Hilfea : first step on Francesco Griffo typefaces

ivan louette

Hi ! I was working for about a week on another Quatrocento project. Some times ago I designed Uccello based on the Jenson Roman. Now I am trying to do something around Griffo's typeface for Pietro Bembo. While I had fairly high resolution of Jenson prints I don't have of Griffo's one. However the reproductions of De Aetna on Archive.org are a good base to improvize something different. Another interesting challenge in my learning curve.

At the moment there is no kerning ; this is more visible on capitals.

Thanks for any comment !

John Savard

It is beautiful, although you may want to give the credit for that to Griffo rather than yourself.

ivan louette

@John Savard  Thanks for your comment. Be reassured that I don't give credit for myself here. The only problem is that I may not give it, for example, the name "Bembo" because it's already in use and registered, like that was the case for "Jenson",which forced me to give it the name "Uccello". I am too grateful to Jenson and Griffo as the so great teachers they are for me And at the moment I feel still very far from what looks really Griffo's font for De Aetna. My next goal, after working on characters spacing is enhancing details to make them better melt togheter into words and lines as the Renaissance type engravers did so nicely. I spent a long time at that task on Uccello.

Dave Crossland

I recommend you read Frank Blokland's https://www.lettermodel.org and play with the software he developed/inspired to operationalize his theory.

ivan louette

@Dave Crossland Yes you are right ! That could help me a lot. Thanks ! However it seems that Griffo used different systems for De Aetnea and Polyphillus.

Craig Eliason

I'd say the same here as I've said on previous threads of yours: seems light and delicate, which makes it lovely at larger sizes but perhaps less useful for running text sizes. 

I like these threads because they lead my eye to quirky details I'd overlooked (here like the difference in /B serif lengths, the tall /c, the different heights of /m humps...)

Linus Romer

I am not sure if it is intended, but your typeface looks rather rough and sketchy to me. From far or at low resolution it looks quite fine (like "de ætna") but I would recommend you to work much more on the outlines. Let us take a look on the \S for example. Your version has some bumps in the outline, but you could have done something like this (the choice of serif style is not important here):

Compared to an \S from «de ætna» it still owns the same spirit and feels the same at low resolution:

Take the time to improve your outlines starting from your sketchy version:

• make curves less bumpy by adjusting the bezier curves (you may take the help of "Harmonize" or something similar) and/or removing unneccessary points
  
• reconsider the thin lines and bows: e.g. the top of the \o is significantly stronger than the bottom right of the \q at the moment
  
• make the overshoot of \C,\O,\G bigger
• make the serifs more consistent (e.g. the serifs of the \S are stronger than the serifs of the \T)
  
• rework the \A and the \B (consider taking other examples of \A and \B from the original)
• do not bend the \h that much at the bottom right

Keep it up, you are on a good way!

ivan louette

@Craig Eliason The case is somewhat different from my previous thread about "Renner Type". I must reckognize that my method produces too skinny results at first steps. But Griffo's font for De Aetna is also more slender than his other productions for Manuce. At the moment I am testing with some succes extrapolations to other weights. But at the same time I try to understand what I can improve at the x height area to give more life and more strength to this fairly light font and to melt characters enough togheter into words and to better draw the eye.

@Linus Romer many thanks for your detailed comments ! Of course this first step is somewhat rough and schematic. I only began today to improve details like serifs and refined curves transitions. However it seems that Griffo himself went to the essentials and did not use "fioritures" as some engravers and designers did in next centuries. Paradoxally that's the first time I work on a so stylized and "modern" font ! But that's also the reason why subtle adjustments like the ones you told are so important.

John Savard

Do remember to save your work. Sometimes one tries to improve something, and ends up making it worse.

I am no type designer myself, so the fact that I did not notice what Linus Romer did means nothing in itself, although I wonder at how he could have seen what he did from the images you provided.

Your images are a bit larger than they appear on the page, but that still didn't make any roughness or bumpiness visible to me. I thought perhaps he downloaded a beta version of your typeface, but except for a free dingbat font you designed, none of your typeface designs seem at the moment to have any existence as free or commercial fonts.

John Savard

One minor nitpick I did have with the typeface was the way the lower-case h bulged out on the right. I've just checked now: while neither Bembo nor IBM's Aldine Roman had this characteristic, the original did include it, so I assumed it was there for authenticity.

Well, sort of. The bulge seems to be subtle when it is present, and, as well, it appears, both in the case of De Aetna and in the case of Hypnerotomachia Poliphilii that both kinds of lowercase h were present, mixed, in the typecase.

ivan louette

@John Savard Yes from time to time I save the steps of my works when I have some satisfaction, and to avoid worse evolution.

At the moment the images I provide are the source of the comments. My fonts will be free and open source. I registered a domain name some weeks ago and will create a micro-website to share them. Type designers will be free to improve them to their taste

About /h bulge it's also present in other Renaissance typefaces. And there is also a (less visible) /n right bulge in Jenson and Griffo typefaces. While /m doesn't show it, and contrarily the right vertical seems very slightly slanted to the left in the two cases. In Jenson and Griffo the right part of the /m is also distinctly lower and in the two cases the /h hump is slightly higher and wider than those of /n and /m. What I must do is not exagerate these features, and it's not so easy ! However I tested to remove them when I worked on Uccello (my jensonian trial) and the overall look was completely broken.

On the other hand there are (very worse to my taste) a beak on /e and a prolongated top end on/r in Morrison and numeric Bembo which are NOT present in Griffo's design. The /e beak seems erroneously inspirated from Jenson's design where it's assumed (coming from manuscripts) and harmoniously integrated. But other Renaissance designers like De Libri and De bruschis didn't use it.

John Savard

I hadn't noticed the less visible right bulge on the n at all when looking at the Aldine originals. My suggestion to you would be to reduce the right bulge on the h to about 2/3rds of its current value; the result would still be slightly more of a right bulge than the actual Griffo design seems to me to have had, but I think that can be justified on the grounds that if one is going to preserve this characteristic of the original typeface, it can't be considered preserved if it's invisible in the context of current printing processes.

I think that will be enough so that the typeface will still be useable for general-purpose typography given current tastes; too much of a bulge, and people will tend, on some level, to react by thinking of the typeface as a sort of latter-day uncial.

ivan louette said:

On the other hand there are (very worse to my taste) a beak on /e and a prolongated top end on/r in Morrison and numeric Bembo which are NOT present in Griffo's design. The /e beak seems erroneously inspirated from Jenson's design where it's assumed (coming from manuscripts) and harmoniously integrated. But other Renaissance designers like De Libri and De bruschis didn't use it.

Interestingly enough, Monotype's Poliphilus (based on Aldus' typeface for the Hypnerotomachia Poliphilii, which as we all know here differed only in its upper case from that of De Aetna) didn't have either of those characteristics.

Looking at Jensen's original type, I found it very hard to see the beak on the lower-case e, but it is there, and it's also there in (a more pronounced form in) Ludlow's Eusebius, for example.

IBM's Aldine Roman, surprisingly, doesn't have those characteristics either. Part of the credit for not having a prolonged top-end on the lower-case r, of course, goes to the fact that the widths are exactly the same on all Selectric Composer typefaces.

konrad ritter

I applaud your initiative, and I really look forward to seeing the first usable versions. 

However, I must agree with Mr Eliason. The sample you've shown us looks too skinny and a bit washed out. I've seen good scans of De Aetna where the heftiness of typeface really comes through, and it's lovely. An unbelievable balance of thick strokes and thin joints. I don't know how they did it, but I've never seen it in any modern revivals. 

John Savard

konrad ritter said:

I've seen good scans of De Aetna where the heftiness of typeface really comes through, and it's lovely. An unbelievable balance of thick strokes and thin joints. I don't know how they did it, but I've never seen it in any modern revivals. 

Since the good scans I've seen have copyright notices on every single page, I give here an image taken from a lower-resolution scan where what he is referring to seems to be visible:

Here, the lower-case h, m, and n all show the stroke of the right-hand stem coming very nearly to a point where it joins the left-hand stem. One thing this indicates is that the impression is not causing a great amount of ink spread at this point.

I can see why modern revivals fail to preserve this characteristic of the typeface. To modern eyes, it looks as if the character is broken due to an excessively light impression, and so it's assumed the characters are designed to correct for a certain amount of ink spread.

However, while both Monotype's Poliphilus and IBM's Aldine Roman completely avoid making the joints thin, this characteristic is present to a degree in Monotype's Bembo, although it has the characteristics copied from Jenson noted above.

ivan louette

Hi ! Some more steps : addressind some of the evocated drawing issues on lower cases, finalizing serifs and curves transitions, testing more weights (+4 and +7 units) and even a 101% width added to a special version of the +7 units on the last picture. In the future I will test more weight again.

ivan louette

Working on a +14 units weight. The width of this one has been increased at 102%

ivan louette

Better with a little bit more room around /e and /a, narrower /u /b /d /k /p /q stems and removing some weight to capitals.

Craig Eliason

Are the "hairline" diagonals in /v that relatively thick (low contrast) in the original?

John Savard

There is no original for the lower-case v. You get upper-case V but not U, and you get lower-case u but not v. At least in De Aetna itself; Aldus may have used the typeface on books in languages other than Latin, however. Absent an original, though, I'd suggest using lower-case y as a model for the correct thickness.

An image of the characters in the typeface on a page about incunabula is missing those two letters, so apparently there indeed is no original to go by.

ivan louette

@Craig Eliason You are right about the too low contrast and too big thickness of /v, /w and even /y ! I only saw it after giving more weight. I should slightly change /x and /z in the same direction.

@John Savard Of course /y is the model when /v and /w are lacking like in latin texts where /v is replaced by /u. And /z is fairly rare (and often fancy) in latin lower cases.

ivan louette

Working on a short ascender and descender form (just an alternate). Correcting characters width and rythmics and working on better details and weight extrapolation settings in FontForge. From thin to regular.

kfeuerherm

I came upon this page as I have been looking into the work of Frank Blokland and am interested in testing his method and that of his students using Griffo's Polilphili type.

I'm not necessarily concerned with creating a revival of it as such, but I am interested in studying the metrics and so on. I've scoured the net for resources and have found some museum scans as well as a facimile, but none are of particularly high resolution nor are the scans terribly flat.

In light of comments I've seen above, I was wondering whether anyone could point me in the direction of some better scans? even just a few pages?

Many thanks for any help anyone can offer!

Karljürgen

kfeuerherm

For anyone else who might have the same question, Frank Blokland kindly pointed me to the Incunabula Short Title Catalogue, here.