openSMILE 2.3 available now for download
The brand-new version 2.3 of audEERING’s well-known audio analysis toolkit openSMILE is now available for download on the openSMILE page. It includes several new features:
Android JNI integration: While version 2.2. did provide first support for compilation as static binary for Android, we now offer integration into Apps via the JNI. An example project and a tutorial for Android Studio are provided. The tutorial shows how to build a live audio analysis app that displays audio parameters in real-time in the App UI.
Updated configuration file interface: Commandline options for audio input and data output formats have been standardized through modular config files.
Batch feature extraction GUI for Windows and batch extraction scripts for Linux: These make it easier to extract audio features from a collection of audio files.
Improved backwards compatibility: All standard feature sets are verified to be backwards compatible to Version 2.2 and 2.1. The original feature sets of openEAR are also backwards compatible with the original openEAR models.
Updated Version of ComParE 2013-2015 baseline acoustic parameter set: Several optimisations to the feature extraction code were applied and an updated version of ComParE was released. It has the same features as the previous version, but is now numerically optimized.
Various fixes and performance improvements