bootphon / sustained-phonation-features

Python package for the extraction of speech features for sustained phonation

Sustained Phonation Features

Code to replicate the features from the paper https://arxiv.org/abs/2006.05365 to appear at Interspeech 2020.

Installing the environment

To run the experiment's notebook, the creation of a conda environment is required. Make sure you have conda installed, and are cd'ed into the project's folder, and run

conda env create -f environment.yml
conda activate pho_features

You should be then good to go

Extracting the features

Make sure you have installed and activated the aforementioned conda environment, then launch a jupyter notebook, and reach the Sustained_Phonation.ipynb file in the interface.

jupyter notebook 

References

.. [1] Riad, R, Titeux, H, Lemoine, L., Montillot J. Hamet Bagnou, J. Cao, X., Dupoux, E & Bachoud-Lévi A.-C.
       *Vocal markers from sustained phonation in Huntington's Disease.*
       In: INTERSPEECH-2020

If you use the Modulation Power Spectrum features, please cite also:

.. [2] Elie, J. E. and F. E. Theunissen 
       *Zebra finches identify individuals using vocal signatures unique to each call type.*
       In: Nature Communications. 2018