Skip to content
No description, website, or topics provided.
Python MATLAB
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
mds_based_analysis
optimized_metrics
README.md

README.md

corticalTimbre

Supporting scripts for the submitted paper: "Cortical modeling of context effects in perceived differences among complex sounds" by Thoret E., Caramiaux B., Depalle P., McAdams S.

This project is separated into two parts:

  1. a mds based analysis of the 17 datasets in folder './mds_based_analysis/'
  2. a part concerning the optimisation of gaussian kernels in './optimized_metrics/'

Instructions

  1. Paste the './ext/' folder (private link provided to reviewers) in the main repo './corticalTimbre/'

  2. MDS based analysis (MATLAB)

    • Run analyses with main_MDS_BASED_ANALYSIS.m
    • Generate figure by running Figure2.m
  3. Metric learning: optimisation of gaussian kernels written in python (python 3.X). We are providing 6 different scripts corresponding to the different analysis of the paper.

    • '01_optimize_metrics.py': run optimisation of between-sounds metrics (optimised metrics are logged in folder './out_folder')
    • '02_optimize_decimated_metrics.py': run optimisation of between-sounds metrics with decimated representation.
    • '03_optimized_kernels_analysis.py': run analysis of the optimised metrics
    • '04_EuclideanDistance.py': compute correlations between perceptual dissimilarities and euclidean distances between STRFs
    • '05_acoustic_interpretation.py': generalizability analysis
    • '06_correlation_with_variability.py': correlations between optimised metrics and the variability of sound representations

Depedencies

Python dependencies: tensorly, numpy, matplotlib, aifc, tensorflow

You can’t perform that action at this time.