Speech-To-Text in German

[Marcirianer]

Hi,
I want to program a speech-to-text recognition in Java.
I am using the following configuration code, the file are downloaded from the recommended download link of the cmu sphinx website:

configuration.setAcousticModelPath("resource:/de-de");
configuration.setDictionaryPath("resource:/cmusphinx-voxforge-de.dic");
configuration.setLanguageModelPath("resource:/cmusphinx-voxforge-de.lm.bin");

In the feat params are:
-lowerf 130
-upperf 6800
-nfilt 25
-transform dct
-lifter 22
-feat 1s_c_d_dd
-agc none
-cmn current
-varnorm no

If I start my program many info like

The dictionary is missing a phonetic transcription for the word 'flex'

appears.

The following code

while ((result = recognizer.getResult()) != null) {
            System.out.format("Hypothesis: %s\n", result.getHypothesis());
        }
        recognizer.stopRecognition();

is executed but without result. On the console appears:

23:30:00.126 INFO speedTracker         # Name               Count   CurTime   MinTime   MaxTime   AvgTime   TotTime   
23:30:00.126 INFO speedTracker         Load LM              1       15,5220s  15,5220s  15,5220s  15,5220s  15,5220s  
23:30:00.126 INFO speedTracker         Score                4       0,0000s   0,0000s   0,0150s   0,0037s   0,0150s   
23:30:00.126 INFO speedTracker         Prune                10      0,0000s   0,0000s   0,0000s   0,0000s   0,0000s   
23:30:00.126 INFO speedTracker         Grow                 14      0,0000s   0,0000s   0,0020s   0,0003s   0,0040s   
23:30:00.126 INFO speedTracker         Frontend             4       0,0000s   0,0000s   0,0150s   0,0037s   0,0150s   
23:30:00.126 INFO speedTracker         Compile              1       7,9330s   7,9330s   7,9330s   7,9330s   7,9330s   
23:30:00.126 INFO speedTracker         Load AM              1       1,1100s   1,1100s   1,1100s   1,1100s   1,1100s   
23:30:00.127 INFO speedTracker         Load Dictionary      1       0,0310s   0,0310s   0,0310s   0,0310s   0,0310s   
23:30:00.127 INFO speedTracker            Total Time Audio: 10,20s  Proc: 0,00s 0,00 X real time
23:30:00.127 INFO memoryTracker           Mem  Total: 868,00 Mb  Free: 244,46 Mb
23:30:00.127 INFO memoryTracker           Used: This: 623,54 Mb  Avg: 622,97 Mb  Max: 623,54 Mb

What I am doing wrong? I am using the tutorial java program.
Thank you very much,

[Nickolay V. Shmyrev]

Try Vosk

https://github.com/alphacep/vosk-api/blob/master/java/demo/src/main/java/org/vosk/demo/DecoderDemo.java

[Marcirianer]

Thank you very much for your answer and solution, I could test it with english wav-file, all fine.