Generative Field Recording is a sound experiment project utilizing machine learning technology that allows manipulating field recording sound more flexibility. 
With the system, people can generate field recording sound which is not existing in the world using recorded sound.
Generative Field Recording was created utilizing machine learning technology, specifically T-sne which is a classification algorithm. The system chopped recorded sound into small fragments, and T-sne recognizes the characteristic and similarity of these fragments. Recorded sound divided into short sound fragment and plot into a two-dimensional field so that people can manipulate fragments of these sounds flexibility like a musical instrument. 

Back to Top