A DSP-based synthesizer that takes sensor input from a seismometer to generate a set of sine waves to drive transducers.
Sound modules for the walking sticks used by visitors of the “Speed of Light” project by NVA, Glasgow. Barometric and acceleration sensors control a small DSP that runs a composition by Chris Weaver of Resonance FM.
Sound track for a ‘line-flicker’ video, generated by feeding the line patterns directly into an iFFT.
Contracting work for twosuns, a Berlin multi-media company. Twister is part of the sound subsystem of enclued, an interactive multi-media development environment. Initially a multi-channel soundcard with embedded spatialization functionality, which was later moved to the host PC.
Computer aided score generation of Balinesian Gender Wayang music, based on multi-channel envelope recordings. The recording system required a custom acquisition system as multi-channel audio interfaces were not available in the early 1990s. It was implemented by Thomas Ruoff and Pierre Dutilleux at FH and ZKM, Karlsruhe. I was involved in the later stages that dealt with the analysis of the accumulated data. Most of this was done in Matlab and its signal processing toolbox. Finally, a custom notation system was devised to take the specific playing styles into account.
Traguisma – the TRAnsient GUItar State MAchine. Or, if you prefer the official title: ‘Algorithms for Combined Sound Analysis and Synthesis and their Implementation in a Multi-Processor Environment’.
This was to become a ‘Diplom-Informatiker’, my graduation project. In short it was about improving guitar synthesizer, with a focus on the attack phases of a pluck, in particular re-plucking a string that’s already vibrating.
The intended target platform was GWIZ’ Infinity Box, but as this project was cancelled, I implemented it on CNMAT’s multi-DSP box Reson8.
Here’s the PDF
Early predecessor of the gluiph. A small 56k module to fit inside the SensorLab.
A four-channel audio converter for the NeXT cube and the ispw (i.e. the IRCAM signal processing workstation).
This started as a hack with a couple of evaluation boards and some blobs of glue logic. With help from Atau Tanaka, and the original design by CCRMA.The second version though had its own board and features analog inputs as well as AES/EBU and S/PDIF. It was designed around an FPGA (by Xilinx) to allow for easy modification in the event that other customers might want to run it off a different machine. These days, however, with tons of multi-channel soundcards available it’s hard to thrill people with quadro…