Recurrent Morphing Radio (RMR) is a generative neural sound production system driven by the most popular genres on digital music platforms. The system extracts and restructures the key sound characteristics found in these styles, generating new sounds that are continuously released. Interspecifics used computer models such as Recurrent Neural Networks (RNN) to develop an automated system that uses platforms such as OpenAI's Jukebox, SampleRNN, and Wavenet, among others, to deconstruct musical structures according to their original properties.

In the live version, Interspecifics interacts with the samples generated by artificial intelligence, modifying and intervening with them using granular and spectral sampling algorithms in a multichannel format with a custom patch built in the Max/MSP software. The sound process is presented with visual live representations as well as metadata from the selected samples.

 

Learn more about Interspecifics