Voice Controlled Gaming
Voice Controlled Gaming, as the name suggests, comprises of two different application domains merged together. The primary focus of this project is the development of a voice recognition algorithm (which can recognize a definite number of pre-defined words that can be used as the game controls). The voice recognizer has then been integrated with a game (acquired as an opensource code) to
demonstrate its functionality as a voice-based controller.
The design therefore involves the speech input device to accept audio, which is then processed by the voice recognizer and transformed into plain text. The game accepts text based inputs from the recognizer, processes the control and shows the relevant player movement on the GUI. The game is also capable of speaking back (voice synthesis) some feedback to user, along with the GUI
rendering, for a more interactive user experience.
The voice recognition algorithm is written in Java and uses JSAPI Library for speech technology interface. The JSAPI uses the Operating System's onboard Native Speech Engine (the Microsoft SAPI in case of windows environment) to access the core functionalities of speech technology. The library itself acts as an interface to the Native Speech Engine (NSE).
Design Specification [PDF]