Lecture Silvia Rosani (IT)
Voices mediation through neural networks
The lecture focuses on the application of neural networks to Silvia Rosani’s most recent works with voices. A web-application assesses the affinity between vocal utterances and audio feedback sounds as well as controls four fans according to the emotion detected in the sounds of a violinist’s voice and instrument. A neural network is trained gradually in a live performance to articulate a structure which moves from heterogenous sonic material to homogenous material in terms of timbre through the generation of the processing parameters and the selection of recordings extracted from traditional folk songs from Northeast Italy and Hungary.
