VoiceIn Game Edition Features

Fonix VoiceIn GE SDK 5.0 provides greater functionality, flexibility, recognition accuracy and control to game developers, enabling them to deliver a more advanced voice driven interface to enhance a player's game experience.


PDF specification sheetSDK Features

  • 'Push-to-talk' no longer required: The recognizer automatically detects speech without pushing a button first, and reports what is recognized when the player stops speaking. Players can pause for any length of time before speaking the next command and still be recognized, all without buttons.
  • Adaptation tool added: Developers can now improve and optimize speech recognition accuracy for similar words or for difficult-to-understand dialects, resulting in dramatic improvements to speech recognition. In some very difficult cases error reductions from 20 to 50 percent have been achieved.
  • Neural networks improved: Newly architected neural networks provide better recognition accuracy in noisy game environments, while maintaining the low memory and MIPS required by game developers.
  • Create new vocabularies during game play: Developers can now design games that enable players to add new command sets or vocabularies during game play. Players can create nicknames for other players or for existing voice commands by simply adding new text.
  • Sony Playstation 3 Spurs SPU support: Spurs Synergistic Processing Unit (SPU) support enables the speech recognition engine to run on a dedicated SPU to offload speech computations from the main processor. Speed increases from 30 to 60 percent have been measured, enabling developers to create larger, more complex vocabularies or command sets.
  • Added language support: Native Unicode support for Japanese and Korean characters has been added to VoiceIn GE 5.0. Game developers can now use Katakana, Hiragana, and Hangul characters without conversion. UK and US English, French, German, Italian, Japanese, Korean and Spanish continue to be supported.
  • Independent memory manager: For games that enable players to create new vocabularies during play (dynamic memory allocation), developers can use their own memory managers, giving them greater flexibility and control over memory usage.
  • Noise threshold: Developers can now design games that give players the option of adjusting the noise threshold that triggers recognition, which may improve speech recognition accuracy in noisy environments. The noise threshold can be adjusted to optimize performance across different headsets and game consoles.