Sakrament company Introduces Sakrament SpeechBOX for its TTS Engine. A High Quality solution for Automotive and Devices Market.
MINSK, Belarus —
BUSINESS WIRE — August 14, 2002 — Sakrament company, the leading software developer in the sphere of speech recognition and text-to-speech (TTS) technologies for the Russian language, today introduced Sakrament SpeechBOX for Sakrament TTS Engine, an evaluation board that designed for demonstration purposes of an embedded TTS engine. It is ideal for embedded solutions such as automotive, hand-held computers, and home entertainment systems.
TTS Software
Sakrament TTS comes with male and female voices and can read any text-based information.
Software was written on ANSI C, except hardware support of addresses. It can be easily embedded in any existing type of microprocessor platforms. In addition, it was programmed without use of any TI libraries. The size of the executable module is 48 KB. Source data volume for speech processing (allophone database) with high sound playback quality is 800 KB. Allophone database was digitized at 16 bit with sampling frequency 11 kHz. Allophone database can be installed into RAM, ROM or to flash memory, giving users flexibility in storage requirements. The database can be compressed in 4 and 8 times using real-time ADCPM algorithms.
Evaluation board
Sakrament SpeechBOX was designed for deploying directly out-of-the-box but also tailored to specific customer requirements by addition required components on the board or removal add-in components from it. Evaluation board provides ability to integrate Sakrament SpeechBOX in/with a various types of electronic devices, PDAs, home entertainment systems and vehicles.
There are 16-bit TMS320C5402 100 MHz microprocessor (without hardware support of float arithmetic), 2 MB flash memory for the program and allophone database, codec for digitizing and playback, UART for communication with PC or other device, keyboard and microphone inputs, LED display (for debug purposes) on the board.
