Ressources

• IDyOM

  I implemented a Python version of the IDyOM model designed and originally implemented in Lisp by Marcus Pearce (2005). This model is a statistical model of melodic expectation that, from a corpus of midi files, trains a variable-order Markov chains model which can be used to predict the probability (or surprise aka information content) from an arbitraty melodic midi file. The code as well as examples and documentation is available on GitHub.

• EEG Data: Music of Silence

  Here is a Dryad repository for the data from The Music of Silence: Part I: Responses to Musical Imagery Encode Melodic Expectations and Acoustics . published in JNeuroSci. This dataset contains EEG recordings of 21 professional musicians listening to and imagining four Bach chorals along with the expectation signal computed by IDyOMpy and the acoustic envelopes.

• DeepJazz

  DeepJazz builds upon the DeepBach project realized by Gaëtan Hadjeres at Sony CSL from Bach chorals to Jazz standard generation. The code is available on GitHub.

• Source Separation

  This project realized with Valentin Bilot, Gabriel Dias Neto, Clement Le-Moine and Yann Teytaut for IRCAM master's degree presents a quick survey of the problem of source separation. We studied three state-of-the-art signal processing algorithms and compared them for remixing purposes using objective measures and a perceptive experiment. We show that these factors are not correlated and that low-separated sources with low artifact tend to be better perceived than high-separated ones with high artifact sources. Code and report are available on GitHub.

• Teaching: Statistics Classes at CogMaster

  I am currently teaching assistant for the Statistics: do it yourself! class at CogMaster cognitive science master's degree at ENS. I put all the relative information here:
  Time Slots
  Exo Intro, A little warmup.
  TP1, Music analysis: A statistical approach to the definition of musical genres.

• Bach Correli? The Game.

  Here are the audio support for the Actes de Colloque of the CREEA in Strasbourg:
  Round 1: Extrait 1 - Extrait 2
  Round 2: Extrait 1 - Extrait 2