Deep Audio Transcription

Published 4/12/2021


Audio transcription is a complicated problem in the field of music information retrieval. In this work, we attempt automatic transcription directly from audio waveforms using a uses a BLSTM network and a time-series MIDI representation.

Team Members

  • Team member portrait
    Victoria Ebert


    Victoria Ebert

    B.S. Computer Science

    Doctoral candidate (NLP), University of Washington