Type Design Revivals
Didot Nouveau
Project Overview
The following is a documentation of my process for a revival typeface based on a source. This was completed as part of the class Type Design Revivals (51–313) at Carnegie Mellon University in Fall 2023 — Instructed by Kelsey Elder.
We were given six sources from the special collections to choose from (each with different typefaces), of which I chose Didot (from Analyse des Équations Déterminées).
Milestone 1
Goals: Determine patterning and grid, define idioms, and gather letter images.
Monday, September 25th
To start on Milestone 1, I took one of the full pages of type (so nothing was too distorted from the zoom lens), and started trying to find out what size of grid would best line up with all of the letters. I decided to go with each unit being half a stroke width, since this pretty much aligned with everything (for the most part), and if anything was off it was likely by half a unit. I then started dividing up words into different character widths, accounting for the space in between them to see if I could see any relationships of how much space was on each side of the letters, and what seemed to make the most sense.
At this point, I also started collecting character images, and I figured it was a good idea to make them all a consistent size (which saved me time later because this was actually a goal of Milestone 2). For now, I only used the extremely zoomed images, and the partially zoomed images, as they had the best quality. However I was still missing some characters.
I then worked on the idioms. I already had a good idea about the character of the type, but it was helpful to put some of it into words.
Milestone 2
Goals: Create consistently scaled character images, define a digital grid system, and create the characters / n / h / o / a / p / e /.
Wednesday, September 27th
To start Milestone 2, I already had many character images, so I used the rest of the more zoomed out images to get any characters I couldn’t find before. I also refined my grid a bit, as I didn’t fully account for any bleed in the height of my characters (even though I had been in the width).
I also defined what this “revival” was going to mean, and decide on some guides for myself:
- I don’t intend on making a pixel perfect recreation, and instead want to make a typeface that has the same spirit as the original. This means that sometimes I might make changes that aren’t necessarily accurate to the source material, but still reflects the nature of the typeface.
- I don’t want to create the typeface to look like it’s printed form. By this, I mean that I want to create it as I believe it would have been if they had our technology at the time in both digital creation and printing. In fact, after seeing very similar (or even the same) metal typefaces, I can see that the metal form is more precise (able to make very thin and straight lines like the serifs), and I want to allow the precision of our current technology to modernize this typeface.
- More technically, I want to make the ascender and cap heights equal. This is partially a personal preference of style, as I think the consistency would suit this typeface well. However, I may come to change my mind on this later, so we’ll have to see.
Monday, October 2nd
Next, I started working on my initial characters. Part of this was me learning the method of drawing in glyphs. I found it helpful to be able to overlap shapes, and limiting myself to the 90 degree angles was helpful in forming consistent curves. I worked first on the characters with repeated shapes (from the n to the h, then the o & h to the p). The most difficult were the a and e, since I didn’t have many similar shapes to work with, and so there were some issues in those forms. (Inconsistent stroke weight, incorrect curves, etc.)
To create the letters, I used a mix of “tracing” the forms, and using the research I did on the stroke widths and spacing. I used the images as a general guide, but always used the grid for the actual points (while keeping the overall shape of the original).
Milestone 3
Goals: Expand, iterate, and refine an initial character set with the following characters:
/ H / O / h / a / m / b / u / g / e / r / f / o / n / t / i / s / v / p /
Wednesday, October 4th
Before expanding my character set, I took the advice from our critique on Monday and refined what I had. This mostly involved correcting some curves (to avoid flattening), majorly fixing the e with new eyes, and fixing the stroke weight on the a.
Monday, October 9th
I then started expanding, first based on the shapes that were easiest based on what I already had (b / i / l / m / r / u).
At this point, I also created an initial proof document setup, although I ended up changing this a bit on Monday according to our demo, and would need to change it more in the future once I had capitals so that I could add in what I was missing.
Wednesday, October 11th
Next, I finished off my final set, adding in the characters I was missing.
I also finished my proof setup, which helped me spot some issues with my type, especially when printed. For example, I decided to make the capital stroke 5 units wide instead of four, as this seemed to fit better at first, especially in something like the “O”. However, when viewed at small scale in context, they seem to stand out a bit too much.
At this point, I had also been thinking about over/undershoots in my curved letterforms. My grid unit is 22, and when I started with my initial character set (o / h / a / p / e / n), I thought that the default 16 units seemed okay (about 3/4 of my unit). However, as I went along, I realized that this was way too much, and was especially noticeable on the baseline (letters like / b / u / in particular).
So I went back through all of my letterforms, and made the over/undershoots 11 units instead. Most letters were easy to change, since I could move just the highest points of the curve without moving anything else (n / m / p), but others required more adjustments (e / a). After this, I had my final milestone 3 proof:
Milestone 4
Goals: Continue drawing and spacing, only adding new glyphs when the set is working well, and drawn and spaced with best practices. Expand the set to have the following characters: / a / b / c / d / e / f / g / h / i / j / k / l / m / n / o / p / q / r / s / t / u / v / w / x / y / z / / A / H / O / E / F / I / L /
Thursday, October 12th
After going back to revisit the source materials, and discuss our progress within our groups, I had a lot of thoughts. There were a lot of smaller details that I was able to notice which I noted down to fix. There were also some larger issues, even across our whole group — most notably with the spaces. None of us had changed the default width yet, and especially when looking back at the source material, we realized it needed to be much wider. It was also interesting to note the different ways we approached certain things, such as the serifs on the s (which is different since it is vertical instead of horizontal), maintaining shapes of terminals (a / r / g / f).
The issue that I was most conflicted about was stroke contrast. I had decided that both the thinnest points of my forms and the serifs would be 1/4 the width of the thick strokes. Both of my other peers made their thin strokes thinner then their serifs, and upon looking back at mine, I thought it was lacking a bit of that contrast. One of the reasons I picked Didot in the first place is because I loved the high contrast. However, I had already made a lot of progress. Another reason I was conflicted is because of how in this typeface, the thin strokes/serifs maintain the same physical size, regardless of font size. Therefore, larger font sizes had higher stroke contrast while smaller sizes had lower contrast.
Although I was advised to continue on with what I had, I just couldn’t give it up. So I experimented with the glyphs I currently had, knowing I had a bit of extra time with fall break, and that I could always go back. I knew it would be a lot of work to make the thicks thicker (and it would change all of my spacing), and I wanted to keep the grid I had. So instead, I made the thins thinner.
I started with my initial proportions of 1 to 4 (or 22 to 88):
At first, I thought that going down from 22 to 16 (about 3/4 of a unit) seemed about right. So then I had a ratio of about 1 to 5 (really 16 to 88). However, It didn’t seem like that much of a difference, especially when printed out small (or anything except display sizes):
So I made another version, this time with serifs half as thin as the original, now with proportions 1 to 8 (or 11 to 88):
I printed out proofs of all three versions to compare. It looked like a large difference on screen. However, there wasn’t actually that much difference when printed. In fact, with smaller sizes, the printer didn’t even have the resolution to make any meaningful difference. It was only noticeable to me when at about 72pt, and most people I asked couldn’t even tell a difference at 200pt.
So, as the week before fall break came to a close, I took a bit of a break to step back, and contemplate my experiments a bit later.
Wednesday, October 18th
After taking some time away, I came back to make my decision. And in the end, I stuck with my original design. Partially, this is because I knew I had spent more time and took more care and attention when creating the curves in my initial version. Also, I realized that even though it may have felt like it could use more contrast at larger sizes, I knew that I cared more about designing this so it looked good at small sizes, and if I made things too thin, it would begin to be harder to read. So I went back to my initial version (2.1) to move forward.
I started by trying to finish up the letters for the goal of Milestone 4, of which I need:
/ c / k / w / x / y / z / A / E / F / L /
I started by finishing up the lowercases, which presented some challenges. Most of these involved new shapes, since they had diagonals, which was only similar to the / v / I had already. I also didn’t have reference for / k /, although I was able to find one from the same book online (although in very low quality), just so I could take some more educated guesses about the form.
For the capitals, it was mostly straightforward, although I had to figure out how I wanted the larger vertical serifs to look in the capitals, and also how to adjust the proportions (since I was making the caps as tall as the ascenders).
Friday, October 20th
Before moving on, I knew I needed to take a second pass at spacing. Nothing was terrible (apart from maybe the f, which just likely needs ligatures/kerning), but it wasn’t great either. So I started from the beginning with o and n, and messed around a lot until I was happy. Then I started spacing characters one by one.
Overall, it’s hard to tell a difference, although it does seem better when in text form. And since I couldn’t help myself, I went ahead and made a few more characters ( B / D / K / P / T / U), and some punctuation (. / , / : / ; / — / — / — / _) so that I could get some more English looking text in my proofs. In fact, I even asked chat GPT to generate some text for me, so that I could get only the characters I wanted, but still have real English sounding sentences and punctuation (Although it seemed to have a bit of trouble at first before some convincing).
Wednesday, October 25th
Before our group critique, I mad a few small adjustments like adding in the / Q / and / R /, and adding in and altering some punctuation (from our demo on Monday).
Milestone 5
Goals: Continue drawing and spacing, only adding new glyphs when the set is working well, and drawn and spaced with best practices. Expand the set to have the following characters: / a / b / c / d / e / f / g / h / i / j / k / l / m / n / o / p / q / r / s / t / u / v / w / x / y / z / / A / B / C / D / E / F / G / H / I / J / K / L / M / N / O / P / Q / R / S / T / U / V / W / X / Y / Z /
Thursday, November 2nd
After small group critiques last Wednesday and this Monday, I had a list of things that I wanted to work on before adding any characters:
- Fix my construction — ̶n̶a̶m̶e̶l̶y̶ ̶t̶u̶r̶n̶i̶n̶g̶ ̶c̶u̶r̶v̶e̶d̶ ̶n̶o̶d̶e̶s̶ ̶i̶n̶ ̶t̶o̶ ̶c̶o̶r̶n̶e̶r̶ ̶n̶o̶d̶e̶s̶ ̶w̶h̶e̶n̶ ̶t̶h̶e̶y̶ ̶t̶u̶r̶n̶ ̶i̶n̶t̶o̶ ̶a̶ ̶s̶t̶r̶a̶i̶g̶h̶t̶ ̶e̶d̶g̶e̶ ̶(̶I̶ ̶t̶h̶o̶u̶g̶h̶t̶ ̶t̶h̶e̶y̶ ̶w̶e̶r̶e̶ ̶s̶u̶p̶p̶o̶s̶e̶d̶ ̶t̶o̶ ̶b̶e̶ ̶c̶u̶r̶v̶e̶d̶ ̶a̶n̶d̶ ̶s̶o̶ ̶I̶ ̶d̶i̶d̶ ̶t̶h̶i̶s̶ ̶e̶v̶e̶r̶y̶w̶h̶e̶r̶e̶)̶,̶ ̶a̶n̶d̶ moving handles off of corners (I couldn’t get shapes with perpendicular lines and didn’t think it was okay to have angled handles on a corner). Note from the future — I was right the first time, I misunderstood what I was doing wrong.
- Fix my thin diagonals (and verticals in uppercase) by making them thicker.
- Fix my curves (they are metrically the same width, but optically light).
- Fix the widths of my capitals (some are narrow because I am making them all taller).
- Experiment with the “flared” (not rectangle) serifs in more of the capital letters to help with weight.
- Try turning auto-hinting off to fix printing on some characters at small scales.
I also wanted to change some things about my proof that wasn’t working, or could be more helpful. Specifically, making the pages with just lowercase, capitalized, and uppercase have smaller text size (I didn’t have any more medium text in my document currently), and adding in alphabetic spacing strings (every character next to every other).
After making the adjustments mentioned, I also started making the remaining uppercase characters (although they still need quite a bit of refining). I also kept versions in the layers panel of the “standard” rectangle serif versions, so I could revisit them at a later date as I figure out what direction to go.
Monday, November 6th
At this point, I still needed to do some work on the capitals. I had some stylistic decisions to make, and I also needed to fix the issues in narrow capitals (especially after making new capitals without adjusting enough for the increased height).
Then, I needed a bit of a break from the capitals, so I went back to punctuation. I didn’t have a reference for the exclamation mark, and although I had a reference for a question mark, I decided to change it quite a bit.
I decided to remain on my break away from capitals, so I started on rough versions of the numerals.
Then I updated my proof so that I could get better feedback, which I think I needed to help with some of my current issues that I might be blind too after looking at it for too long.
Milestone 6
Goals: Keep expanding the set by adding glyphs, including old style numerals, and basic punctuation. Begin considering open type features such as ligatures. Try to have the following characters:
/ a / b / c / d / e / f / g / h / i / j / k / l / m / n / o / p / q / r / s / t / u / v / w / x / y / z /
/ A / B / C / D / E / F / G / H / I / J / K / L / M / N / O / P / Q / R / S / T / U / V / W / X / Y / Z /
/ 1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 0 /
/ . / , / : / ; / ‘ / “ / ! / ? / — / — / — / _ /
Wednesday, November 6th
On Monday, we critiqued our proofs in groups, which I found helpful to spot issues which I hadn’t been able to see (ex: “s” is now too wide, the “g” needs contrast in different places, or the “P” crossbar is too high). However, everything seemed to be going quite well, so I took a bit of time to think about ligatures. I was a bit concerned about the spacing becoming an issue, so I used the advice to follow the pattern of an “n”. I focused on just ligatures beginning with f, which is what I have seen most frequently in early metal type.
For most of them, I elongated the f to merge into the next letter. However, it felt a bit too strange on “ff”, so I tried just changing the height on it and merging their crossbars. Since my normal “f” is so narrow at the top, they all seem a bit wide, so I’m not exactly sure what to do, as I don’t want to put them closer, or the spacing would be too narrow.
Monday, November 13th
On Wednesday during class, we had a bit of fun experimenting with different styles of our letters. We focused on a small subset of our letters for this / o / h / a / p / e / n / to see what a full set may look like. In addition to the regular, I created a display, bold, slab serif, and sans serif.
I had a lot of fun with this, and really wished I had more time to devote to this project (maybe I will in the future after the semester), and could even look into variable fonts.
Final Project Goal
Goals: Identify your final font goal, and if it includes characters beyond the required set. Consider characters such as:
/ 1 / 2 / 3 / 4 / 5 / 6 / 7 / 8 / 9 / 0 / (lining, proportional, etc…)
/ | \ ( ) { } @ # $ % ^ & * + < > ≤≥ ¡ ™ £ ¢ ∞ ¶ •
Language Support: / à / á / â / ã / ā / ă / ȧ / ä / ả / å / ǎ / etc…
Historical forms: / æ / œ / ſ / etc…
Symbols or Illustrative Glyphs
Monday, November 20th
On Wednesday, we got a chance to look back at our source material, and discuss our current progress with our groups. I went on a bit of a hunt for any special characters I could find. Some were more practical (ligatures, prime marks, parentheses, brackets, and braces). Others were more for fun, as I didn’t anticipate being able to add something like greek letters.
I then took some time to think about extended character sets. I first started with quotation marks and brackets, since I felt they were most necessary.
Then to keep with the language support, I added things like upside down question marks, and other dots. I also edited my previous lines to make them similar in style (and weight) to the brackets, and add slashes.
And since my source included a lot of math, I couldn’t help but add some math symbols.
Monday, November 27th
Before leaving for break, we got to do a variable font demo in class. Since I already had the characters / o / h / a / p / e / n / in a regular, bold, display, sans, and slab from the demos on styles earlier, I experimented with these styles. And since I was having too much fun with variable fonts, I finished what I started over break (although it’s still only 6 characters).
I started by editing the / a / so that each section of the character had the same amount of points, and would match when going from serif to slab/sans. It did work, however because of the way the linear interpretation worked, combinations such as both display and bold at the same time didn’t look right. This makes sense as to why you usually design each combination at their full extent, so they can better transition at any value.
So, I got rid of the slab (it was my least favorite, and it’s exclusion meant I only needed to make 4 new versions and not 11), and made the necessary combinations (Bold Display, Sans Display, Bold Sans, & Bold Sans Display).
Over the break, I did a bit more experimentation with symbols. Since these symbols included contrast (as opposed to the math symbols), they were more time consuming, especially as many didn’t reuse shapes from existing characters.
I also started thinking about the end of this semester in general, and in the scope of this class and my goals. With the completion of these symbols, that is pretty much where I initially planned to stop in terms of characters (with only a few exceptions). However, I’ve decided to go a bit further and include language support. With the addition of working on the final poster, and also attempting kerning (on at least the alphabetical characters), I think this will end the project in a good place.
Wednesday, November 29th
On Monday, we made our last visit back to the library to take a look at our sources, and discuss in groups. At this point, it was very easy to become blind to certain issues. This included things such as:
- Tight spacing in numerals
- Strange alignment of many math symbols
- Prime marks being angled (I had used an image of italic ones)
In addition to fixing these issues, I also now had to think about my poster. We would all be using the Risograph, and to make things easier and more cohesive, we could choose 2 of our six colors to use (in addition to black). We chose red and blue (which was my first and second choice respectively), and I went with red.
I didn’t have many ideas to begin with. I started with the name. Admittedly I was helped by Chat GPT to brainstorm some ideas, most of which were bad. However, I liked Didot Nouveau, as neither word looks like it sounds, but they both rhyme. I also knew I wanted a strong grid (as my type itself is so perpendicular). Going off this, I used a thick line on the side, and a thinner one horizontally (they are even in the same ratio as my typeface, 1:4). I then knew I wanted a description of the project/typeface, so I added that as a subheading.
Then, I decided to include text at various sizes (like in the specimen posters we saw at the library), as well as a full character set. All of the type itself is in black, and I used the red as a way to caption everything.
In class last week, we had a demo on language support, and talked about some things like math symbols and arrows.
I started by adding arrows, as I really just think they’re fun (and they weren’t too time consuming.
I knew I wanted to add language support since my original material was in French, and especially if I wanted to include anything from the book on my poster, I would need at least a few accented characters. I decided to start through the “Western European” tab in glyphs (since that’s where French’s required characters are).
After making my dot-less / i / and / j /, I started one accent symbol at a time (or two for accents/graves). I made the symbol, added anchors, and then added all of the characters which used that symbol into the font. After checking that everything looked good, I moved onto the next.
That got me a decent portion of necessary characters done, so I decided to finish out the Western European tab (and I couldn’t resist the challenge of making these new characters fit with the set like any other upper or lowercase). Throughout all of these new characters that I wasn’t as familiar with, I gathered inspiration and “standards” from online.
First there are the combined figures (and while I was at it, an / fj / ligature).
The / IJ / and / ij / were pretty much given to me and didn’t require anything special.
For the rest, I started with both glyphs next to each other. For / æ / and / œ /, the main collision issues I fixed were at the bottom to middle center. I tried to join them smoothly, while also maintaining the ideal stroke width. For / Æ /, I left the E as it was, and slanted the / A / more so that it could take up enough width. For / Œ /, I tried different approaches, but in order to get the right width and make it still recognizable as an / O / and an / E /, I took the serifs off the left side of the / E /, and basically truncated the / O /.
Most of the rest of the characters were newer, but had some or a decent basis of shape in what I have drawn.
For starters, / Ð / Ø / ø / were all modified off of existing characters, just with an added line. Then / Þ / and / þ / were based off of my characters / P / and / p /, just with the bowl shifted or stem height added.
The other characters required me to visit some of my numerals for shapes. For / ð /, I started with a backwards / 6 /, cut off the terminal, and added a slash. For both / ẞ / and / ß /, I used part of my / f /, and a heavily modified / 3 / or / 5 /, based on how I saw them commonly drawn online.
I also did a bit of ideating on my poster, although it wasn’t my main focus. I added in my new characters, simplified the number of text blocks, and tried some alternatives to the solid red bar. However, I still didn’t think any of these were the right direction, and I needed to take a bigger pivot.
Wednesday, December 6th
Before moving on, I got some ideas for directions for my poster, and my typeface. First, I fixed some of my issues in type, which mainly consisted of fixing the placement of anchors in my / i / and / j /, accentuating my uppercase horizontal serifs / E / F / T / Z /, and adding even more weight to the thins on my letters which relied heavily on thin vertical or diagonals / A / K / M / N / U / V / W / X / Y / k / v / w / y / z /.
Although I had asked earlier in the course, and I had already looked into some resources about kerning (and kerning groups) in glyphs, we got a real demo in kerning on Monday. I had been looking forward to this, but I didn’t want to do it before I was ready, as tempting as it was for pairs like / AV /.
I knew I needed to use kerning groups to save my sanity due to all of the accented characters. So after a couple of tries and reduplicating my last file to start over, I got kerning groups to work, and so I started on my way. I didn’t want to go in with reckless abandon, so I used the spacing strings for lowercase only, uppercase only, and mixed case.
To make the daunting > 2.5k combinations sound more reasonable (with the extremely short time frame), I set myself some ground rules.
- Only kern in multiples of 11 (half my unit)
- Make as many kerning pairs as is reasonable (such as the left side of / O / and / C /)
- Focus the majority of time on pairs that really need it (like / AV /) and minimal energy on more minor adjustments
And after forgetting about foreign characters multiple times, I finally kerned everything to a level for my own satisfaction.
It’s a bit hard to tell with the small images here, but it definitely makes a difference in some pairs, and can be seen when toggling between the two, and when looked at up close.
Friday, December 8th
Now that I was mostly happy with my typeface (I had to stop somewhere), I started focusing more on my poster. Currently my poster felt more like a proof, so I had some things to work out.
First, I decided that it wasn’t that necessary to show my type at different sizes (this was useful when we had metal type, since each size had to be cut separately). There was also an issue with size and hierarchy, as nothing was really the focus of the poster.
I tried to experiment with one of the book pages which featured math, and also showing some of the construction of my letters.
However, there were some issues with these iterations. First, I had to confront some of the constraints of the RISO printer. Any “gray” values would basically appear as less dense patches of pigment. This means that extremely light grays don’t work well, and all gray values will pretty much appear very textured. Because of this, I decided against showing the original page on the poster (which also allowed for more room), and not tint the page with my text on it, so you could more clearly see the text.
Also, small text sizes wouldn’t work well, so the page needed to be fairly large if the text was to look decent.
There were also some specific design elements that weren’t working so well (I asked classmates for some feedback). First, the subhead was too small. I had made it exactly one line long, but this made the font small. Two lines was too big, so I added some extra text to make it work.
Then for the construction diagram, the / Ag / made the points too small (for both the RISO and for general hierarchy purposes). So I decided to use a single letter which was only as high as the x-height so that it could be larger (I chose / a /, as I felt the points were the most interesting).
I then separated out the colors for the RISO printer.
I also had to work on my proof a bit. I had slowly been editing it to include more useful things for a more “final” proof rather than a process one. This meant showing off more of the features of my font, rather than pages which could help with the process.
I also had to plan out the space I was using. I knew there was some extra things I wanted to include for my main font, but I also figured out there was enough room to show off my pixel font a bit.
