Perrin OT

nepenthe's picture

Hello. This is my first stab at font building. It is not an original design, but my attempt at recreating the fonts found in the Oeuvres de Descartes series published by J. Vrin in the 1970s. It was pointed out to me by someone named Claude on a fonts newsgroup that the Titling capitals were derived from the work of Louis Perrin (The Lyons Capitals). If any one has any additional information on these fonts or any critique of how to improve the design, I would really appreciate it! Some of the glyphs (like the italic `k') I could not find in the book, and I am not quite sure how to make it fit with the other letters. Also, I know that the Italic caps are butt ugly, because I haven't been much interested in improving them (they rarely come up in use for me). But I'd still appreciate suggestions on them. You people seem to know a lot about good typography!

application/pdfPDF Sample Specimen
EPerrinPro.pdf (69.7 k)

hrant's picture

> This is my first stab at font building.

Wow. Are you sure? ;-)

A while back I translated a conference presentation by Ren

jfp's picture

Its a good typeface but not Perrin revival. It sound more, from what i see in your interpretation, a Deberny & Peignot or Fonderie Bertrand Elzevir. I can't find exact reference, but perhaps with a scan of your sources, it will be more easy.

In the past, we have done during the course, some typefaces based on similar sources: look at Jannet and Bertrand samples.

hrant's picture

I did some comparisons, and it seems your text face is close to Baudoire's stuff*, although not exactly. The thing is, I could swear I've seen that characteristic narrow-top "e" somewhere before, and for some reason it makes me think of ATF. Dunno.

* Which Audin notes was actually a copy of the Augustaux in the UC but not the lc. So if the titling and body face in that book were part of the same font, the likelihood that it is a Baudoire (but how did they print it in the 1970s - was it a facsimile of a much older work?) I would say increases.


jfp's picture

sound really possible Hrant. good research.

nepenthe's picture

Access to Scans

Hrant and Jean, thanks so much for the input! I have put together a page of scans as well as a link to the PDF that is a bit updated (I noticed a few details that needed to be changed). It is interesting that you describe these kinds of fonts as being a bit wunky, which is what initially inspired me to make them. I like escpecially the narrow top on the e and c, as well as the backward slope of the b, the lovely g, etc. But especially I love the Capital R.

The first time I saw the titling font was on the cover of Pensees Diverses sur la Comete. I noticed the spine while browsing through the philosophy books at school. I just absolutely loved these letters, especially the capital R. The cover of this book is what first got me interested in typography last winter!

In the same section I found the series on Descartes, which is quite an important resource for Descartes scholars. The editions definitely look like they were produced photographically to me and contains many different fonts. One of the problems in identifying the fonts is that there are so many different ones used in the series.

I chose to work from Tome Six, since this is the cleanest copy we have. I based the font off the largest of the type, which is about 14pt. The original's italic has a noticeably lower x-height than the regular; I kept it lower but not as much as in the original.

The font in the introduction and in the footnotes is different than the main body text, but I did not reference this in the design (in fact the small fonts seem to resemble very much JFP's Bertrand). Also, the small caps do not seem to be based on the capitals, but are possibly from another typeface. I decided to retain their flavour anyway. The original numerals are very ugly, and I tried to make them prettier (this is where I made the most change).

Interestingly, the italics used at different sizes all seem to be the same design, but drawn slightly differently. For example, the smaller ones are curvier, while the stems on the full size are straighter.

I hope that we can achieve a correct identification so that it can be named properly and I can learn more of the history of the designs. I would like to look at those books that you recommended, Hrant, but they are typography books are not available here in Winnipeg. Hopefully I will be able to buy some books at some point, but they are generally prohibitively expensive (I am a philosophy student). If you can post any scans from your books I would really appreciate it!

hrant's picture

Good scans.

That "a" is really shockingly contemporary, and I guess you did a great job with it after all. The only thing I'd change would be to firm up the very outside-bottom-left.

> I love the Capital R.

It's a charming sort of structure - many people like it, including Roger Black from what I've noted.

> The cover of this book is what first got
> me interested in typography last winter!

Life is strange!

> [numerals are] where I made the most change

I can't see what the original numerals look like (the only ones I see, on the "original cover", seem fine, and similar to your new ones actually!) but one thing I would suggest is to make them "classic French" style, not conventional, since this is after all a historical font - so the "3" and "5" ascending rather than descending. Look at the old Didot stuff - quite nice:

I would have bowed the diagonal of the "4" though... and I will, heh heh... :->

As for scans, I will gladly post them - just know that they're from photocopies of reproductions of who knows what! I'll have to do it tomorrow though.


hrant's picture

Whew - that was more work than I anticipated!

This is almost everything I have. All the scans are 300 dpi, gray balanced. The filesizes are about as small as I could make them without losing important detail.

Deberny's "Romain Ancien", from F Thibaudeau's "La Lettre d'Imprimerie", p 379.
(This scan is cropped - the top half was mostly redundant.)

Various Elzevirs, from Thibaudea, p 447.
111: Beaudoire
112: ?
113 & 114: Deberny
115: ?

The Beaudoire from Audin '28.

The Deberny from Audin '28.

The Perrin caps from Audin '72, p 244.

The Perrin from Ovink, p 284.
(The chopped glyph is just an "el".)

More Perrin from Ovink.

The Beaudoire from Ovink.

The Jannet:
I'll wait for the Typophile archives to be properly accessible.


nepenthe's picture

You're amazing! I don't know how you can know so much about type. Your scan of The Beaudoire from Ovink is exactly the same one in the book, numerals and all. In my book there are the same 11pt and the same 14pt. The only differences are that in the Descartes books the lc a, c and s in the 14pt size are different. I guess Floch (who printed the book) made his own?

Googling `ovink beaudoire' turns up nothing useful. Can you suggest to me where I can get access to this book? Is it purchase-able? I am so happy, because I have been wanting to know the story of these designs since I first saw this book.

hrant's picture

I don't know that much - collecting stuff simply helps one appear smart.
Audin was the one who knew stuff. But I'm glad I could help.

The Ovink thing is a 3-part article titled "Nineteenth-century reactions against the didone type model", published in the Dutch journal Quærendo - a really useful series. The Beaudoire bit is in part 2. I don't seem to have kept an exact reference for that, but fishing through some old emails I can see that part three is in V2 #2 (1972). If you have trouble finding it, try to find an index to Quærendo (I know they published indeces, plus it might be online), or look through the complete Ovink bibliography in Quaerendo Volume 14 #1 (1984), pp 8-17. If you still can't find it, let me know.

So are you trying to find a good name for the font? It's tricky, because Beaudoire's work* was apparently derivative to some extent. But I wouldn't call it Perrin [any more]. The only thing I can think of is "Théophile". Or maybe something related to mid/late-19th century France.

* At least that design; there's actually a font in Audin '28 by Beaudoire that's a quite charming upright cursive - even though Audin himself calls it "bien laide" (quite ugly).

BTW, check this out:
But looking through my stuff, that "g" -and the lc in general- makes it look closer to the "Old Style" of Miller & Richard of Edinburgh (an "Anglo" take on the Elzevir) than the French stuff. And tellingly, the same guy who made Romana made ATF De Vinne!


hrant's picture

Looking closer at Ovink's Beaudoire showing, two things strike me:
- Of the four fonts (11 and 14, roman and italic for each) the 14 Roman is different than the others, most obviously in the numerals*, but also in the "a" and pretty much all-around.
- Considering what the caption says, "Marquet" might actually be a good name for this face!

* BTW, that Perrin "5" is one of the sexiest glyphs I've seen in a long time.


nepenthe's picture

Good Point. I will will rename it thusly. Also, your numerals are came out really clear, so I will use them as a reference.

Hrant, is this your website? You make so many posts here, but I don't see your name on the members list. I wonder if you are some kind of typography angel?

hrant's picture

> is this your website?

Now that's funny.
Angel? I guess it's worth noting here that in Armenian culture the Grim Reaper is considered one.

The numerals: Tell me exactly which one(s) and I'll make 600dpi (my max optical) scans for you. Although if you got the actual Qu

speter's picture

I wonder if you are some kind of typography angel?

Actually, in typography they are known as devils!

(And Hrant, not just in Armenian culture: "the angel of death")

hrant's picture

> "the angel of death"

Oh, yes - I forgot.
The literal translation of ours is "soultaker angel".


nepenthe's picture

Hmmm... looks like no library in Winnipeg has Quaerendo, and they sell online for $150 per volume. If you can just scan that whole page at 600 dpi I'd be thrilled, but if you can't, then I am wanting the 14pt upright numerals.

hrant's picture

A raw 600dpi scan of the Beaudoire from Ovink, saved JPEG(7).

The Perrin-style numerals from that page, scanned at 1200dpi* and gray balanced.

* I've rarely used this before, because I don't trust it too much (plus I don't like the asymmetry), but my scanner is supposed to do 1200 in one dimension (the drive's) and 600 in the other (the element's). I tried both directions and this was less bad.


nepenthe's picture

Taking into account the wealth of new information courtesy of Hrant, I have redesigned the numerals and the Italic caps to make Marquet as a replacement for the inappropriately named Perrin. You can check out the new PDF at my site:

hrant's picture

This is really looking quite good. It looks like we'll be getting perhaps the best digital Elzevir around! It's especially nice to see titling caps in conjuction.

I can't tell for sure (because your text blocks are justified) but it looks like your blank space is too wide.

- Note that those are the Perrin-style ones.
- I think the proportional divergence is cool, but the "8" might be going too far.


BTW, I hope our frantic historical mining hasn't distracted others from doing a critique of this design on its own merits!


nepenthe's picture

Hrant, my school tells me that they can get Quaerendo from another university. But I cannot find the index for which issue part two of the series is in... is there any way you can find this out? Does anyone else know the specific issue and pages that the bit on Beaudoire is in?

Also, I'm not sure what you mean about the eight. I thought that the design is too heavy on top, which is one of the reasons why I thought the numbers are overall quite unpleasant looking. Do you think that I should make it more properly balanced?

hrant's picture

I used to live off of Inter-Library Loans! :-) Then I sort of ran out of things to request. Well, not really, stuff still pops up once in a while.

It suddenly hit me that I have a copy of Ovink's obituary (he was da man), which contains a comprehensive bibliography of his writings*. So the Elzevir stuff is in Volume 1, #4, pp 282-301 (1971).

* And actually looking through Part 3 of the series I now see a reference to part two with a full citation.

The numerals:
I don't know if it's better to go with the Perrin style or the Beaudoire/Marquet style. Or maybe something else entirely, even your own design. But even if (or maybe especially if) you go with the Perrin style, don't be too shy about fixing things. When I look at what you had in your most-recent iteration, it strikes me that the proportional divergences are quaint and probably in character, but the "8" simply is too ugly because of its huge top.

BTW, I got some more goodies for you:
While I was looking for something else I ran into the single most important publication about Perrin that I know of: Audin's book on the 1923 Perrin exhibition in Lyon. It's quite rare - it seems only 61 copies were printed. There's a very extensive text (120 pages), a complete catalog of works, and some great facsimiles (as well as actual prints -like pressmarks- from Perrin's own engravings). The paper is very yellowed though. There are two things in there that will probably interesting you most:

1) A facsimile of Perrin's famous specimen sheet, showing two sizes that are basically Marquet's designs: the 11 and the second 14. Here are scans:
The second scan is a collage from two sections of the sheet.
If you want more details about the specimen let me know.

2) This is cool. A character set of floriated caps, from "A" to "V". Here's the "R":
BTW, that's a RIP-killer right there...

Concerning the Marquet fonts, I guess it's safe to conclude that Perrin spliced in his own caps and numerals - although I can't be 100% sure. And Beaudoire seems to have partly parroted this mix-and-match deal for some reason.


hrant's picture

About the numerals again:
If you look back to Perrin's original drawings
you'll notice some differences from what was actually made in metal, including a much better "8". Maybe he changed his mind, or maybe his punchcutter* had his own ideas (not unusual), but if you're concerned about historical fidelity in terms of changing the "8" the precedent is actually there.

* Francisque Rey.

BTW, I've moved those most-recent three images to the "Elzevir" subfolder, so they're now:


jtombeur's picture

In a hurry...
Frank Jalleau made a font.

hrant's picture

And how many potential benefactors do you have who only read Armenian? Come to think of it, how many people in the world only read Armenian? Pretty close to zero.

Which doesn't mean that anybody who cares an iota about type shouldn't sign the petition!


jtombeur's picture

> [Hrant, it's not a question of who reads Armenian. It's a question of who cares about either the Armenian language or the Cabinet des poin

nepenthe's picture

Thanks for the link to that book, Jef. I will be interested in buying it if I can manage to cough up the seventy buck or so it would take to get it to Canada.

Also, I signed the petition. That collection seems like a gem and I hope some day I can get to see it if it is not thrown away by then...

Also, I'll post my final PDF when I get the chance, but now I am studying for exams!

nepenthe's picture

The nice final version is now done, but the specimen sheet is still pretty plain. I made the titling numbers based closely on the nice Perrin numerals from one of Hrant's scans.

Since there has been very little comment on the design itself, I sould welcome it. Especially for some of the finer technical details you might notice.

Hrant, I remember that you suggested to `firm-up' the very bottom outside left of the lc `a', but I do not know what you mean. I did change it a bit, so maybe you can take a look at it and tell me if it still needs to be changed. Also, I did try reducing the word spacing bit and I do think that it looks nicer this way. Do you think that it is too close now, or just right?

nepenthe's picture

I just realized that the spacing still looks wide, which I think might be because InDesign is not very good at hyphenating French and Latin. So I uploaded a new version with English on the last page and standard leading so you can see how the colour looks in that setting.
Specimen Sheet

hrant's picture

I don't get why nobody is reacting.
Did I come down too heavy before and scare people?
Come on, guys.


The "a": looking at it again, I think it's just fine. I was being too Modernist before.

The new "8"s are great.

- The tail of the "Q" is frail.
- The head of the "t" is out of character.

- The descenders of "p" and "q" are too long.

That BP sign is very nice, and that Asterisk in the Expert set surprising (quite possibly in a good way).

Spacing: it needs work (like look at the right side of the "t") but the overall tightness and the size of the blank space seem good.


Jon Whipple's picture

I think it's because we may have been rendered speechless. I have never seen a G with such an articulated spur. I am fascinated. The ? 3 5 9 are gorgeous and I like the dagger.

Comparing the scans from Hrant and notice that the lc s is different. The scans show an almost ball terminal and you are going down a serif path. This is intentional?

The asterisk in the expert set make me grin. I felt a ping.

I agree with Hrant that the Q's tail is too frail. It definitely needs a little more thickness for its length. But then it would be too frail if it was shorter (DON'T make it shorter) so some more weight would most likely help. Add it at the extrema of the curve.

The tails on p and q look good in the alphabet presentation, but are too long when you look at them in lines of text.

I think this looks great.


nepenthe's picture

Thanks for the critiques. Right now I am on a holiday and so I cannot work on the design. But I will definitely make a list of the suggestions and try their implementation when I get home.

Foremost will be the spacing and kerning of the characters. I have been meaning to redo this entirely, since I figured how to achieve better spacing with less kerning while making the italic; but this will be a big task for the number of glyphs in the roman (regular, sc, titling, accents...). I didn't notice before how wide the spacing on the `t' is, so any more comments about spacing woulr be great. It is hard to tell what is off after I have become so used to the spacing.

Looking at the `Q', I can see what you mean about the tail being too frail. Sometimes I was unsure about whether to follow the perrin, beaudoire or version from Les Oevres de Descartes most closely. But I from now on I will try to just balance out the family over all.

Jon: The lc `s', c e and a are based on the ones in Oevres de Descartes, which are somewhat different than all the others. I think that they look much better anyway, especially the s which matches more nicely to the uc `S.' But having ball terminals does match better with the italic; which do you think looks better in the roman?

Re: the asterisk, I wish that I could take credit for this, but it is exactly the one in Oevres de Descartes. It is one of my favorite glyphs from there since it is different from most. But thanks about the dagger and the BP glyph. Those are ones I actually did make.

My descenders on p and q from the italic are actually a bit shorter than from the orinigal ones. I know that they do not blend perfectly with the roman for this reason, but in the original the two were never set on the same line. When I did make them shorter, I felt that a lost of their distinct character was lost. When I get back I will try making it with shorter descenders on those glyphs and you guys can tell me what you think.

Hrant: I do not know exactly what you mean about the top of the t being out of character. Can you elaborate? Do you mean that the sloped part should be more curvy than it is now?

grod's picture

Beautiful. Please, set the entire text of the Meditations in French and English and make it available as a free pdf download :-) I've done a lot of work on Descartes, the Cartesian circle in particular, so it would be really cool to have a copy in such a gorgeous face.

hrant's picture

> from now on I will try to just balance out the family over all.

Good idea.

> When I did make them shorter, I felt that a lost of their distinct character was lost.

As in life itself, the individual-versus-whole dilemma is at the core of type design.

It's hard to say where a good balance is. Just know the issues (like proper distribution of Cartesian* area usage, glyph easthetics, etc.) and take the design decision that makes most sense to you. Just one thing: besides the quite large depth, there's the issue of the descenders being different lengths. This could actually be a plus, but know that it's quite unconventional, especially since the ascenders are pretty much even.

* Speak of the devil!

> the top of the t being out of character

Not sure. Try making it concave and more rounded at the edges.


nepenthe's picture

Noah: I actually did an OCR of the seminal 1967 Ross edition of the Meditations (English). I did not do the original latin or french editions though, because these books from school do not open up all the way to fit in the scanner. Currently the PDF of the meditations is not online, but I can make available when I get back. I find it really useful to have this in PDF so that if I want to write a critique of Descartes description of God as infinite, I can just do a search of it for `infinite', `infinity', etc.

If anyone wants to see the latest online version of the font in action you will have to read my last paper on Hume. (Any philosophers here will have to take it easy on me because I am only in second year at university!)

application/pdfJP's Hume essay (set in Marquet w/ LaTeX)
hume-essay.pdf (71.6 k)

eomine's picture

I think it's just harder to comment on a revival. For example, I think the italics do have a nice structure; they're not 'calligraphic', but on the other hand I think it looks odd (especially when compared to the roman). Maybe because of the weight distribution.

The figures look very quirky to me, and they're also set too tight (look for '1993' or '1980' in your last sample, for example).

My point is, these are characteristics that you're bringing from the original typeface, so you'll eventually have to choose between being faithful to the original and adapting it even though it may require you to remove some of its quirks.

nepenthe's picture

Eduardo, you're right about the figures being too tight. I have trouble getting them right because they are oddly proportioned compared to the regular text. Similarly with the ascender/descender in the italic, I will have to take another good look at the original and rethink how I want to deal with it in relation to the rest of the glyphs in the family. Thanks again for all your input!

nepenthe's picture

application/pdfA major update
marquet-specimen2.pdf (202.6 k)

I've been having trouble connecting to the internet from home lately, so I've decided to burn the new file to disk and upload it from school. I've just printed it as well and am very happy with the results.

I have decided to keep the traditional proportions on the italic. If I ever feel ambitious I will produce a second version with shorter ascender/descenders like in Trinit

Forrest L Norvell's picture

I love the titling characters; the Q and the R in particular are almost startlingly elegant. I also love the lowercase z. Overall, your text setting makes the font look unusually graceful. There is one major stumbling block for me, though, which is the lowercase r

nepenthe's picture


Thanks very much for the insights. I actually kind of like the steep angle on the r, but I will take a look at it and try to improve it. The square brackets are indeed too thin. I have not edited them in a long time and since I made them the rest of the font has become darker and more consistent. So I did revise these to still be thin, but now more in proportion to the rest.

Also, good points about the italic. I've been hating the lc k for some time now, but wasn't sure what to do with it. I recently came across a book that was set in Fournier, and seeing how similar it is Marquet in many ways, I decided to adopt a similar style. There was no lc k in the original, so finding this was helpful. I also modified the uc K to be more similar to the R and Q as well. Regarding the y, you are quite right. I would say that overall the spacing on the italic is not very good because I just haven't had the time to do enough work on it. I find that the cursive was easier to draw but more difficult to space!

After much thought, I also decided to revise the ascenders and descenders to be a bit more conservative. I am still a bit uncomfortable with the ascenders of the longs glyphs, so I will change them to be shaped more like the r.

I'm also working on expanding the titling set to include a lc that is sort of a blend of the Beaudoire lc and one by Perrin that matches his Titling figures more closely. I will post that in the Display section when more is done. It looks very French!

Forrest L Norvell's picture

I have to reiterate that the titling R and Q are unbelievably hot. I've never seen anything like them, but they aren't so mannered they draw undue attention to themselves. They are also, after looking around, incredibly French. That said, a little reflection tells me you need to work on the upper right curve on the R some more; it looks a little squashed.

Keep working on this one! I really like it!

nepenthe's picture


I'm glad about your comments about the Q and R being so hot :-) The R from this design was what got me interested in typography. I have spent countless hours on the titling caps trying to achieve maximum hotness and I really don't think that I can change any of them for the better at this point without disrupting the orignal character. But I can see that the part you mention declines more swiftly than its P counterpart. I will revisit this letter as I continue to work the display set. I am more sure now that I will want keep these as a separate font altogether rather than put them in the main font as a Titling alternate.

I am also continually bothered by the italic even after having fixed the k, descenders, etc. That is primarily why I haven't posted any more updates. Lately I have been working on rewriting the opentype code to make it more elegant and to allow for more comprehensive kerning classes on the roman. Once it is figured out how to do this properly, I will apply the same methods to the Italic. And if anybody can tell me how to fix the ugly w, please let me know!

BTW, Thomas, if you read this, thanks again for posting those OpenType feature files on the Adobe site. I am just in awe of how perfect and sophisticated your fonts are!

Forrest L Norvell's picture

I think the w and the v are both fine. I had to search through your Hume paper (what did your professor think of the over-the-top typography, by the way?) to find the italic w in a setting, but there it made perfect sense. I think in general this face could stand to be letterspaced a little more loosely, and if so, you could widen the italics a little and lower the contrast. I think my objections to the y would diminish with a little outward curvature in the descender coupled with a thicker stroke.

I think the italics would be easier to space if the overall letter spacing were looser, too. One of the things I've been learning lately is that we (contemporary users and designers of type) like to bunch our letters up much more tightly than in the past; if this is a revival typeface (which it still kind of is), maybe you should try revival spacing as well.

_Palatine_'s picture

I hate to revive a thread this old, but I'm absolutely captivated by this pdf:

Has this font been released? If so, under what name?



Stefan Seifert's picture

Infact, as hrant said in the first return:
This is great stuff.

Very classic and excellent specimen style!


Stefan Seifert's picture

Brillant title!!

Tomi from Suomi's picture

I just remembered that I made a bold version of Fournier, and I've also forgotten that I'm talking with Monotype with turning this into a font family:

This is a font with plenty of variation.

_Palatine_'s picture

@ Tomi

A Fournier revival that is actually usable?? I have to pinch myself . . .

Please get that released as soon as possible! Would love to license it.

Tomi from Suomi's picture

Hello, Christian-

This will take some time; I've already studied for some time my copy of "Modéles des Caracteres de l'Imprimerie"; and as you can see, I've already made a bold version of the typeface.

But to make a usable revival of Fournier is like making a mould from brilliantly formed child, while he is growing.

Stefan Seifert's picture

Hello Tomi,

you have copies of Imprimerie specimen? Let us see some JPG.
I have an old letterpressed probe of Garamond by Imprimerie.


piccic's picture

Tracking… :-)

Syndicate content Syndicate content