Pomegranate Fonts

Carolyn's Process

Process 1

The model for the Carolyn Puzzovio’s ‘classic’ Lagoon typeface was a book printed in 1810: ‘Liakatar vark ew vkayabanut'win srboc'’ – (Complete lives of saints) by M. Awgerian. This was printed at the San Lazzaro Mechitarist monestary in the Venetian Lagoon:.


The typeface was cut around 1805 by the monks and used for the renowned Zohrapian Bible.

Process 1.2

Close-up photographs of many of the book’s pages were made using a Sony SLR digital camera.


A database of the best images of individual characters was made and these were printed out at a cap height of 50mm. These were studied, compared and pencil tracings made by hand at that size, attempting to interpret and remove imperfections caused by ‘ink squash’ and damage to the soft metal of the type: ie. to recreate the original intentions of the monks as far as possible.

Process 1.3

Drawings made from the tracings using black ink were the first stage in correcting the shapes and making the characters harmonise with each other.


This picture shows the lower case character: ‘djeh’

Process 1.4

The drawings were scanned and placed into the background layer of Fontlab program (on a Macintosh computer). Their outlines were drawn and further refinements made.


These Armenian characters then became the basis for a previously non-existent Latin version, which was drawn on-screen. Next, the special characters, mathematical symbols, punctuation etc. were created, working largely digitally with occasional test drawings.

Process 1.5

Then came the lengthy process of kerning and fitting the characters into words to test legibility – which then creates problems due to the many long descenders (and some awkward ascenders) in the original, clashing and creating both aesthetic problems and legibility issues when the type is set into words. 


The first solution was to create thirteen alternative forms for some of the problematic characters which would be automatically substituted when certain keyboard combinations arose using the relatively newly-available OpenType ‘glyph substitution’ features. The second was to create a number of glyphs (95 in all) which were ‘ligatures’ of the two clashing characters – these problematic combinations were identified by Ghabuzyan’s tests and through his knowledge of the Armenian language.


The help provided by a type consultant in the UK – Jeremy Tankard – was invaluable at this stage of the project.