I tried to use pyAudioAnalysis.