TU Logo   IFS Logo Vienna University of Technology
Institute of Software Technology and Interactive Systems
Java Audio Feature Extraction
[Audio Feature Extraction Home] [People] [Publications]

Step-by-Step Guide

This step-by-step how-to will guide you through the process of extracting various features from audio files using the Java Audio Feature Extraction.


1. Start the application

You can start the Java Feature Extraction application from in a terminal by invoking
java -jar AudioFeatureExtraction.jar
or by just double-clicking the jar file.

2. Select input files

You can choose to search a directory tree for audio-files or select a text-file containing one filepath per line

screenshot 1

Further select the output directory (the directory where the feature-files will be placed) and enter a common basename for the featurefiles. The filenames will be <basename.rp.vec, <basename.rh.vec and <basename.ssd.vec respective.

3. Choose features to extract

Check/uncheck the features-sets you want to extract.

Further select the number of parallel extraction threads and whether you want compressed feature files (saves up to 52% diskspace).

4. Start the extraction process

screenshot 2

You can dynamically change the number of parallel extraction threads.

5. Enjoy your features

screenshot 3

E.g. try the Java SOMToolbox with your features!