Acoustic parametrization

This part of the documentation details the different tools for acoustic parameters extraction, storage and usage. In SIDEKIT, low level interface for acoustic parametrization is implemented in the frontend module. Two high level classes allow a fast and simple extraction of acoustic parameters:

• FeaturesExtractor

• FeaturesServer

Before introducing those objects, we give a brief description of the HDF5 format that is used to store and exchange acoustic features. The HDF5 format is the prefered serialization format in SIDEKIT.

• 1. Save the features in HDF5 format
  • Saving features per audio channel
  • Saving features for a collection of audio files
  • Loading features from a HDF5 file
• 2. The FeaturesExtractor object
  • Extract features with standardized input and output filenames
  • Extract features with non-standardized filenames
• 3. The FeaturesServer object
  • 3.1 Options of the FeaturesServer
  • 3.2 Get features from a single file
  • 3.3 Get features from several files