android - Android 上的 Pocketsphinx 找不到声学模型定义 mdef

Question

我已经将 so 文件复制到 libs 文件夹和 sync 文件夹以及 Assets 中的 assets.lst 文件。我还在应用程序中添加了 pocketsphinx-android-0.8-nolib.jar。我还补充说:

<uses-permission android:name="android.permission.WRITE_EXTERNAL_STORAGE"/> 

list 文件的权限。

然后我添加了用于设置 SpechRecognizer 的代码

private void setUpRecognizer(){
    Assets assets = new Assets(context);
    File assetDir = assets.syncAssets();
    File modelsDir = new File(assetsDir, "models");
    recognizer = defaultSetup()
        .setAcousticModel(new File(assetsDir, "hmm/en-us-semi"))
        .setDictionary(new File(assetsDir, "dict/cmu07a.dic"))
        .setRawLogDir(assetsDir).setKeywordThreshold(1e-40f)
        .getRecognizer();
    recognizer.addListener(this);
    File modelsDir = new File(assetsDir, "models");
    recognizer = defaultSetup()
        .setAcousticModel(new File(assetsDir, "hmm/en-us-semi"))
        .setDictionary(new File(assetsDir, "dict/cmu07a.dic"))
        .setRawLogDir(assetsDir).setKeywordThreshold(1e-40f)
        .getRecognizer();
    recognizer.addListener(this);
    recognizer.addKeyphraseSearch(KWS_SEARCH, KEYPHRASE);
    File menuGrammer = new File(modelsDir, "grammar/menu.gram");
    recognizer.addGrammarSearch(MENU_SEARCH, menuGrammer);
    File digitGrammer = new File(modelsDir, "grammar/digits.gram");
    recognizer.addGrammarSearch(DIGITS_SEARCH, digitGrammer);
    File languageModel = new File(modelsDir, "lm/weather.dmp");
    recognizer.addNgramSearch(FORECAST_SEARCH, languageModel);
}

运行应用程序后，出现异常:

08-26 13:42:38.301: E/cmusphinx(14405): ERROR: "acmod.c", line 90: Folder '/storage/sdcard0/Android/data/com.perscitussln.sherbet/files/sync/hmm/en-us-semi' does not contain acoustic model definition 'mdef'

谁能帮我解决这个问题或给我一些建议？

Answer 1

这个:

   File modelsDir = new File(assetsDir, "models");
   recognizer = defaultSetup()
        .setAcousticModel(new File(assetsDir, "hmm/en-us-semi"))

必须改为

   File modelsDir = new File(assetsDir, "models");
   recognizer = defaultSetup()
        .setAcousticModel(new File(modelsDir, "hmm/en-us-semi"))

或对此:

    recognizer = defaultSetup()
         .setAcousticModel(new File(assetsDir, "models/hmm/en-us-semi"))

下次注意点。