This block aims to reflect on the Afraid to Own a Body composition by Lena Meinhardt and Eva DörrIt, which is a fixed media composition based on a poem by Emily Dickinson from around 1866. Eight channels were used for the performance, making the whole experience very immersive, and IRCAM’s Rave software was used in order to generate their sound material. Rave offers models for realtime deep learning based processing, and the model they trained for this piece learned both of their voice timbres, melting them together and generating a new modeled voice. The artists even mentioned they wanted to create something that would develop its own “consciousness”, as a reference to the artificial intelligence capabilities of the mentioned engine, and also a result of various hours feeding voice material to it, including a full spectrum of textures and also intelligible material. Max MSP was used for much of the processing, where the various sound files would be fed into the models. Regarding the composition, it was intended to contrast something that develops and then falls apart again, which matched with their use of intelligible sentences from the poem, such as the often heard “I am afraid to own a body” sentence compared to completely unintelligible voice material that was generated by the artificial intelligence. The whole sonic experience really suited the original poem, and it was a good representation of artificial intelligence capabilities applied to sound compositions.
