Audio Metaphor

Generative Soundscape Composition

natural language processing, audio segmentation, AI composition

  • Home
  • System
  • Implementation
  • Audio Metaphor generates unique soundscape compositions. It processes natural language texts retrieved from current Twitter posts to search for audio files in online databases that represent the text.

    Audio files are cut up using a machine learning algorithm to separate parts of the audio file with different qualities. Cut up segments of audio are then used by an artificial intelligence soundscape composition engine to create unique audio representations of the natural language texts.

    System flow


    Audio Metaphor was designed to continually create unique soundscape compositions that represent natural language requests. The system generates soundscape compositions by first transforming a natural language request into segmented audio file recommendations. Thereafter, segmented audio files are used by a composition engine. more...

    Concept Map for performance In[A]moment.
    Concept Map for performance In[A]moment. Courtesy of Jon Bantandos


    Audio Metaphor has been employed in the contemporary dance and technology productions as well as in realtime generative performance based application systems. The application of the Audio Metaphor, as a performance application, demonstrates a positive direction in using natural language and social media for generative soundscape in real-time contexts. more...