Skip to content

KingOfBrian/VocalKit

Repository files navigation

VocalKit

I no longer advise using VocalKit, as a much better project, Open Ears has come out. http://www.politepix.com/openears/

VocalKit is a wrapper for available open source Speech related packages. It's goal is to ease the development of voice recognition solutions for the iPhone by providing a nice, simple Objective-C API.

Currently VocalKit is in an Alpha version and just wraps Pocket Sphinx. When enabled, it will post notifications for the recognized speech. It does not currently configure Pocket Sphinx programmatically, it just configures ps from a file. It also does not trigger speech processing when the user stops speaking. Anyone interested in helping is greatly appreciated.

Please review the wiki if you run into any problems.

Example:

    [vk startListening];
    ....
    [vk stopListening];

....

- (void) recognizedTextNotification:(NSNotification*)notification {
    NSDictionary *dict = [notification userInfo];
    NSString *phrase = [dict objectForKey:VKRecognizedPhraseNotificationTextKey];
    [textView setText:phrase];
}

If you want to help, the following area's could use input:

  1. Help define the API for Finite State Gramar creation

  2. Add another (or better) pocket sphinx configuration. An FSG example would be great.

  3. Add better HMM models or models for other languages.

Discussion and Questions about VocalKit should go to the Google Group or as an issue here.

Allowing support for Julius is a design goal, but not priorities.

Getting Started with VocalKit

To get started, clone the resository and open 'VocalKit/VocalKitTest/VocalKitText.xcodeproj'. This project is setup with the WSJ 5k word corpus. The test is not a great representation of Pocket Sphinx's capabilities, and I hope to get a better tuned example up soon. Copying this project is the simplest way to get started with Pocket Sphinx.

Adding VocalKit to your project (Thanks to Three20)

VocalKit is compiled as a static library, and the easiest way to add it to your project is to use Xcode's "dependent project" facilities. Here is how:
Estimated time: 5 minutes.

  1. Clone the vocalkit git repository: git clone git://github.com/KingOfBrian/VocalKit.git. Make sure you store the repository in a permanent place because Xcode will need to reference the files every time you compile your project.

  2. Locate the "VocalKit.xcodeproj" file under "VocalKit/". Drag VocalKit.xcodeproj and drop it onto the root of your Xcode project's "Groups and Files" sidebar. A dialog will appear -- make sure "Copy items" is unchecked and "Reference Type" is "Relative to Project" before clicking "Add".

  3. Now you need to link the VocalKit static library to your project. Click the "VocalKit.xcodeproj" item that has just been added to the sidebar. Under the "Details" table, you will see a single item: libVocalKit.a. Check the checkbox on the far right of libVocalKit.a.

  4. Now you need to add VocalKit as a dependency of your project, so Xcode compiles it whenever you compile your project. Expand the "Targets" section of the sidebar and double-click your application's target. Under the "General" tab you will see a "Direct Dependencies" section. Click the "+" button, select "VocalKit", and click "Add Target".

  5. Now you need to add the Audio Toolbox and iconv framework to your project. Right click on the "Frameworks" group in your project (or equivalent) and select Add > Existing Frameworks. Then locate AudioToolbox.framework and libiconv.dylib and add it to the project.

  6. Finally, we need to tell your project where to find the VocalKit and pocketsphinx headers. Open your "Project Settings" and go to the "Build" tab. Look for "Header Search Paths" and double-click it. Add the following:

    • The relative path from your project's directory to the VocalKit directory.
    • The relative path from your project's directory to the VocalKit/include directory.
  7. While you are in Project Settings, go to "Other Linker Flags" under the "Linker" section, and add "-ObjC" and "-all_load" to the list of flags.

  8. You're ready to go. Just #import "VKController.h" anywhere you want to use VocalKit Wrapper class

Modifying VocalKit to build from Pocket Sphinx or Flite sources

VocalKit comes with multi-architecture static libraries for Pocket Sphinx and Flite.
Follow these steps to configure VocalKit to compile from your custom built libraries.

  1. Build Pocket Sphinx and Sphinx Base using the 'build_for_iphone.sh' shell script.
    This will place the libraries in a platform selectable directory. Follow the README for more information.

  2. Remove include, libflite*, libpocketsphinx, and libsphinxbase from the Frameworks folder

  3. In project settings, search for "Header Search Paths" and add these paths:

  • "$(HOME)$(SDK_DIR)/include/pocketsphinx"
  • "$(HOME)$(SDK_DIR)/include/sphinxbase"
  • "$(HOME)$(SDK_DIR)/include/flite"
  1. In project settings, search for "Library Search Paths" and add "$(HOME)$(SDK_DIR)/lib"

  2. While you are in Project Settings, go to "Other Linker Flags" under the "Linker" section, and add "-lpocketsphinx", "-lsphinxbase", "-lflite", "-lflite_cmu_us_slt", "-lflite_cmulex", "-lflite_usenglish"

Packaging Considerations

Pocketsphinx and Flite are LARGE -- 16MB for all the multi-architecture static libraries! Bundled into your app it should strip out unused symbols and shrink down to around 5MB. If you are submitting to the app store, only include both flite and pocket sphinx if you are actually using features from pocketsphinx and flite.

About

Objective-C shim layer for Speech Recognition

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published