During the ALMAT residency Jonathan is working on several aspects of live coding practice, addressing technical, aesthetical and poetical questions. For example, with the aim of solving a specific limitation that derives from the ephemeral quality of live coding - namely that many processes get lost in the flow of writing and rewriting - he developed a SuperCollider library that allows to collect and catalogue synths played during live performances. These synthesis processes are then available for being reused and modified at a later point, filling the gap between live coding and (out of time) composition.
Live coding and story telling
Jon is also interested in incorporating storytelling into live coding performances. He wants to experiment what musical performative possibilities there are for story telling. "I really like what story telling is, as a performative practice. It's a bit like weaving. Usually there are anecdotes that are sort of woven together, and it's very improvisational ".
Voice and language
Working on this concept, he came to the idea of exploring voice and language, in particular analysis of the voice. After surveying different open source speech recognition packages, he wrote a ⬀python frontend / supercollider backend - based on ⬀Sphinx - for getting some of the speech analysis data into the computer. Feeding the phonemes from the recognition algorithm into SuperCollider he can use them to transform parameters of the synthesis process, using his voice to modify sounds in real-time.
Being confronted with the limitations of tools like Sphinx, Jon has been trying to find ways that are a little less discrete and a little bit more open to deal with his voice. Trying to create a more complex classifier or feature detector, he wants to build a tool capable of capturing specific inflections and peculiar vocal utterances, pulling speech apart into its phonetic, sub-phonetic or non-phonetic qualities.