#ViZa

tool for visual exploration of music

We implemented a visualization tool, “ViZa”, enabling to show 3d relationship between aggregated features of audio samples. This application was developed for use in a research context, specifically for short audio samples (< 10s), with the following goals: to rapidly visualize distribution and correlation of features to quickly spot outliers, in a straight way or based on t-sne’s dimension reduction algorithm to check perceptual validity of feature distribution by playing instantaneously with the mouse in the 3d space to generate user test validation of extracted features, as a simple satisfaction study on given axes and cluster distribution, or by allowing them to modify the underlying 3d space for iterative research

A self - explanatory video can be seen at : http://youtu.be/iF8_3dIZ_34

• unzip example folder
• launch Viza.app
• select example/annotation folder when asked
• wait a bit ...

essentia : branch Viza essentiaX : https://github.com/MartinHN/essentiaX and openFrameworks : https://github.com/MartinHN/openFrameworks.git : on branch Viza

#documentation can be found here