Transcribe .mp3 files with Transcriber.java

[Gorka Perez Sánchez]

Hello , I have tested the Transcriber.java file and I have tried to trascribe
a .mp3 file . When the code :

URL configURL = Transcriber.class.getResource("config.xml");

ConfigurationManager cm = new ConfigurationManager(configURL);

Recognizer recognizer = (Recognizer) cm.lookup("recognizer");

AudioFileDataSource dataSource = (AudioFileDataSource)
cm.lookup("audioFileDataSource");**

dataSource.setAudioFile(audioURL, null);

is executed I obtain the following Exception in the comand
dataSource.setAudioFile(audioURL, null) :

Audio file format not supported:
javax.sound.sampled.UnsupportedAudioFileException: could not get audio input
stream from input URL
javax.sound.sampled.UnsupportedAudioFileException: could not get audio input
stream from input URL
at javax.sound.sampled.AudioSystem.getAudioInputStream(Unknown Source)
at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setAudioFile(AudioFileData
Source.java:150)
at edu.cmu.sphinx.demo.transcriber.Transcriber.main(Transcriber.java:56)
Exception in thread "main" java.lang.NullPointerException
at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setInputStream(AudioFileDa
taSource.java:178)
at edu.cmu.sphinx.frontend.util.AudioFileDataSource.setAudioFile(AudioFileData
Source.java:162)
at edu.cmu.sphinx.demo.transcriber.Transcriber.main(Transcriber.java:56)

How can I configure it for transcribing .mp3 files (the config.xml file or
other parameters in the project) ?.
Thanks
Gorka**

[eliasmajic]

Mp3 is compressed data. You need to convert it to a .wav file first.

You can use ffmpeg to do this.

[Nickolay V. Shmyrev]

Or you can use this to add mp3 to Java Sound

http://www.tritonus.org/

[Anonymous]

In my case, I added tritonus_mp3-0.3.6.jar and
tritonus_share-0.3.6.jar fom
http://www.tritonus.org/ to my java lib/ext
folder. This should add the MP3 capabilities to Sphinx. But the above
(Transcriber.java) code will not work as it is because AudioFileDataSource
is not accommodating MP3 formats. To take care of that part, either you modify
AudioFileDataSource.setAudioFile() method or just not call this method.
Here is how I changed Aligner.java to allow MP3.

public class Aligner {

    public static void main(String[] args) throws IOException, UnsupportedAudioFileException {

        if(args.length<2){
            System.out.println("Error: two parameters required <audio-file> <script-text>. Stopping further processing");
            return;
        }

        ConfigurationManager cm = new ConfigurationManager("config/aligner.xml");
        Recognizer recognizer = (Recognizer) cm.lookup("recognizer");

        TextAlignerGrammar grammar = (TextAlignerGrammar) cm.lookup("textAlignGrammar");
        grammar.setText(args[1]);

        recognizer.addResultListener(grammar);

        /* allocate the resource necessary for the recognizer */
        recognizer.allocate();

        // configure the audio input for the recognizer
        URL audioFileURL = new URL("file:" + args[0]);
        AudioFileDataSource dataSource = (AudioFileDataSource) cm.lookup("audioFileDataSource");

        if(args[0].endsWith(".mp3") || args[0].endsWith(".MP3")){//if non-MP3
            AudioInputStream audioStream=null, decodedAudioStream = null;
            AudioFormat baseFormat=null, decodedFormat = null;          
            try {
                audioStream = AudioSystem.getAudioInputStream(audioFileURL);
                baseFormat = audioStream.getFormat();
                decodedFormat = new AudioFormat(AudioFormat.Encoding.PCM_SIGNED,
                        baseFormat.getSampleRate(),
                        16,
                        baseFormat.getChannels(),
                        baseFormat.getChannels() * 2,
                        baseFormat.getSampleRate(),
                        false);
                decodedAudioStream = AudioSystem.getAudioInputStream(decodedFormat, audioStream);
            } catch (UnsupportedAudioFileException e) {
                System.err.println("Audio file format not supported: " + e);
                e.printStackTrace();
            } catch (IOException e) {
                e.printStackTrace();
            }
            dataSource.setInputStream(decodedAudioStream, null);            
        }else{
            dataSource.setAudioFile( audioFileURL, null);//it does everything it needs to for the supported audio formats
        }

        Result result;
        while ((result = recognizer.recognize()) != null) {

            String resultText = result.getTimedBestResult(false, true);
            System.out.println(resultText);
        }
    }
}

You may need to compare with the stock Aligner.java that comes with Sphinx to
see my changes.

I hope it helps

Pannu