Technical Details

Creaceed will provide basic instructions to set up CeedVocal SDK in your iOS application.

CeedVocal SDK:
  • is based on the open source projects Julius (decoder) and Flite (phonetizer).
  • is provided in the form of a static library with header files. The programming language of the SDK is Objective-C.
  • is a multi-locutor Automatic Speech Recognition SDK, meaning that no training is required.
  • supports isolated words recognition from a user provided list of up to 5,000 words (can take from 0 to 2 seconds to output result, depending on word count).
  • supports keyword spotting. No grammar based recognition at this time.
  • is optimized for fast launch times (about 3 seconds for loading acoustic model, excluding phonetizations)
  • provides a phonetizer which is sufficient for common words, but won't phonetize neither abbreviations nor numbers, and won’t give access to phoneme sequences.
  • is available in 6 languages: English, French, German, Dutch, Spanish and Italian.
  • provides Voice Activity Detection (VAD), but may need manual cut in noisy environments.
  • is provided as a Universal Binary library (armv7/armv7s/arm64/i386/i86_64) for easier setup within the Simulator.
  • iOS 6.0 SDK required for development, and deployment possible on iOS 5, 6 and 7 devices. Deployment on iOS 4.3 still possible but with armv7 architecture only.
  • supports 32-bit and 64-bit architectures.

En | Fr | De | Nl | Es | It