Nothing happened after recording audio

Android release: 10 QKQ1.190828.002

Nothing happened after recording audio. The audio manager works, records something, stops after a while, but there is no display of the recognized sentence. Clicking on menu items except language one does nothing.

comprise_sample_app.log

Do a git pull as there have been some changes.

The record() function accesses the personal server STT. Normally you need to set this up first.

But: package.json now has an attribute "personalizedServerURL": "https://94.130.225.120:8443" which is the server we set up remotely for easier testing. So after adding this you should be able to start within setting up the server.

I replaced record() by record2() in home.page.html

Now I don't see result of SpeechToText, but I always have the same answer "Have you tasted our new ice cream ?"

I see some minor problems in logs:

W/TextToSpeech: Couldn't retrieve ISO 3166 country code for locale: en_EN
    java.util.MissingResourceException: Couldn't find 3-letter country code for EN
        at java.util.Locale.getISO3Country(Locale.java:1776)
        at android.speech.tts.TextToSpeech$9.run(TextToSpeech.java:1492)
        at android.speech.tts.TextToSpeech$9.run(TextToSpeech.java:1477)
        at android.speech.tts.TextToSpeech$Connection.runAction(TextToSpeech.java:2312)
        at android.speech.tts.TextToSpeech.runAction(TextToSpeech.java:763)
        at android.speech.tts.TextToSpeech.runAction(TextToSpeech.java:753)
        at android.speech.tts.TextToSpeech.setLanguage(TextToSpeech.java:1477)
        at eu.comprise.COMPRISE_TextToSpeech.speak(COMPRISE_TextToSpeech.java:151)
        at eu.comprise.COMPRISE_TextToSpeech.execute(COMPRISE_TextToSpeech.java:73)
        at org.apache.cordova.CordovaPlugin.execute(CordovaPlugin.java:98)
        at org.apache.cordova.PluginManager.exec(PluginManager.java:132)
        at org.apache.cordova.CordovaBridge.jsExec(CordovaBridge.java:59)
        at org.apache.cordova.engine.SystemExposedJsApi.exec(SystemExposedJsApi.java:41)
        at android.os.MessageQueue.nativePollOnce(Native Method)
        at android.os.MessageQueue.next(MessageQueue.java:336)
        at android.os.Looper.loop(Looper.java:181)
        at android.os.HandlerThread.run(HandlerThread.java:67)

and

D/SystemWebChromeClient: ng:///HomePageModule/HomePage.ngfactory.js: Line 142 : ERROR ReferenceError: any is not defined
I/chromium: [INFO:CONSOLE(142)] "ERROR ReferenceError: any is not defined", source: ng:///HomePageModule/HomePage.ngfactory.js (142)
D/SystemWebChromeClient: ng:///HomePageModule/HomePage.ngfactory.js: Line 142 : ERROR CONTEXT [object Object]
I/chromium: [INFO:CONSOLE(142)] "ERROR CONTEXT [object Object]", source: ng:///HomePageModule/HomePage.ngfactory.js (142)

In assets/i18n/fr.json, I wrote in last line "Réponse traduite".

Is your function record2() calling vosk-api via SpeechToText with the model you put in models/src/main/assets/sync/model-android ?

Please pull for the next version (inserted your translation and fixed the TTS error, thanks for that!)

I also published a new cordova-plugin-comprise-speech-to-text, feel free to uninstall and install the latest version again.

Note: As the model with record2() method is stored all locally, i only attach one, which is english in this case. But maybe it is enough to replace the english by the french one.

So to your last question: Yes, the model is stored / used within the path you mentioned.

Thanks, it works!

I have to try another model, because this one has a relatively high WER (> 50%) comparatively to models used with https://kaldi-web.loria.fr with a WER near 0%.

closed