The typeface Ambroise is a contemporary interpretation of various typefaces belonging to Didot’s late style, conceived circa 1830, including the original forms of g, y, &; and to a lesser extent, k. These characters are found in Vibert’s typefaces. Vibert was one of the appointed punchcutters of the Didot dynasty. Punch cutting at the time was a hard and long task. In the case of the various Didot styles of typefaces we find in the company’s catalogues of typefaces that it was more the job of a team over a long period rather than the unique work of one designer.

In the past, (before the 20th century) you generally find only one style of typeface, Garamond at the time of Claude Garamond, Didot at the time of Didot, according of the country were you live. There is also a certain amount of influence between European master-punchcutters, an example being that of Bodoni, who cut and printed from fonts derived from those of Fournier. In the past, typeface trends remained in vogue for long periods of time. For instance, the Didot family, who began working in the middle of the 17th century, continued until the 20th century. Several punchcutters, such as Pierre-Louis Vafflard, Léger, Vibert, Molé and Didots themselves such as François-Ambroise or Firmin, worked on Didot typefaces with different styles. The Didot cut by Ambroise Didot, and used in the ‘Virgile’ published by Pierre Didot in 1798, is generally described by the specialists as the most accomplished. After some years of interest in Didots, it seems that it was Vibert who cut the some unusual lowercases to improve the rationality of the forms. Theses unusual forms endorse the general effect of verticality which is the prime characteristic of the Didot.

Extract from the Specimens des divers caractères, vignettes typographiques de la fonderie de Laurent de Berny, 1828.

It is the weighty Black, which was the basis for the conception of the family of display Didots. In the second half of the 19th century, it was normal to find fat Didots in several widths in the catalogues of French type foundries, mostly alphabets of capitals only. The narrow versions were widely used for heavy titlings in theatre posters. These same typefaces continued to be offered by French foundries such as Deberny & Peignot (in Spécimen général des fonderies Deberny & Peignot, Paris, 1955) until the demise of the last type foundries in France at the end of the 1960s.

Pagefrom Deberny & Peignot typefaces catalogs (1910).

Typeface revivals

It is know that Didot faces are highly contrasted, generally more than their italian counterparts designed by Giambattista Bodoni. Original Bodoni counterforms are more round compared to the much later interpretations such as the Morris Fuller Benton (1910) or Henrik Jost version for Bauer (1927). In fact, it is the Didots typefaces which are truly vertical in their counterforms. The general flavour of the Bodonis is more soft and curvy, whereas Didots are more strict in their design.

Compared to Bodoni, there are fewer Didot revivals. The Didots were used in France until the 1970s because of their availability in Deberny & Peignot catalogues. Today, Didot is not so much used, but the name Didot remains a reference in France because of the long duration of their dynasty. For example, there are many places and streets in Paris which are called Didot but only one very tiny street is named after Garamond! Didot is also the name used by the older generation of French typographers when they refer to didones (or Modern) typefaces. And there is the Francis Thibaudeau classification (La lettre d’imprimerie, Paris, 1921) which uses ‘Le Romain Didot’ as a category.

Today, there are two major revivals of Didot typefaces. Adrian Frutiger designed a Linotype Didot in 1991 based on the work of Firmin Didot (1764-1836). Frutiger used Voltaire’s book, ‘La Henriade,’ published by Firmin Didot in 1819, as the basis for his design. Jonathan Hoefler designed HTF Didot originally as a titling family for ‘Harper’s Bazaar’ in 1992, but quickly the magazine asked for more sizes and weights, and for versions more adapted to text composition. HTF Didot is now part of the Hoefler typefoundry collection. Hoefler had two sources of inspiration: one seems to be Molé le Jeune (1819); and the other Pierre Didot and Vibert. Both of these revivals, HTF and Hoefler, take the most common model of Didot.
I try to focus on uncommon sources as a complement to these previous revivals rather than a direct competitor. Ambroise is a titling family which exist only in roman form but which includes many different weights and widths as used in late 19th century French publishing and early advertisements.

Jonathan Hoefler designed HTF Didot originally as a titling family for ‘Harper’s Bazaar’ in 1992. Adrian Frutiger designed a Linotype Didot in 1991 based on the work of Firmin Didot (1764-1836) and used Voltaire’s book, ‘La Henriade,’ published by Firmin Didot in 1819, as the basis for his design.


Ambroise typeface attempts to reproduce more of what we see printed on paper in the 19th century; a more accurate representation of Didot punches. So, the unbracketed serifs are not truly square straight-line forms but use tiny transitional curves instead. The result on the page appear more softer and less straight, particularly in larger sizes. The high contrast of the Didot is indeed a strong feature, so I decided to focus only on a display version. But who knows? I might decide some day to finally design a text version complete with italic.

Uncommon letterforms

Another reason for a display version only is the adoption of unusual glyphs such as the g, k, and, perhaps y. After a few months of work on this project, I fully understood that their unique forms came from the idea of affirmation of sustained verticality within the style. The usual g presents a strong emphasis on the horizontal because of the position of the median stroke, following from reference to equivalent pen-stroke. In the case of the new g, it follows the then current trend (at the time of Didot) to move from broad-edge pen tradition to fine-pointed flexible pen. The new g and the curvy k display a perfect integration with the verticality of the others glyphs. As it is always difficult to design perfect, narrow Didots (because of the disturbance of diagonals and horizontals) these characteristics helped me to integrate these particular glyphs with the rest of the character-set.

Unusual Didot glyphs in Ambroise.

Vertical construction versus humanistic construction

This was a big step for me; representing the last step, from student type designer to becoming a professional type designer. I started to design my first typeface in 1989! I learnt type design via calligraphy as a teaching method. This is the most efficient method, and one which I now use to teach type design to my students. It is an efficient way to ensure that beginners understand the direct relationship between writing and the construction of letters, the evolution of the type forms, the relationship of counterforms to forms and so on. This very important part of the learning process is the key. Particularly because most of graphic design teaching focuses on the geometrical construction of letters, no help at all for students who might wish to become type designers. Geometrical construction of letters is more the art of reproduction than creation itself. Type design is an intellectual process more than a simple manual process.

In the same way, knowing how to design good and creative type forms has little to do with knowing what is a Bézier curve. This is exactly what happened with the appearance of the Didots, which began about a century before with the Jaugeon Commission (1693–1704). This Commission was composed of ‘Académiciens’, who called themselves experts — but who had, in fact, no direct understanding of the arts of calligraphy, punchcutting or type design. They presented to Louis XIV a grid-based method of construction for a new typeface, a romain du roi, based on a geometric construction which followed the same principle as our bitmap fonts of today. Didots are the eventual resolutions of the earlier attempt by Grandjean and his colleagues. The Didot style is the expression and fulfilment of mathematics and logic described by René Ponot in his article ‘Du Romain du Roi au Didot absolu’ (L’affiche en révolution. (1) written by Alan Marshall and Thierry Gouttenègre, Musée de la Révolution Française, Vizille, 1998).

The Jaugeon Commission (1693–1704) presented to Louis XIV a grid-based method of construction for a new typeface, a romain du roi, based on a geometric construction which followed the same principle as our bitmap fonts of today.

Perfectly demonstrated by Gerrit Noordzij, the concept of translation versus expansion in ‘Dossier A – Z 73,’ (2) under the direction of Fernard Baudin and John Dreyfus, Association Typographique Internationale, Copenhagen) describes very well the differences in the interaction of the forms with their counterforms. In France, without any connection to The Hague school where Noordzij has been teaching, it’s Bernard Arin at the Scriptorium de Toulouse, established 1968, returned to the humanistic approach to type design. Translation is what the broad-edge pen produces on paper; expansion is what the flexible fine-pointed pen produces on paper. When you design a typeface via a humanistic approach, the counterforms are round and the outside forms more square. There is also a stronger emphasis on horizontals. During the design process, you balance all your type forms by exploiting the horizontals and the open counters. With this in mind, it is quite revealing to design a typeface based upon a geometric, vertical concept. In this case, the outside forms are very round and the counters narrow and square. The general balance comes from the overall vertical rhythm.

Bernard Arin detailed explanation of the design of a typical Garalde.

Gerrit Noordzij describes very well the differences in the interaction of the forms with their counterforms.

Interpolation limits of Ambroise and Anisette

It is obvious today that type designers do not draw all the variations of a family, because interpolation features are such a great help in our design work. But it is important to be aware of the limitations of this wonderful technology. In 1995, when I designed Anisette, an all caps typeface family, one of my objectives was to go up to the limits of the interpolation. My two master-sets of character elements were very far apart in term of weight; light to black. When I interpolated them, the resulting medium weight was not very good. In the end, I had to redraw it completely before I could use it as the basis for the interpolation of the other intermediate weights between regular and bold. But there is an explanation for this phenomena. Thin sans have no contrast at all between horizontal and vertical strokes. The small modifications or minor mistakes are immediately noticed by our eyes. The job is not only conducted by vision, but also by geometry and rules. Black sans are generally designed visually, and many optical adjustment need to be made. The main adjustment resolves the strong differences between horizontals and verticals. The latter are normally bolder than the horizontals. In medium weights, the thin line with little contrast between horizontals and verticals is closer to what you need to follow. When you interpolate, you can establish the right weight, but the strong optical adjustments from the black to the medium weight remain too prominent in the interpolated medium. You need to increase the weight of the horizontals. The overall result will tend to look bolder after modifications, so, you need, again, to use interpolation with the thin weight to obtain a perfect medium.

Anisette interpolation problems.

Two main problems appeared during the interpolation process of Ambroise. The way the serifs are designed with Bézier curves puts a point in the middle of each end of the serif. Because the serifs are very thin, there is a limitation put upon the PostScript grid. When I interpolated them for intermediate weights, and intermediate widths, the middle point appears too high or too low because of the grid and the low number of units. So, I carefully adjusted all the serifs in most of the new series. The second problem results from the interpolation of the glyphs a, k, K, R. The endings of these glyphs are similar in construction. In the bolder and wider versionss, the design is more open and horizontal, in the lighter, narrower weights, it appear more vertical in form. The interpolation process provides strange results; for instance, the forms appeared lighter in the middle of the stroke for no apparent reason. I tried different point placements, different software, with no improvement in the result. So, again, I redesigned all theses forms in the whole interpolated series.

Ambroise interpolation problems.

Software procedure

The punchcutter always improved his method by his creation of new tools, better mixtures of metal, etc. The contemporary type designer is now in front of the computer technology and associated software. All software has its limitations as demonstrated in the previous paragraphs. Despite the globalisation of our activity, none of us uses the same design procedure but this is a good thing because it encourages individuality in design outcomes. In my case, after learning to design letters on tracing paper, I first used the German software Ikarus in the early 90s. Despite not being updated by Macromedia since 1996, I like the American program Fontographer to draw my forms and to create the spacing. The Dutch Robofog is a great help for the major proceedures of interpolation and several other programs fulfil very useful functions via the Python scripting interface. Lastly, the Russian Fontlab is a very good piece of software for font production, cross-platform naming, and so on.

Theses days, I draw most of the time without paper, working directly on the computer. I don’t print each glyph, one at a time, during the design process, because screen interpretation is generally enough. I prefer to wait until the end of the day to printout several letters at different sizes via a layout program such as InDesign. The PostScript hinting was more difficult than usual because of the high contrast and the vertical stress of Ambroise. The hinting worked correctly for low-resolution screen size but at high-resolution medium size, all the thin parts disappeared, not because they were too thin in their design but because of the hinting process, This tends to ignore the overlapping of round letters relative to base-line and x-height horizontals. In the case of Ambroise, the effects included cutting of the thin parts: the hinting program could not recognise the difference between the outside outline and the inside outlines.

Size limitation

Unlike most type families, the Didots do not work equally well at all sizes without problems. For each group of sizes, a special design with new contrast needs to be designed. But, it is not always the case because the printing conditions, variety of paper and colour can influence the printed result. At the end, as the Multiple Master system from Adobe (discontinued in 2000) demonstrated to us:
it can be a nightmare for the user. I prefer to focus on a larger palette of widths that are really useful to the users than subtle modifications required to adapt, successfully, size by size. So, I designed only one master for all sizes generally larger than 14pt and available in many weights and many widths.

Didot ornaments

Ornaments, or Vignettes as we call them in France, are the fun part of typeface design. They add value to the typeface because their presence improves the overall appearance of a page. Ornaments were an integral part of the job of the punchcutter in the past. You could not, so easily, mix Fournier ornaments with Garamond, or Bodoni rules with Bembo. Each typeface required its own set of ornaments to perform well on the page. Without them, the layout lacks humanity. For Ambroise, I played again with interpolation. Generally, only one set of ornaments are designed for a type family. This time, I designed fewer ornaments but ensured each was capable of multiple variations. Each series features the same set of ornaments, and none of them are similar, they always follow the weight and width of the typeface reference. The result is a wonderful tool for the user, as the variations help to create different rhythms and colours to the layout.



I like to consider Ambroise as my first type revival. Three years ago, the project started with just a few black letters designed quickly on screen and stored somewhere on my Mac until summer of 2000 when I decided to start the project in earnest. By September, I had designed the Thin and Black completely, showed them to some close friends during Leipzig ATypI conference. A few months later, the project was finished. In the early days, it was not much fun to work on. But, day after day, I took more pleasure in designing it. I think I have now discovered a new way of designing typefaces, one which had never previously occurred to me. I really hope to discover other unexplored areas in the future.

Copyright 2001. Jean François Porchez. Thanks to David Jury. Published originally in Typographic 59, the Journal of the International society of typographic designers.


Didot family (in French).
HTF Didot
Linotype Didot
Didot Elder


1. René Ponot in his article Du Romain du Roi au Didot absolu (L’affiche en révolution. Written by Alan Marshall and Thierry Gouttenègre, Musée de la Révolution Française, Vizille, 1998).
2. Dossier A-Z 73, under the direction of Fernard Baudin and John Dreyfus, Association Typographique Internationale, Copenhagen).
3. La réforme de la typographie Royale sous Louis XIV, Le Grandjean by André Jammes, Librairie Paul Jammes, 1961.
4. Les Didots, trois siècles de typographie et de bibliophilie 1698–1998, catalogue by André Jammes, Paris 1998.
5. Conoscere Bodoni under the direction of Stefano Ajani and Luigi Cesare Maletto, Torino, 1990.