Speech Recognition SDK for Playstation 2

The ScanSoft Games Software Development Kit (SDK) for PlayStation®2 is a flexible and easy-to-use middleware. Designed for games developers, this SDK enables easy integration of speech recognition functions into games and edutainment titles. The Games SDK is based on the ScanSoft ASR-1600 V2.0 a phoneme-based, medium vocabulary speech recognition engine. Speech recognition is well suited to numerous applications including adventure games, role playing games, strategy games; moreover, speech enables increased player interactivity and an improved overall gaming experience.

New Games Speech API (GSAPI)

The SDK ships with a runtime engine, all the necessary components required by the speech engine for accurate speech recognition, as well as a complete set of development tools, sample programs and documentation. Additionally, it offers developers a high-level of control of the Automatic Speech Recognition (ASR) engine.

The unique structure of the Games Speech API (GSAPI) allows:

Speech development to happen independently of the compile/build/debug cycle allowing contexts to be created off-line and parameters to be tested within the development environment. An evaluator for PC allows for quick testing of developed contexts.
Quick and easy integration of the GSAPI code into the game code
Advanced memory management
Easy localization without the need for new code
Key features

Accurate recognition – A phoneme-based speech recognition engine that uses proprietary signal processing techniques and advanced speech recognition algorithms to provide accurate recognition in any environment.

Speaker Independent Models – No requirement to train the system, the end user can simply plug in the microphone and start playing. End-users can even teach the system new commands for personal customization.

Flexibility – Actions can be easily attached to recognized words or phrases. Automatic detection of voice activity and trailing silence.

Scalable Vocabulary – Using the ScanSoft BNF format, new vocabularies can be easily created.

Gender (in)dependence – Option to select gender-dependent and gender-independent models, allowing optimization and maximum customization to application.

Multiple Recognition Modes – Push-to-talk, automatic stop, open microphone, and continuous operation mode.

Updated Voice Models – The language models have been trained with the voices of children as well as the standard adult male and female voices.

Optimal memory usage – The system can omit the orthographic string information, integer word IDs can be used instead.

Customizable – Full customization of speech parameters available.

Multiple languages available – Castillian Spanish, French, German, Italian, Japanese, Korean, as well as UK and US English.

Full flexibility at run-time – Multiple contexts can be loaded, several recognizers can be supported.

Easy integration – Memory management and audio signal acquisition can be done from outside the ScanSoft GSAPI (hooks) for optimal integration within other game code.

Optimized – Only one language per engine. Integrated lexical expert system includes language independent phonetic transcription system.

System Requirements

ScanSoft ASR context and grammar development:

Microsoft® Windows® 95/98/2000/Me or Windows NT® 4.0 or higher
Microsoft Windows Multimedia compatible sound board (16 bit) and CD-ROM drive
64 MB RAM
Intel® Pentium®-based PC (200 MHz) or higher
PlayStation® 2 development:

PlayStation®2 development TOOL (DTL - 1000)
PlayStation®2 runtime library release 2.x
PlayStation®2 tool chain release 2.96
Download Datasheet

NOTE: This datasheet is in the PDF format, which can be viewed with the Adobe® Acrobat® Reader® available FREE at http://www.adobe.com/.

Games SDK for PlayStation®2
(0.99 MB)