irregularities are not to be automatically discounted as imperfections of the mechanical processes.

Well, I've come to agree; and I think Dwiggins was one designer bright and innovative enough to consciously use irregularity to advantage. In fact, one could even explain why the smaller sizes have more variance than the larger: readability versus aesthetics. I hope you find "proof" of this one day - that would be such a nice standard to brandish against The Grid (which is as present in Noordzij as it is in Crouwel, BTW).

hhp

BTW, if there's enough interest, I can put up scans from a couple of Linotype specimen books that Gerald has loaned me.

hhp

Kent,

Dwiggins did write something about deliberate idiosyncrasies. I read this a long time ago, but remember it distinctly, because at the time I thought it was a surprising concept. He said something about too much regularity inhibiting readability. I've been looking for it for the last couple of days, because this keeps coming up in this thread, but have not found it yet.

Digital Electra: what Dwiggins was desiging towards was not the patern drawings or the metal type itself, but the inked image of the type on paper. His pattern drawings were not the finished product; they were what would give him the letter forms he wanted when printed. This is something completely different than desiging a face that looks like it was printed via letterpress.

Electra italic vs. cursive: the italic is interesting to see in mass, because its almost exactly like the roman in color. The cursive, because of all the rounder shapes, has a different effect. An single italic word in a sentence set in roman is nearly invisible; there is not enough contrast for it to stand out clearly, while the cursive contrasts well.

Hrant: if you want to play with blurred letters, just put a fine gray stroke around them. Pretty much any page layout or drawing program will let you do this. This gives the effect of a slight blur. It's a trick I use a lot when doing illustrations; an outline in about a 25

> An single italic word in a sentence set in roman is nearly invisible

I think some rare words would be hard to catch sometimes, but overall it's totally visible.

> just put a fine gray stroke

But an imagesetter doesn't know gray: a dot is either on or off. That's why you need a defocus - it's analog.

BTW: the scale I'm talking about is way smaller than even your top example.

> press hard on your eyeballs

:->

hhp

"The classic printing red is crimson, I believe, but now I see darker reds generally used instead. Is this related to the difference in inks on letterpress and offset, and the interaction with the color of the paper showing through the non-black inks? Or is this just a change in taste?"

William, in the early years of incunabula printers would have rubricted (printed in red) simply because it was the traditonal thing to do, as manuscript books were often rubricated. By Moxon's time (1683), they were using a number of standard colors. The reds he mentions are Vermillion and Red-Lead. Usually paper would effect the color. With current standardization of colors (PMS) there would be no signifigance difference between any method. There are transparent whites and opaque whites, however, that when used with colored pigment will provide variance of opacity. Don't really know how to answer your question though.

The digital Electra is based on the 14 point master.

hhp

Kent, for a more extreme example of irregularity succeeding you can compare the original Caslon with something like Adobe Caslon. I gather that 'Founders Caslon' is an attempt to capture the flavor of the original, but without the hoakiness of bumpy outlines to mimic early printing. Caslon Old Face has lots of irregularities in the matrices I believe. Adobe Caslon is a fine face, no doubt, but the more irregular versions to me are more inviting and more readable at small sizes.

I feel that Caslon is a little too irregular, but you can go too far the other direction and get something that is beautiful and uniform, but less inviting and less readable at small sizes, like Bodoni.

I am wondering whether this is happening now with digital type...

But there's a huge difference between intentional "irregularity" versus just... crappy quality.

hhp

William

"I gather that 'Founders Caslon' is an attempt to capture the flavor of the original, but without the hoakiness of bumpy outlines to mimic early printing."

I recently had an article published in _Parenthesis_ where I provide a number of enlargements of the Founders Caslon. Justin reproduced the face based on existing type and specimens. The best he could find, yet it was done purposely _in situ_; ink spread and all. As I remember he initially was not intending to even provide kerning pairs. The 8 point is particularly revealing.

Why in the hell isn't there more currently in print about Dwiggins? I just spent $60 for "MSS by WAD" and you rarely see the reprint edition of the Agner "Books by WAD" for less than $75 (never mind the first edition). Why doesn't somebody like Dover just compile all of this info and more into a single affordable volume? You can get Bruce Rogers' "Paragraphs on Printing" from Dover for less than ten bucks...

Andrew,

I've been interested in Dwiggins for a few years now, and it's frustrating that you need to buy books that are 50 and more years old to read his words and see his work.

There is a really nice overview of his work up to the time (1934) in the second volume of the Dolphin, published by the Limited Editions Club. It's not much, but there is not any other overview of his body of work that I know of.

You guys must be worried that the Linotype specimen scans contradict your stance... :-/

hhp

Hrant -- What scans? My stance remains the same: the digital Electra does not yield results like the original -- for whatever reason. It wasn't too spindly in the designer's day, but it is now; I don't like it and in my opinion it is no longer a text face suitable for trade book printing. I do not think that the difference is due solely to overinked letterpress printing versus offset. More carefully printed specimens of metal Electra are not likely to convince me that I should like the digital just the way it is, stop complaining, and start using it in my books. But please, post away with the scans and present your argument. I'm all ears.

Meanwhile, I'm still tracking the history of the digitizations.

Andrew and Paul -- Paul Shaw is at work on a biography of Dwiggins. I have had a preliminary conversation with a publisher about a new collection of some of Dwiggins's writings, but I can't say whether that will lead to anything. I don't know what Dover's process is, but there could be a lot of different permissions to track down. In the case of MSS, I think all of the individual rights holders for most of the different essays would need to be tracked down -- not all of the pieces were Dwiggins's copyright. For those that were, until recently Dorothy Abbe would have been the person to ask, and I understand that she was very protective of WAD's legacy. The situation with BR's Paragraphs would undoubtedly have been much simpler.

So you may have to just sit tight for the time being.

-- K.

> What scans?

The scans from the original Linotype specimen book(s). So I'll have them up soon.

> the digital Electra does not yield results like the original

Well, I agree with that. And I enjoy your "complaining"!

I do however see a couple of dangers (not necessarily pertaining to you):
1) A certain fetishism of letterpress clouding judgment.
2) The assumption that Dwiggins himself agreed with this or that "rendering" of Electra. One thing you see is that the Linotype prints are much lighter than even the ones in "Emblems and Electra", much less the usual work we see, which tends to over-inking/pressing, as Gerald explained. The other thing is that Dwiggins might have preferred lighter type had he lived in our time, like most people -if not you and me- seem to. What does this all mean in terms of the merits of Dwiggins versus Electra itself?

> in my opinion it is no longer a text face suitable for trade book printing.

Agreed.

> I do not think that the difference is due solely to overinked letterpress printing versus offset.

You're right: it's also due to the digital being based on a large size of the metal (14).

--

And good luck with all your WAD efforts! He deserves a lot of attention.

hhp

My stance remains the same: the digital Electra does not yield results like the original -- for whatever reason. It wasn't too spindly in the designer's day, but it is now; I don't like it and in my opinion it is no longer a text face suitable for trade book printing. I do not think that the difference is due solely to overinked letterpress printing versus offset. More carefully printed specimens of metal Electra are not likely to convince me that I should like the digital just the way it is, stop complaining, and start using it in my books.

Kent

Since the digital face appears to be almost identical to a letterpress printed specimen presented in the way that the manufacturer intended, your stance does seem a bit odd. Most likely Electra may as well have met with consumer resistance when it was issued as a metal face for the very same reasons for which you are now rejecting the digital. It wasn't exactly the most popular face of its time period. Given only the minor discrepencies between the two specimens I furnished Hrant, I'd say, that with a font-editing program, you could easily adjust the digital to that of the original. That would not solve the problem though. It would still be even more spindly. Since there was a bold version issued in metal as well as digital, I'd also suggest that a meatier version (demi-bold) could possibly be put together using Fontographer's or FontLab ability to interpolate between two adjusted masters. Then you'd have the color you prefer, without the outline disturbance of inkspread, as in the earlier specimens shown here.

> the digital face appears to be almost identical to a letterpress printed specimen

But only the 14 point. The 10 point for example sets almost 10% wider. And since the color is maintained across the sizes, the smaller sizes must be getting darker too.

> It would still be even more spindly

Why?

hhp

the digital face appears to be almost identical to a letterpress printed specimen

But only the 14 point. The 10 point for example sets almost 10% wider. And since the color is maintained across the sizes, the smaller sizes must be getting darker too.

> It would still be even more spindly

Why?


Yes, only the 14 point. But, you do realize that if you were to purchase Linotype matrices you would have to buy each size separately. 7, 8, 9, 10, 11, 12, 14 point. And the italic would be a separate purchase per size, as would the bold. A sizeable investment. One of the reasons the idea of offering sizes during the photofilm years was abandoned was because the consumer resisted buying extra sizes when they could just as easily photographically enlarge the one size they did buy. The same resistance would be met with today with digital type. The onus is less here on the manufacturer than it is on the consumer. Monotype, for one, did try to offer various sizes or a font during the photofilm years and it was an unsucessful effort. There was no good market reason to continue this practice with digital type.

It would be even more spindly because my reading of the metal and digital specimens indicated that certain strokes and crossbars on the metal were actually thinner than the digital. If you were to replicate the digital exactly from the metal specimen, you would have an even more spindly face.

> The same resistance would be met with today with digital type.

But there's a difference: it's much cheaper to do now. Assuming you have the know-how, you just need to make an extra master (for each style), and inter/extra-polate.

hhp

>But there's a difference: it's much cheaper to do now.

Oh yeah. I recently gave a fellow printer two sets of Linotype Optima matrices in the 10 point roman. He told me that in the early 70s these would have cost around $2000 to $3000. Not sure how accurate that is but I wouldn't doubt it was somewhere in the ball park. I know that replacement character mats for Monotype were priced at $20 or $25 each in the mid 80s. When letterpress was commercially viable it was prohibitively expensive.

Digital type is dirt cheap by comparison.

>1) A certain fetishism of letterpress clouding judgment.

Possibly. I'm trying to see past the nostalgia ;-)

I do not long for the bumpy edges. I do not require squishy outlines or rounded serifs or anything like that. I recognize that there will be a range of stroke weight between any given samples, which is greatly influenced by vagaries of printing technique. I happen to think the stroke weight of the digital is a bit light, but that is not my foremost complaint.

My primary objections to the digital Electra are its excessive contrast and some of the changes in the character widths.

>2) The assumption that Dwiggins himself agreed with this or that "rendering" of Electra.

I agree. Which is why I try to be very careful to qualify my opinions and to provide as many primary source citations as possible -- to show how I arrive at my interpretations. But they are, of course, always just that -- my interpretations, colored no doubt by my own biases, despite my best efforts at objectivity.

>The other thing is that Dwiggins might have preferred lighter type had he lived in our time, like most people -if not you and me- seem to.

This may be true; tastes evolve and change over time. We can only speculate. Indeed, one of his later typefaces, Arcadia, was intended as a very light face -- "round and crisp -- like the new moon one day out -- a trimming of Diana's toe-nail . . ." If I'm calculating WAD's dimensioning correctly, it would be about 60 em units in stem width at 12 point -- slightly lighter than Joanna. But it had less contrast than the digital Electra.

Returning to that. Dwiggins's own comments in the introduction of Electra include this statement: "The face, as may be seen from this specimen, falls within the 'modern' family of type styles, but it is drawn to avoid the extreme contrast between 'thick' and 'thin' elements that marks most 'modern' faces." I'm not sure this applies so well to the current version.

It is an excess in this contrast -- the ratio of thicks and thins -- that I interpret Dwiggins to mean when he refers to "dazzle" in this assessment of an early proof of the 24-point Electra experiment: "The 2/17/37 experimental l.c. seems to me to have a 'dazzle' due to too much variety of color. . . . I have kept close to this stem weight, but have tried to iron out dazzle by getting the range of color more uniform as you can see from the thin-papers marked 'revised'." ["Thin papers" refers to Dwiggins's working drawings.]

A screen-resolution scan may not fully convey this, but the contrast ratio in the letters that WAD refers to is less than that in the digital Electra.


Electra no. 55, 24 point experimental proof


Digital Electra

If Dwiggins felt that the 24-point proof was too dazzling, I imagine that he might have thought the 12-point digital to be even more so. I certainly do. [Full disclosure: my scan is taken from a photocopy of the original proof; I believe it fairly reflects the contrast of the proof, but I can't be certain.]

[You know, I have to admit, though, that the digital looks a whole lot better when you scan it in and blow it up like this. Much nicer than when printed at text sizes on paper.]

>Since the digital face appears to be almost identical to a letterpress printed specimen presented in the way that the manufacturer intended, your stance does seem a bit odd.

Obviously, Gerald, I am not looking at what you are looking at, and you are not looking at what I am. And we are surely looking with different eyes. It is probably time for us to agree to disagree. I remain a little puzzled, however, -- and forgive me if I am misrepresenting you -- but your statement seems to imply that the way that the type was presented to the public by the manufacturer when it was first introduced was not the way the manufacturer intended, or that the way the type was presented in projects designed and printed under the direction of the type designer were somehow less accurate representations of the typeface than the showing in Big Red. I am not trying to denigrate the showing that you are referring to; I'm just not convinced that it necessarily lays a greater claim to the ideal representation than other examples.

BTW, when was the big red book published?

-- K.

"your statement seems to imply that the way that the type was presented to the public by the manufacturer when it was first introduced was not the way the manufacturer intended, or that the way the type was presented in projects designed and printed under the direction of the type designer were somehow less accurate representations of the typeface than the showing in Big Red."

I'm not saying that at all. I haven't addressed the issues you mention (the initial presentation, Dwiggins' direction, etc).

But you are comparing the digital to Dwiggins' 24 point, which was never produced. I believe my point was that the specimens in Big Red were (produced). And the digital is a near match for the 14 point. And the specimens are very clean. I think that's about it.

Dwiggins' intentions are obviously of interest, and you've brought that out very well, but the history of type design is full of intentions fouled up by the economics and production practices of the manufacturer. The digital face evolved not from these intentions, that's clear to both of us (I think). My "goal" here was simply to find out where it did evolve from, for better or worse.

BTW, when was the big red book published?

The supplements which I mentioned previously are dated (and I did include this information) but not Big Red itself. Most of the great "fat" specimen books appeared in the 1920s to early 1930s (ATF, BB&S, Linotype, Monotype, etc) but did not necessarily carry date of publication, copyright or whatever. These were all published presumably before copyright became a significant issue and the Depression was reality.