Standard di Fatto
Standard di Fatto was created in response to the Call for Cover of Progetto Grafico Journal no. 43, issued by AIAP, the Italian Association of Visual Communication Design, under the title “Back Then Is Back Now: Design History and Education in the Age of Artificial Intelligence.” The brief asked designers to consider what the discipline’s craft traditions and material processes might offer a field now organized, in large part, around generative systems whose internal operations stay opaque even to the people who rely on them daily. Submissions had to take the form of a modular pattern with no single focal point, able to hold its own once a masthead and other editorial elements were placed anywhere across it.
The work takes its subject from Annex K of ITU-T T.81, the 1992 standard that defines JPEG compression. Annex K lists two matrices of 64 integers, one for luminance and one for chrominance, that set how much of each spatial frequency in an 8×8 pixel block gets discarded before the human eye is expected to notice. These values were not calculated. They were measured, in a series of psychovisual experiments in which observers were asked to report the point at which each frequency became visible against a uniform field, at a fixed viewing distance. The standard is explicit that the two tables are offered only as examples. Their adoption as the default setting in libjpeg, the reference library behind most later implementations, made them the de facto standard anyway, cited unchanged in software and research for over three decades.
Standard di Fatto sets these 64 values as a typographic grid. Each digit is composed in Bell Centennial, the typeface Matthew Carter drew for AT&T’s telephone directories in the late 1970s. Printed at six points and at speed on newsprint, its letterforms carry ink traps, notches cut into the counters that anticipate the spread of ink under a high-speed press. The type only reaches its intended shape after the printing process has degraded it. Over each digit sits a halftone screen whose spacing is set directly by the value it renders. Where the table preserves information the screen is fine and the digit holds. Where it discards, the screen widens until the number breaks apart into scattered dots. Three technologies from three different periods meet here on the same premise: the halftone screen, developed in the 1880s; Bell Centennial, drawn in 1975; and JPEG’s default quantization matrix, fixed in 1992. Each decides in advance, before any specific image exists, what will not survive its own reproduction.
The composition is mirrored across both axes at the page’s center. The reading order a conventional 8×8 grid imposes, low frequency to high, upper left to lower right, is replaced by a field with no beginning or end, built instead around two independent axes that meet at a single point. It is there, at the center, that the most aggressively quantized values converge.
None of this happened in secret. The numbers are published, documented, and available to anyone willing to read a technical standard. What Standard di Fatto proposes is only that almost no one does this, and that a value written into an example table, carried into general use by a piece of open-source software, and never revisited since, now resembles most design decisions made inside automated systems more closely than any single deliberate act of authorship does. The call’s question about where authorship resides once a process is automated may not need a new answer for the age of artificial intelligence. It may already have one, running quietly underneath most images anyone has ever looked at.
Note: quantization values and their psychovisual basis as specified in ITU-T Recommendation T.81 (1992), Annex K.