Specifications

69
PF1070-04
Voice recognition engine middleware
● Speech recognition middleware for the S1C33 Family
● Real-time execution in an S1C33 chip
● PC tools for making speech recognition ROM data
■ FEATURES
• Speech recognition middleware for the S1C33 Family is provided in the form of a linkable library.
• The speech recognition engine and a listening routines using an A/D converter are included in the library. Data
decompression and speaking routines with a D/A converter or PWM can be implemented using the
S5U1C330V1S library.
• Using the Seiko Epson exclusive isolated word recognition technology, 20 to 100 words can be recognized in
real-time.
•2 types of recognition methods are available; Speaker Independent Recognition for which data is created on
a PC, and Speaker Dependent Recognition that allows direct word-registration on products.
• PC tools for evaluating the speech recognition engine are provided.
•Evaluation boards (S5U1C33209D1, S5U1C330A3D1, S5U1C330M1D1) are provided.
• Suitable for PDAs, digital stationeries and digital toys.
■ HARDWARE CONFIGURATION
S1C33xxx
S1C33000
Core CPU
Built-in RAM 16-bit timer
External RAM External ROM
Bus interface
10-bit A/D
converter
D/A converter
or PWM
MIC
AMP
Mic
Speaker
POWER
AMP
In case use speak
■ MEMORY REQUIREMENTS AND SPEED
• RAM ................................................................................ 5K bytes
• ROM
Program ....................................................................... 27K bytes
Dictionary data for recognition (Cepstrum) .................. 1,280 bytes/second
Dictionary data for recognition (VQCode) .................... 128 bytes/second
• Recognition speed
20MHz operation, recognition dictionary: 30 words ..... 0.54 seconds
(Approximately 0.7 seconds for single dictionary entries such as "Good morning" or "Good afternoon")
S1C33 Family S5U1C330V2S Middleware