Recognize the speech via wav file by pocketsphinx_batch

[Toan Nguyen]

Dear cmusphinx,

I noticed that the tool pocketsphinx_batch from pocketsphinx can recognize
the speech via wav file.

Could you please tell me how to use this tool in android os or write some
source code to implement exactly like it does.

Yours sincerely, Toan

[Nickolay V. Shmyrev]

The code to recognize file on Android is outlined here:

http://stackoverflow.com/questions/29008111/give-a-file-as-input-to-pocketsphinx-on-android

[Nickolay V. Shmyrev]

Also here

https://sourceforge.net/p/cmusphinx/discussion/help/thread/e8ea1585/?limit=25#9501/528d

[Toan Nguyen]

@Mr. Nickolay V. Shmyrev
I already tried your solution. But my app gets crashed at the command line below

Config c = Decoder.defaultConfig();

Please tell me how to fix it.

[Nickolay V. Shmyrev]

I'm not sure what do you mean by "crashed", you need to provide more details.

[Toan Nguyen]

It means my android app is forced close instantly as soon as running the command above

[Nickolay V. Shmyrev]

It must print something in logcat, you should check it out first of all.

[Toan Nguyen]

The attachment file is the log error
I am using the source code from github
https://github.com/cmusphinx/pocketsphinx-android-demo

[Nickolay V. Shmyrev]

And, did you try to think on this error? Did you try to google for it? This error is very easy to solve if you turn on your brains for a second.

[Toan Nguyen]

I already added jni so file to your project properly but it does not work

[Toan Nguyen]

Could you please try to run your sample code like you suggest to check whether it works or not

[Nickolay V. Shmyrev]

I already added jni so file to your project properly but it does not work

No you did not load the JNI library, this is what your log says.

Could you please try to run your sample code like you suggest to check whether it works or not

It works fine

[Toan Nguyen]

Sorry, I already figured out the bottom line here is

I have to add the command below

    recognizer = defaultSetup()
            .setAcousticModel(new File(assetsDir, "de-cont"))
            .setDictionary(new File(assetsDir, "voxforge.dic"))

                    // To disable logging of raw audio comment out this call (takes a lot of space on the device)
            .setRawLogDir(assetsDir)

                    // Threshold to tune for keyphrase to balance between false alarms and misses
            .setKeywordThreshold(1e-45f)

                    // Use context-independent phonetic search, context-dependent is too slow for mobile
            .setBoolean("-allphone_ci", true)

            .getRecognizer();
    recognizer.addListener(this);

[Toan Nguyen]

Before the command line

Config c = Decoder.defaultConfig();

And It works