de.sciss.strugatzki.FeatureExtraction
The input audio file to extract the features from.
The channel behaviour determines how to handle multichannel files.
The channel behaviour determines how to handle multichannel files. Currently the feature vectors are calculated on a mono signal only. This setting determines whether multiple channels in the input audio file are mixed together, or if just the first or just the last channel is used in the extraction process.
The output "audio" file to write the feature vectors to (this will be in AIFF format).
The FFT overlap factor used to step from vector to vector.
The FFT overlap factor used to step from vector to vector. This equals fftSize / stepSize, so a value of 2 means the window step is half of the fft size (windows are 50% overlapping).
The FFT size used to calculate the feature vectors.
An optional file to which the extraction settings are saved (in XML format).
The number of MFCC used for the spectral feature.