/UnityAndroidSpeechRecognition

This repository is a Unity plugin for Android Speech Recognition (based on Java implementation)

Primary LanguageJavaGNU General Public License v3.0GPL-3.0

UnityAndroidSpeechRecognition

This repository is a Unity plugin for Android Speech Recognition (based on Java implementation). This is just a basic implementation which uses the inbuilt Android Speech Recognition. I have implemented some additional functions to give some basic functionality.

Features

  • Set Recognition Language
  • Set the max number of results
  • Set the prompt question

Usage

  • You can first test out the SpeechRecognition.apk from here. You need a minimum of Android Lollipop (API Level 21) to use this apk.
  • Next, import SpeechRecognition.unitypackage into your Unity project. (Tested latest on Unity 2018.1.0f2)
  • To include SpeechReconginition in your own project copy the contents of Plugins/Android/ into the same folder in your project and have a look at Test.cs and ReceiveResult.cs for an example usage. Test.cs has function for calling the speech recognition popoup whereas ReceiveResult.cs has function which receives the recognized text.
  • When you press the button, a speech recognition itent should pop up. This intent sends the recognized text to 'void onActivityResult(string recognizedText)' function of the object 'Main Camera' (You can change the object to which ReceiveResult.cs should be attached by changing the object name in Test.cs. The default return object is set to 'Main Camera').
  • You can change the following line pluginClass.CallStatic("setLanguage", "en_US") in Test.cs to change the language. A list of languages along with their tags is present in the languages.txt file. This file is not exhaustive.
  • Similarly you can also manipulate other features like maximum number of results etc (In case of the max number of results, this is literally maximum and the number of results you get may not always be this number, but it would be definitely less than this number).

Experimenting

  • You can import SpeechPlugin to Android Studio to experiment with it (pluginlibrary module is only important).
  • Once you make changes and build the module you should have a pluginlibrary-release.aar file in SpeechPlugin/pluginlibrary/build/outputs/aar folder. Rename this to speechassist-debug.zip and then unzip it. On unzipping you will get a classes.jar file which you will have to copy into Plugins/Android folder in Unity along with the AndroidManifest.xml. The manifest specifies the activity class which extends the UnityPlayerActivity and hence is necessary. If you already have a manifest in plugins folder, I would recommend against copying this manifest as it can cause confilict. Instead, you can modify the manifest which you already have to specify the activity class which extends the UnityPlayerActivity.
  • One more thing to observe is that I have classes.jar in my libs folder of my AndroidStudio project. Have a look at this post to know where you can find it.

Note

  • I have extended UnityPlayerActivity with my custom activity and forced to subsitute it by changing the manifest file. Though, this works it is undesirable when you want to make/use multiple plugins. An ideal way would be to create a custom activity that will be short-lived and its sole purpose would be to get created, do the job (user picks photo from gallery e.g.), deliver result (picked image) into onActivityResult and terminate itself. Please refer this post for more details on this approach. I will update the plugin once I have more free time.

  • On experiencing a problem, feel free to create an issue.

  • If you want any additional features too, feel free to create an issue.