Hrant, do you have Meridien?
Somebody stop me from licencing the dam thing just to test this.
Okay, again waaay too much time spent. But what's done is done. I post this in the hopes it continues to be instructive.
Naturally, I've selected a particularly egregious example in order to strengthen my point. ;-)
Here is an example of Adobe InDesign's Optical metrics applied to 10 pt Trump Mediaeval. I'm not claiming that TM's spacing is exemplary (though certainly reasonable), but what InDesign does to it is my focus. Also, this particular example seems to exhibit a similar strategy to what I perceive going on in the Meridien sample -- probably because Trump Mediaeval shares some characteristics with Meridien, in particular the long sharp serifs.
The top is the natural fitting. Beneath that is the Optical treatment. Below that is an annotation showing the numerical adjustments that InDesign is applying. Note how it pushes apart the straights and pulls together the straight-rounds, creating an uneven spacing -- E i n bi ldu n g.
At 12 pt, the numbers are different, but the basic strategy is similar:
If you look at Nina's full Meridien page, you'll find this same word seven lines up from the bottom:
This exhibits the same odd fitting quirks.
Compare this with what I must assume is Meridien's natural fitting from the Linotype tester:
We can argue about whether the natural fitting of Meridien is too loose for text or not, or about the effect of various other design attributes on readability. But I hope we can agree that something funky is happening with the fitting in that example, and I'm pretty sure Adobe Optical metrics is the culprit.
Kent, that wasn't just educational, I'd say illuminating! And it looks like you're right on the money about your optical spacing suspicion.
Here's the word from the scan compared with the sample you posted from the Linotype site (or was that MyFonts? In any case, I'd hope they have kerning activated in their Flash).
I tried to scale the two to the same size, though that wasn't quite easy (sorry, my scanner sucks).
This shows a remarkably similar pattern to what you posted, Kent (with the notable exception of n_g).
The crazy thing is, while the word obviously got quite a bit looser overall, there are some pairs that still got tighter (in absolute terms) – those being mainly round/straights (b_i, l_d) – whereas straight/straights like i_n and i_l got looser.
Could this be a case in point for a combination of optical scaling plus tracking?
(I'm getting the creeping impression somebody was consciously trying to wreck the spacing on this one. :-/ )
(PS: James, that was evil. ;-) )
First, a clarification: please do buy Meridien if you actually need it. But not
just to test it out; that money would be better spent buying a font you need.
Second: William, below is a rough-and-ready snapshot of the Meridien passage in the Norton book. I think it's telling that Carter focuses his praise on the demi, that being a weight I'm sure can very well be used to great effect at sizes around 9.
Third: thanks Kent and Nina for the judicious testing. This is my Typophile.
It's all good, Hrant, you were spot on: I don't need it, and I don't think I'll want to use it; in fact I don't like it too much so far (as you may have guessed). That was my "need to figure this out now" urge talking. :-)
Re the Carter passage: Does he actually misspell Meridien, or was it ever sold as "Meridian"?
Btw, what's this set in? Feels nice and tight after looking at all this Meridien stuff.
And, thank you guys for the great input. I'm beginning to feel I've been learning way more on here than in that uh, rather expensive type design class I'm taking… dunno if that's primarily a sad thing to say about the class, or a great thing about Typophile. But I tend towards the latter.
Probably just a Nazz fan. Thanks Kent, Nina, Hrant. I’ve seen so much contradictory opinion on kerning. This is really solid and timely for me to look at.
The heading is spelled "Meridian" as well, and I suspect all that was the editor's fault. The book has a long errata list (but doesn't correct "Meridian"). I should know the font used for the body, but I'm not very good at ID-ing, and quite strangely the book doesn't have a colophon. Frankly it's a wishy-washy, "precious" typeface, with one of those open-bottom binocular "g"s that serves as an excuse for novelty. Nina, I think your desire to see something you like after having to look at Meridien for too long is leveraging the lack of quality of that image to deliver to you a desert mirage. :-) The sans however is Olive, a font that some revile but I personally value a good deal.
"I think your desire to see something you like after having to look at Meridien for too long is leveraging the lack of quality of that image to deliver to you a desert mirage. :-)"
Ha! Nicely said. I guess I should go and unleash that visual hunger for good type on the work on *my* book. :-)
Olive, as in Antique Olive? That has been growing on me a lot, and does seem to look very nice in there. I used to detest it – probably due to over-exposure in crappy display settings (a bit like Kabel, of which I've become quite the fan).
Thanks, Hrant. Honestly, Aesthetically, I don't like the spikiness of Galliard, but it doesn't have the readability problems that Meridien has for me (and Nina.)
I can understand why Carter found Frutiger's shapes so fascinating. I was stuck a few years ago in a waiting in Charles de Gaul airport. It was bare, but the room had these six foot high numbers in Frutiger, labelling areas. In that size you could see every nuance, and admiring and studying the curves was like watching a movie for me, for over an hour while we waited.
In Rocky the spikes are blunted, so it becomes a different animal.
I wonder whether the original spacing of Meridian by Frutiger is also bad, because he was following too much his own theories, which in this case were not so good. --He argues on the Linotype site that the space between n's should be the same as the width of the counter. Tracy says it should be less. It may also be just that these long spikes don't allow good spacing.
I don't know exactly what is going wrong, and it is an intriguing question precisely because it is so well drawn.
Edit: here is the link to Frutiger's comment on spacing.
Kent, I'm wondering what you think of his comment on spacing.
> Linotype site (or was that MyFonts? In any case, I’d hope they have kerning activated in their Flash).
It was in fact the Linotype site.
One useful thing about using Trump as an example is that, since it was originally designed for Linotype composition, it has very few kerning pairs. (Metal Linotype, you may know, could not accommodate any kerns at all). There are none in this sequence. So the values I list can be taken as absolute, no need for math.
I would suspect the same is true for Meridien. So, it wouldn't matter whether the tester could accommodate kerning, since there probably isn't any.
In fact, I would say that in any reasonably competent text font, this particular sequence -- E i n b i l d u n g -- should not require any kerning pairs.
> This shows a remarkably similar pattern to what you posted, Kent (with the notable exception of n_g).
This is undoubtedly due to the different structures of the lower bowl, Trump having a horizontal left thrust & Meridien having a demur angle.
> Could this be a case in point for a combination of optical scaling plus tracking?
No. This is a case in point for leaving it alone and not using optical metrics.
Optical spacing deviation: So I've been looking deeper into InDesign's optical spacing, and I must say I'm deeply disenchanted with it. (I know I could [should?] have started a new thread on this, but since we were kind of talking about this anyway…)
1) The optical spacing algorythm seems to apply the same pattern Kent outlined above to serif fonts, across different fonts and sizes. Note the sequence "bild", which by opening up the straight/straights and tightening the straight/rounds (and vice versa) ends up as something closer to "bi ld".
Here's Swift at 10pt (top "Metrics"; bottom "Optical"). To my eyes, the beautiful visual rythm is completely thrown off by optical spacing.
More or less the same thing happening to Dante at 12pt (a pretty extreme case of "bi ld"):
2) To find out about the assumption that optical spacing can be beneficial for the special case of badly spaced fonts, I did a test with the most badly spaced font I have, which is my own work in progress – I actually haven't worked on its spacing yet, there's just something kind of there.
Again, InDesign applies a similar pattern as on the serif fonts above – and even here, I'm not sure the "optical" version is actually better. To name just one thing, the "ein" sequence looks pretty wacky to me.
3) Interestingly, the pattern that optical spacing applied to the serif faces above changes dramatically for sans faces. Below is Parisine at 10pt. Note how in the "bild" sequence, the straight/straight in the i_l combination is tightened a lot more than the straight/rounds surrounding it, making the "bild" sequence not read like "bi ld" as above, but closer to "b il d".
With this one, I must admit the optical version doesn't look quite as atrocious, so maybe optical spacing does work better for sansserifs?
It does startle me though that this was tightened up a lot even at 10pt, which I'm not sure is such a good idea.
4) This is just for giggles and to see how far the thing could go. It obviously "could not read that".
Kent,"No. This is a case in point for leaving it alone and not using optical metrics."
I probably expressed myself poorly there. What I meant was, could that mean that positive tracking was used in that sample on top of optical spacing*? (I'm not proposing doing that. :-) )
* I meant "spacing" of course, not "scaling".
(Although maybe they didn't even manually track that out, but let their H&Js take care of that …)
Nina, you don't think InDesign helped in #2?
BTW, #4: LOL!
Hrant, about #2: I admit I stumbled over the initial e_i so badly that it probably intensified my negative judgment a lot.
That said, I would agree the optical spacing helps in some places ("inb" tends to look better); but overall, I'm not happy.* I think it's doing too much of the "bi ld" thing (if it tightens the b_i and l_d like that, I'd want less air in the i_l too), and the d_u looks too tight to me too, relatively speaking. But maybe that's just my untrained eye (which is also still very tired of seeing this font).
* Of course, I'm not happy with the top version either. I can't really decide which one is worse; in any case, I'd hesitate to call the optically spaced version a crucial advancement.
> (Although maybe they didn’t even manually track that out, but let their H&Js take care of that …)
There you go. I doubt that there has been any wholesale tracking applied on top of the optical scaling. It's possible that the H&Js are liberal with letterspacing allowance. I'm not convinced that's true. But I'm not looking closely at the original.
I think you're correct in observing that the algorithm works better with sans serif designs. I have not explored it much in that realm. I imagine that the wide range and variety of serif structures can easily throw off the algorithm. For that reason, sans serif is probably more amenable to an algorithmic optical spacing routine. I'm not convinced the InDesign version is better than JFP's fitting, but it's certainly not as dramatically worse.
Regarding the adjustments to your type design: The thing is, while it might improve some things, it also worsens others at the same time. This is the crux of my argument -- it's unreliable, at best. Good to be somewhat disenchanted.
That doesn't mean it can't sometimes be helpful in a pinch. But evaluate the results closely; you can't always trust it.
@Bill: Do you mean this comment?:
"One ground for disagreement was the concept of rhythm in a type line. I tried to show, with the aid of an enlargement of Nicolas Jensen’s roman, that the interior spaces (counters) of the lower-case letters had the identical width of the spaces between the letters, which was easy to demonstrate in a Latin text with its many straight strokes. [. . .] Later on, I designed all my serif faces in accordance with this concept in order to avoid unevenness in the flow of reading."
This seems overly simplistic and I hesitate to take it at face value. Dwiggins used to make some cockamamie statements, too, about even "picket fence" spacing. Best not to take them too literally.
Kent, yes, that's the quote. I agree he couldn't have taken it completely literally, but I'm wondering whether he took literally enough to create problems in Meridien. I think I remember reading somewhere that he now thought he had made the original side bearings of Frutiger (Roissy) too wide for signage.
Well, I don't have enough evidence before me to offer an opinion about the fitting of Meridien in relation to that comment and whether it detracts from Meridien's suitability for text setting. I think I've established pretty convincingly that the sample before us is not the natural fitting for Meridien, which can't be blamed on the font.
After that, it's pretty hard to accurately assess other variables.
There have been allegations that Meridien is fitted too loose. I haven't seen a trustworthy sample from which to draw any conclusion.
The letterforms seem moderately wide. And wider letterforms generally beg for slightly more open fitting. When does it all become a problem . . . ?
The tester sample looks like reasonable fitting to me, but obviously that's no way to measure. I'd have to experience the type first hand.
The only other example I have at my disposal is a mediocre reproduction of something from the late 80s, presumably from phototype, and it's definitely set too tight (as was somewhat the habit then). Again, I don't trust that this can be attributed to the font itself.
Xtian made a plea above for examples of Meridien used well. I'll second that motion.
I have a friend who has Meridien (although it looks like
Florian has it too). What should I ask her to set for us?
I just asked on typografie.info for good examples of Meridien, too – maybe somebody there has something to share.
The best I could find was this sample from "Schriften erkennen" (Sauthoff, Wendt, & Willberg; 1998), which I think looks decent – at least a lot better than the book I posted about. Goes to prove how much comes down to spacing:
(Sorry, more German.)
Larger scan is here.
BTW, I just re-read the Meridien passages in Bringhurst. While, as is often quoted, he does recommend pairing Meridien with Frutiger,* he also seems pretty convinced that Meridien is not in fact a very good face for text.
While on p. 237, he quotes its (then) lacking glyph coverage only in this regard ("in the absence of small caps and text figures, the related Apollo is often more useful for text"), the part on Apollo then gets more illuminating (p. 215, cf. p. 238):
"[In making Apollo] Frutiger used the opportunity to rethink his first text face, Méridien, drawn eight years earlier. Apollo lacks the sharpness of Méridien, but its smaller eye, blunter serifs and reduced modulation can make it a better choice for text […]".
* For the record and since we were talking about this before: "Frutiger also mated well with the same designer's Méridien and Apollo, though such a mixture was not apparently part of the original design plan, and the fonts did not match in weight or body size." (p. 257)
Hrant, that is cool. Has your friend set anything with Meridien in the past that we may look at?
Linotype provides PDFs with sample settings of Meridien on its website: http://image.linotype.com/samples/hirespdf/12948.pdf
I don’t have Meridien*, but I do have Latin 725, which is Bitstream’s version of Meridien – again a very different beast. Nina, if you want to, I could provide a text sample of that.
*) That is, I did use Meridien once when at University, for two booklets. Not really book text, rather annotations/captions in narrow columns – I think it worked quite well for that. I’m fond of Meridien’s italics.
Here’s a comparison of the Regular and the Italic, rendered at 48pt. From top to bottom: Linotype Meridien, Bitstream Latin 725, Frutiger Serif:
It does seem to work pretty well in your booklet, Florian. Any chance you could post a PDF of a page or so? (Though I wouldn't want you to infringe on anyone's copyright of course.)
Also, how was your experience typesetting Meridien (after how Christian described how he struggled to make it look good)?
It's interesting that Bitstream's version is tighter and Frutiger Serif has been blown up on the body (most certainly to match Frutiger Sans).
So do I still need to ask my friend for a setting?
I wonder. Personally, I'd still be quite interested to see an example of Meridien
set well, maybe with the same text as in the sample I posted so we can compare
en détail?* On the other hand, interest in this thread seems to have kind of dwindled, so if I'm the only one who still wants to see this, I guess your friend doesn't need to worry.
* Which might be a bit evil since it's in German. I can offer to type that & email it over, so your friend would only have to copy-paste. Let me know.
So tracking is poison, eh? I must be going mad. What happended in the last line of the sample above (Frutiger Serif Regular)?
Is it a space or what?
I’m at the University today. They have Meridien here, so I could set some text with it and upload a PDF. Nina, if you want it to be the same text as in your sample, please send it to me via e-mail soon.
@Jan (paragraph): The Frutiger Serif sample is taken from the specimen PDF on the Linotype website. I’d advise you to download the PDF yourself and print it – it’s never a good idea to judge the finer points of type on screen. The wonky spacing is most likely due to rendering errors in the process.
Wicked. Thanks for the offer, Florian! I just sent you an email.
Okay, here’s a PDF [0.8mb] with a replication of the two text samples you sent me.
I hope it helps to get an impression of how Meridien looks like out-of-the-box.
Pretty much InDesign default settings, except for the values stated.
I like the colour of the Medium weight better (depending on the way of production, of course). IMO, the ‘f’ in the Bitstream version is an improvement, so I included that one, too.
Thanks, Florian, the PDF sample spacing is fine.
Highly interesting. Thanks Florian.
I'm still not a fan of Meridien, but in your sample it does seem to read much more fluidly, so I'm pretty impressed with how much really comes down to good spacing (and/or how badly the Adobe "optical spacing" can mess up a well spaced font).
There are still some combinations in the metrical setting that seem strangely spaced to me (like the "c" followed by a straight ascender as in the "h" or "k"), but that might just be me.
One thing I'm wondering about Adobe's optical spacing, based on seeing this, is if it might use excessive care not to let anything touch (or if that really comes down to the book I have being tracked out on top of optically spaced). In examples like the following (sorry about the small size and bad quality), the spacing seems to place more importance on equal distance between serifs than on the actual optical space between letters.
For example, in "Knoblauchkette" it looks as if the "K/n" serifs were made to be at approximately the same distance as the serifs in the "n", which obviously makes the space in between the letters too big. Similar stuff is going on in the samples below.
(Although – isn't this last one a hint too tight in the metrical setting?)
Once I look really closely at the spacing of many established commercial fonts, I see all sorts of mischief.
Should I be saddened by this or cheered up? (That I am not the only slack bastard with atrocious spacing :)
Florian, thank you!
No biggie, but if/when you get a chance
you might want to set all that left-aligned.
It's interesting how much variance there is between the two cuts.
Thanks, that's interesting. I'm more and more thinking that my problems with Meridian are not the spacing, but way the horizontal strokes are treated. There's high contrast, and almost all of the horizontal is thin, and those serifs like needles. I think maybe its that the horizontals are starved, and at text sizes it is just too thin to read comfortably. The medium partly remedies the problem, but then it has too much sparkle...
Was it designed for metal? Or for film type? Maybe in those days it worked better, but for me the digital doesn't work as text.
Nina said: ike the “c” followed by a straight ascender as in the “h” or “k”
I assume you think they’re too tightly spaced, right?
Have a look at this thread about ch & ck in German, and this related discussion at Flickr. After all, Meridien is a child of the 1950s. I’ve seen faces from that period with much tighter ch and ck pairs.
BTW, Meridien does with almost no kerning. The Roman has a mere 94 kerning pairs, all of them for combinations involving uppercase and/or punctuation. I’d think this is quite little, and typical for fonts stemming from the pre-digital era.
@William: I wonder if Opera’s horizontals are more robust? Who has a sample?
From Myfonts:Monotype Apollo can be seen as the Monotype version of this design; Sofratype’s Opera is a newstext version.
@Hrant: Okay, here’s a setting, flush-left/ragged right: meridien-left.pdf [0.8mb]. With InDesign’s hyphenation slider set all the way to the left, i.e. ‘better spacing’ over ‘fewer hyphens’. Additionally, a version with some quick line-break corrections. I hope that is what you were looking for.
Florian, sorry for making you do that! :-/
But thanks a lot - it's the only way I for one can see anything.
The Lino Regular is a bit light for its spacing and proportions.
The Lino Medium is well balanced, but shouldn't be used above 10 point.
The BT Medium is essentially a display font (due to its dark tightness).
None of them are paticularly well-spaced (maybe due to light kerning).
"Have a look at this thread about ch & ck in German"
Oh, wow. Thanks for an entirely new perspective. I agree with Ralf's statement that to younger practitioners this just looks weird (which is what happened to me here), though I disagree with his stance on blackletter, and consequently feel a bit bad about this as I'm actually quite familiar with the blackletter ligatures but failed to make the connection. What do you think, will this language-specific exception be lost over time?
Also, thanks for the new PDF. I think the font looks a lot better RR than in the justified setting.
That said:"Meridien does with almost no kerning""None of them are particularly well-spaced"
I'm sort of relieved to hear you guys say that; I was cautiously wondering. Frutiger does have a semi-god kind of status around here (especially at my type design school, where he studied), which tends to make me wary about thinking anything non-optimal about anything he did – which must be BS, as he's human after all.
"The Lino Regular is a bit light for its spacing and proportions"
I'm glad I'm not alone in this!
Quite human, I believe.
Typophile: the sacred cow abattoir.
following without tracking so as not to disrupt interletter spacing :-)
Chris! That almost made me choke. :-D
And it's a reference the afterworld will likely never get…
I'm going to ask an unrelated question:
Why do some of you think you need a license in order to lawfully use Meridien for private study and critique?
I just checked out the link from hrant
BTW, here it is - the Jackson Pollock of Typophile threads:http://www.typophile.com/node/32680
Stopped looking when I hit the My Little Pony pics! ;o)
Florian's link to the digraph thread brought back some vivid memories for me, of my time studying calligraphy and design with Friedrich Neugebauer in 1976-77. Although I was happy to follow along with all of his close-spacing suggestions when we were doing blackletter, we kept having tussles about ch and ck when I was lettering in roman and italic. He was thoroughly accustomed to the use of the digraphs because he had encountered them all his life in his daily reading and therefore felt it was right and proper to write those characters closely; whenever he asked me to follow suit with this, I would argue that when I went back home and was lettering for non-German readers, I would need to put more space between them!
Unless I am mistaken, isn't it true that in German language, if you see a "c" it is always followed by an "h" or "k"?
Also, concerning the "picket fence" spacing that Kent mentioned, to my eye that has never seemed like an ideal worth striving for in anything but blackletter. In fact if followed too literally it becomes e x t r e m e l y b o r i n g a n d r e p e t i t i v e ! ! ! (Perhaps it could be the basis of a new font called "Sominex" . . .)
Here's an interesting metaphor that I got from Neugebauer (and he told me he learned it from Larisch):
As you write your letters, think of pouring water between them. As the water accumulates on the baseline and then rises to the top of the string of characters (either x-height or cap height), approximately the same amount of water should be consumed between all the letter pairs. While one should never follow this slavishly or literally, I think it is a more useful and more concrete way to think about letter spacing than "area".
You guys may disagree with this approach in thinking about typographic spacing, but over the years as I have introduced students to the making of letters with pen and ink it's been a real help.
isn’t it true that in German language, if you see a “c” it is always followed by an “h” or “k”?
Bruce, basically that’s correct. But German language is full of loan words and names from French, English, Greek, Latin and other languages – and has always been. So, a ‘c’ followed by something else than an ‘h’ or a ‘k’ is actually pretty common.
FWIW, we were just discussing the water metaphor in spacing here:http://www.typophile.com/node/56000
Ray Nash, who taught several generations of typographers from his studio at Dartmouth College in New Hampshire, also used the water idea to explain letter spacing. It is interesting to note that this notion has been mentioned from several sources both here and the threat altaira referenced. I never got around to mentioning Nash while reading that thread.