Below the logocentrism of metadata: Kurenniemi's case

As distinct from cultural analysis, sonic analytics in its strictest sense refers to sound analysis performed by technology itself. Technological media provide modes of signal reading that do not immediately aim at understanding. The archaeological method adapts to this ascetic approach to signals, resisting the temptations of premature narrative contextualization. A shift of emphasis takes place from cultural interpretation of audio content towards its analytics or even spectrogrammatics. Sonic signal analysis and data processing is at work for commercial and monitoring uses already – even if automated music analysis mostly still results in generating metadata. With a focus on audio-as-signals, non-logocentristic search engines emerge that by digital sampling randomize what has to this point been known culturally as music.

From 1972 to 1974, the Finnish electronic engineer and research artist Erkki Kurenniemi recorded private entries on audio cassettes, leaving a heap of such records to the Central Art Archives of the National Gallery in Helsinki. How to cope with such a sonic bequest? Algorithms that sort according to spectra similarity can correlate phonetic signals according to loudness and dynamics in decibel in order to find out, for example, the moments when Kurenniemi's dictation switches to singing.2 Significant speech segments can be derived below the philological fixation on semantics. In additionAs well, the actual location where the articulation took place might be derived by signal-to-noise ratio analysis separating the speaker's voice from background noise.

Kurenniemi developed the electronic music synthesizer DIMI-A (Digital Music Instrument, Associative Memory, 1969), which was equipped with the remarkable option to retrieve digitized audio signals according to ‘content’ rather than by first addressing the storage cells – sonic hashing. The sonic archive requires the development of specific search tools that differ from those used for text search in traditional archives. Sonic memory does not consist of single symbols or characters but of micro-tonal objects that lead to an associative memory rather than order from metadata.

Philological hermeneutics knows how to make sense of texts – but how to make sonicistic sense of signals that are addressed to ears? It is a different kind of memory that speaks to the ear when an individual becomes archive in electro-acoustic devices.