Serge Modular Archive Instrument (SMAI): Bridging Skeuomorphic & Machine Learning Enabled Interfaces

The Serge Modular Archive Instrument (SMAI) is a sample-based computer emulation of selected patches on the vintage Serge Modular instrument that is housed at (redacted). Hours of recorded audio created by specified parameter combinations have been analyzed using audio descriptors and machine learning algorithms in the FluCoMa toolkit. Sound is controlled via (1) a machine learning dimensionality reduction plot showing all the recorded samples and/or (2) a skeuomorphic graphical user interface of the patches used to record the sounds. Flexible MIDI and OSC control of the software enables custom modulation and performance of this archive from outside the software. Differing from many software synthesis-based emulations, the SMAI aims to capture and archive the idiosyncrasies of vintage hardware as digital audio samples; compare and contrast skeuomorphic and machine learning enabled modes of exploring vintage sounds; and create a flexible instrument for creatively performing this archive.