False Return
A heterogeneous vocal corpus treated as an unstable memory whose responses repeatedly disappear and return in altered form, driven by a dry saxophone solo as live exciter.
- 2026
- AI & Machine Listening
- Saxophone and live human–machine performance, stereo
- 4:40
False Return approaches a heterogeneous vocal corpus as an unstable memory whose responses repeatedly disappear and return in altered form.
A dry saxophone solo acts as the live exciter. Somax2 listens to its pitch, articulation and spectral changes, using them to navigate and recombine a corpus assembled from sung arpeggios, humming, sustained female voice, breathing, whisper, breathy vowels and chopped vocal fragments.
Machine responses are withheld, fragmented and re-entered, allowing silence and interruption to function as active formal boundaries. The work was realised as a live human–machine performance; extended inactive intervals were shortened afterward, while smaller gaps and abrupt returns were preserved as part of the form.
Vocal corpus sources
womb_affliction / Katarina Rose (#448085), xkeril (#685679), CVLTIV8R (#836586), SkeetMasterFunk69 (#235576), IENBA (#850703), AlienXXX (#168511) and zShwagg (#185524). All source recordings were edited, reorganised and transformed within the vocal corpus.