Word Processor, 2021-ongoing software instrument

Programming and Build: Sean Dockray, with Reduct Concept and Development: Sean Dockray, James Parker, Joel Stern

instrument (n.), from the Latin instruere: to “arrange, prepare, set in order; inform, teach”

The Word Processor is an ‘instrument / compositional / processing tool’ that appropriates methods from speech-to-text and Automatic Speech Recognition as parameters for text and audio composition and experimentation. It was originally built for Unnatural Language Processing, an online performance program at Unsound Festival 2021, in collaboration with Reduct, as a tool for artists to work, think, and compose with and against ASR. The public version of the Word Processor (which works best with Firefox) is preloaded with a range of audio-texts, largely drawn from technologists and corporations talking about ASR. If you would like to add your own, email us.

A short film introducing the Word Processor along with some of the history and politics of speech recognition is available here. This is also where you’ll find early experiments with the instrument by a range of artists and researchers.

An unlisted playlist of Word Processor experiments lives here

Thao Phan’s performance lecture Listening for Misrecognition, which extensively deploys the Word Processor is available here

In 2023, we have begun to develop the next iteration of Word Processor (WP2) for live and offline performance. We plan to share this too. For now, WP2 will be launched live at Unsound 2023 in Krakow.