MCML Speech Interaction SDK (SLC-SDK): Discontinued

    MCML Speech Interaction SDK is a speech application development toolkit for multilingual speech-to-speech translation and dialog systems, such as VoiceTra4U. MCML (Modality Conversion Markup Language) has been adopted as its communication protocol, as the SDK is designed to realize global communication by connecting and converting languages, texts, speech, images, and gestures to overcome the language barriers around the world; regardless the modalities. The communication protocol was internationally standardized under the activities of U-STAR (http://www.ustar-consortium.com/), a worldwide research consortium for multilingual speech translation which NICT has been taken the leading role since its establishment. By using the SDK, users can easily connect NICT's high-precision speech recognition, machine translation, text-to-speech, and dialog management servers to develop their own speech applications.

    Main Features of Ver. 2.2.1 (Released on September, 26th, 2016)
    • The following updates have been made to the Speech Dialogue Engine Package (dm-2.2.1.zip).
      • Editing and building action codes are now available on the DMBuilder.
    • The following updates have been made to the Sample Application Package for Android (sdk_Android-2.2.1.zip).
      • Disabled the GPL libraries.
      • The interval between speech data transmissions have been modified from 1/2 second to 1/8 second.
      • Added information to the document about importing projects into Android Studio and building them.
    • The following updates have been made to the Sample Application Package for iOS (sdk_iOS-2.2.1.zip).
      • Modified the project settings on Bitcode and Provisioning.
      • The interval between speech data transmissions have been modified from 1/2 second to 1/8 second.
    • The following updates have been made to the Sample Application Package for PC (sdk_PC-2.2.1.zip).
      • Refactored the dependent library, "MCMLLib".

    Download (unavailable) | See License Agreement | See README | See Privacy Policy

    The service has been discontinued. We apologize for the inconvenience this may cause and thank you for your cooperation.


Incremental Speech Recognition SDK (ISR-SDK): Discontinued

This SDK is a development toolkit for developing incremental speech recognition applications. Incremental speech recognition is a type of speech recognition that exports the recognition results successively. Therefore, the recognition results can be retrieved successively without having to wait for each user utterance to finish.

This SDK is composed of the Incremental Speech Recognition libraries and sample applications. The supported platform is Windows 7 or later.

Developers can easily add incremental speech recognition functions to their programs by calling the APIs implemented into the incremental speech recognition library.

The SDK also consists of the following 2 sample applications:

  1. A simple program which executes incremental speech recognition as it reads the audio wave file
  2. A program which executes incremental speech recognition while playing a video online, and displays the recognition results as subtitles.
Main Features of Ver. 1.0.0 (Released on September 1st, 2016)
  • Newly released.

Download (unavailable) | See License Agreement | See Privacy Policy

The service has been discontinued. We apologize for the inconvenience this may cause and thank you for your cooperation.