Obviously our own work is filled with fine examples of Advanced Realism. But the examples themselves can become hard to describe with the level of detail that I’d like to use because they tend to spiral into technical complexity. An example that’s about the right size and density is the top half of the “Requiem” column of our installation Breath which is a drawing generated automatically from Mozart’s music and one that I happen to think turned out quite well.

Scholars have indicated that the last music Mozart is known to have committed to paper is the bottom-left hand corner of the second page of the Lacrimosa of his Requiem. Everything else above and for the next 8 pages is by another hand. This is the seed for the formal idea: the diagram presented in Breath is generated by automatically to identifying (distorted) duplicates of Mozart’s music in the subsequent pages; arcs connect runs of material that are similar, given idea of similarity; arcs are drawn in a way that reflects how sure the algorithm is that the passages contain duplicate material; all the music is drawn using glyphs taken from a scan of Mozart’s manuscript distorted to make the new notes; the underlying layout grid, which ought to just be the staff-lines of the manuscript paper Mozart used is itself recalculated from the positions of Mozart’s note-heads; and they too are drawn and accessed in a way that indicates their ambiguity (the further you get away from Mozart’s hand, the more the angular uncertainty effects the grid placement)

Sometimes such “explanations” are met with suspicion. Either I am trying to secure status for the resulting artwork by displaying the quality of the engineering; or I am trying to secure the status of the image by claiming that it is a unique and pure expression of a problem that flows directly from Mozart himself. Both of these ideas I find repulsive, and both stem from a confusion of what we are doing and why.

The diagram is “pure” in the sense that no other lines are drawn, the form is executed and drawn without intervention. But musically and computationally the diagram is absurd — any computational idea of musical similarity is thoroughly ungrounded, even the one used here. Five other diagrams were created that afternoon using different “ideas” of similarity. And we chose the best one (and entertained the notion of combining all five). The point (the motivation, the force, the success, …) of this image isn’t to reveal some sacred and hidden musicological structure, unseen by scholars to date, by the use of sophisticated computer algorithms; nor is it to make art that is necessarily better through the use of better algorithms. Although the approach is technically rigorous — perhaps even state-of-the-art — it isn’t technically correct. It’s a trap to entertain the very existence of a technically ultimate solution to this non-problem. But as an image it gets its force from its form — the dynamics of duplication (of the music), the asymmetries of precedent (the gravity of the gesture on the page), the strange familiarity of the copy (of the handwriting). There’s some texture left by the process that is larger than the specifics of the approach or the confines of the subject matter.

And I know of no other, more direct routes, to create images that possess this trace. (Not closing my eyes and imagining pictures, not pushing paint around a canvas, not …)