Robert Ochshorn, computer scientist & artist, in a talk about art and aesthetics, raw and cooked data, and an audience that may not yet exist.
Schlosspost: Rob, you obtained a degree in computer science and a specialization in fine art. Your background includes media and journalism, electrical engineering, sonic/visual art, and activism. Are you an artist? Are you a scientist? An artist-scientist?
Rob Ochshorn: My computer science education placed a lot of emphasis on efficiency, automation, and confidence – taking a system and making it more robust, replacing an analog process with a digital equivalent or enhancement, proving that your algorithm does what you think it does in a timely manner – but my work doesn’t take the problem as a given. I’m not a scientist because my work is far too subjective. I don’t quantify success. I don’t know if I’m an artist either. I’m not so concerned with fitting myself to a label.
»PDF to cognition« (2014): Word-shape representation of text, trained on the neuroscience literature. By taking PDF journal articles and splitting them into words – without breaking them down into letters – R. Ochshorn saved each word as an image, and then built up a basis set of word-shapes. The distance between words is entirely a function of their visual similarity.
Schlosspost: Why were you interested in art or working with artists and artist groups? What was the collaboration like?
RO: To a certain extent, I’ve been engaged with visual culture for most of my life – photography, amateur video, & computers – but I remember very clearly when I discovered the work of Lissitzky and Rodchenko. I sometimes used to hide in the Fine Arts library at my school to concentrate on math homework, and one day chanced into a section on early Soviet constructivism. It was the first time I thought about art and aesthetics playing a pivotal role in social change. That was the constructivist pretense, at least, I’m not saying it worked, but I never managed to shake the idea of designing for a new sort of person – creating an audience that may not yet exist. I guess that’s an important idea for your magazine as well.
»Chewing« (2013): Robert Ochshorn’s »Chewing« is the result of a computer program written to transform a video from time into space, applied to John Smith’s short film »The Girl Chewing Gum« (1976). Each row of the video represents one minute of time, which is mapped from left-to-right and top-to-bottom as in a book. The entirety of Smith’s film is played from top left to bottom right with the space between the first and last frame filled with shrinking frames at their correct spatial-temporal position.– Text by Anthony Iles
Schlosspost: What is data? As a resource for your art?
RO: Data is surrogate stimulus. Numbers, carefully chosen to be used in lieu of sight, sound, smell, touch, and taste. Not only to simulate our observations, but to measure what we can not. I make interfaces between the computational and social/political processing of data. A well-designed interface is much like an instrument: it augments our capabilities in some way and allows for expression. Perhaps the design of these interfaces is itself expressive, but I don’t think of my work as primarily an artist’s practice. In my video interface, InterLace, authorship is a process of creating interpretive meta-data (Montage Interdit, VideoVortex9), so that not only the filmmakers but also their viewers can create meaningful continuities.
»God« (study): R. Ochshorn developed the software for Eyal Sivan’s web documentary, »Montage Interdit«, from his research on compression at the Jan Van Eyck Academie. As a data-based film, continuity is not a linear narrative, but is rather achieved at »runtime« by the viewer, who can navigate by tag or by source, resorting to make new continuities and montage. The act of authorship then is in the creation of metadata rather than narrative. An early study (pictured) allowed exploratory visual navigation through 15 Godard films.
Schlosspost: What is your focus in working with data?
RO: Language about data-work often assumes a cooking metaphor. Data begins »raw« and needs to be processed before consumption. I made a pair of videos last year, Chewing and Digesting, trying to represent the transformation of time and image into static data (Chewing) and in the latter how we can make meaning out of the accumulation of our own data-traces (Digesting). We’re constantly leaving data behind – maybe how we focus our attention, »reading,« can be twisted into a form of writing (see: Hyperopia Thing).
Very few people see the whole data cooking process. Sometimes the hidden stages reflect social organization, for example a director cutting (»raw«) footage into a narrative for an audience, while other processing stages are opaque owing to technical structures, such as the encoding of a video into a compressed file format suitable for networked distribution. The human and algorithmic processes both concern themselves with determining what is important and what can be discarded; my focus is in exposing and unifying the intuitive and algorithmic modes of processing.
The App and the Territory (2014): With French musician and composer Nicolas Mondon. – The App and the Territory is a prototype representation of a particular musical space. The sounds of this space were recorded from a re-tuned piano and guitar, each specially prepared to produce a bell-like timbre, and variously muted. All of the samples taken together form a sonic territory, from which a possible map is presented for exploration. The map is a two-dimensional reduction of multi-parameter analysis of each sample’s harmonics, such the location of sounds on this map are based on internal similarity rather than stable or external criteria.
Schlosspost: How can machines »compress« images in a way that helps us understand salience?
RO: My Sublimation project is one experiment in a visible, lossy compression. But also how can interfaces allow us an awareness and instrumentation of the powerful statistical techniques like principal component analysis that already underlie machine data analysis/compression?
Schlosspost: What’s next for your work?
RO: While perception is instantaneous, data persists. It is stored on and trusted as memory. Our memories, individual and collective, are externalized as data. I am interested in the dreams of this data. Dreams in terms of (unrealized) fantasy and possibility, but also dreams as subconscious associative access.
Screenshot from Screen Dreams (2014). Hybrid relational database and video timeline explorer, with material scraped from ZKM’s online archives. Unlike most video codecs, which aspire to replay the video sequentially with fidelity to capture, here an internal representation of the entire file can be viewed, manipulated, collapsed, and cross-referenced.
Schlosspost: Dreaming, like all things here in the twenty-first century, is but information processing. Our memory is shuffled as we dream and whether that is a cause or an effect, a means or an end, random or significant – all of these distinctions blur together in this age of correlation.
Dreaming is the time when our consciousness is more virtual than physical and as a consequence many things are possible: possibility is the reality of the virtual. We used to bring our memories into the physical world (paper, celluloid, &c) but now we just transfer them from a lucid virtual space of dreamy possibility to a dull digital one of error-correcting databases.
Dreaming is a wish, fulfilled or not, but never constrained by compromise or convention. How resigned/designed our memories must find their digital overlords! Beware of anyone claiming an artificial intelligence by jumbling our memories into their models, for they may be trying to steal our dreams.
I dream of a not-so-distant future where virtual space permits the most lucid of dreams, singular and shared, that push their possibilities into our waking, physical reality.
Rendering, Pdf2Cognition (2014). Words are digitally represented from an image-form basis (here trained on academic neuroscience papers) instead of character codes. The PCA basis forms a continuous space of text-image.
Interview by Clara Herrmann
Parts of the interview were first published in the Berlin magazine REVUE Magazine for the Next Society No. 14 Rohstoffe (resources).