1

I'm running the LIUM 8.4.1 Jar with this command from the quickstart:

java -Xmx2024m -jar ../diarization/lium_spkdiarization-8.4.1.jar --fInputMask=./pcd.wav --sOutputMask=./showName.seg --doCEClustering showName

However, the program fails with the following Unsupported Audio File Exception. How does this error crop up and what am I missing to solve it? My set up has been to install cmusphinx4 from source, then download the lium jar. 

Diarization    INFO  | Diarization tuning     {main() / 10}
Diarization    INFO  | Initialize segmentation        {initialize() / 10}
MainTools      FINER | showName=showName      {splitHypotesis() / 10}
Diarization    FINER | -------------------------------------------    {run() / 11}
Diarization    FINER | --- showName ---       {run() / 11}
Diarization    FINER | -------------------------------------------    {run() / 11}
SEVERE AudioFeatureSetFa  UnsupportedAudioFileException
javax.sound.sampled.UnsupportedAudioFileException: file is not a supported file type
        at javax.sound.sampled.AudioSystem.getAudioFileFormat(AudioSystem.java:1059)
        at fr.lium.spkDiarization.libFeature.AudioFeatureSetFactory.getAudio(AudioFeatureSetFactory.java:162)
        at fr.lium.spkDiarization.libFeature.AudioFeatureSetFactory.MakeFeature(AudioFeatureSetFactory.java:207)
        at fr.lium.spkDiarization.libFeature.AudioFeatureSetFactory.MakeMFCCFeature(AudioFeatureSetFactory.java:255)
        at fr.lium.spkDiarization.libFeature.AudioFeatureSet.read(AudioFeatureSet.java:898)
        at fr.lium.spkDiarization.libFeature.AudioFeatureSet.setCurrentShow(AudioFeatureSet.java:1322)
        at fr.lium.spkDiarization.system.Diarization.ester2Diarization(Diarization.java:719)
        at fr.lium.spkDiarization.system.Diarization.run(Diarization.java:634)
Nikolay Shmyrev
  • 24,897
  • 5
  • 43
  • 87
user592419
  • 5,103
  • 9
  • 42
  • 67
  • So, do you understand the exception text? It clearly says what is the issue. – Nikolay Shmyrev Sep 23 '13 at 20:38
  • If it's that wav is an unsupported file type, then that would be surprising given the example they provided. Am I interpreting it incorrectly? – user592419 Sep 23 '13 at 20:45
  • I'll try resampling like you mention here: http://stackoverflow.com/questions/7613089/open-source-software-for-transcribing-speech-in-audio-files?rq=1 – user592419 Sep 23 '13 at 20:46
  • That did not work. I'm still unsure why this wav file is not supported given the quickstart. Your help would be appreciated. – user592419 Sep 23 '13 at 20:55
  • 1
    There could be many issues with WAV files, for example they might have extra headers and so on. If you don't know what is the issue with the file you can share it to let others look. Also, you can "standartize" wav by copying it with sox: sox your.wav good.wav – Nikolay Shmyrev Sep 23 '13 at 21:07
  • I tried doing sox pcd.wav pcd_standard.wav but got the same output. I'll check out the header file info to see if there's something off there. Thanks Nikolay. – user592419 Sep 23 '13 at 21:17
  • 1
    try `sox pcd.wav -r 16000 -2 -s -c 1 pcd_standard.wav` – Nikolay Shmyrev Sep 23 '13 at 21:23
  • Also see http://stackoverflow.com/questions/2843847/workaround-for-unsupportedaudiofileexception – Nikolay Shmyrev Sep 23 '13 at 21:33
  • Thanks Nikolay. Your suggestion to do the longer command worked. Much appreciated sir, I wish I could give you cred. – user592419 Sep 24 '13 at 17:49

0 Answers0