How to extract features from opensmile




For installation of openSMILE toolkit please refer to the link 
Installation of openSMILE

How to extract features using openSMILE toolkit?

Demo Feature file to check if openSMILE is working completely
This should generate a file input.energy.csv in the output folder path
If the SMILExtract is not added to the /usr/bin path then you need to use ./SMILExtract from the openSMILE folder path

./SMILExtract –C config/demo/demo1_energy.conf –I <input_folder_path>/input.wav –O <output_folder_path>/input.energy.csv

Chroma Features

./SMILExtract -C config/chroma_fft.conf -I input.wav -O chroma.csv

./SMILExtract –cfgFileTemplate –configDflt cWaveSource,cFramer,cEnergy,cCsvSink -1 1 2>
myconfig/demo1.conf


MFCC Features

./SMILExtract –C config/MFCC12_0_D_A.conf –I input.wav –O output.mfcc.htk

PLP Features

./SMILExtract –C config/PLP_E_D_A.conf –I input.wav –O output.plp.htk





 Prosody Features
  ./SMILExtract –C config/prosodyShs.conf –I input.wav –O prosody.csv

Comments

Post a Comment

Popular posts from this blog

SOX - Sound eXchange - How to use SOX for audio processing tasks in research.

Sox of Silence - Original post - http://digitalcardboard.com/blog/2009/08/25/the-sox-of-silence/

How to get video or audio duration of a file using ffmpeg?