Chip adds speech synthesis to embedded devices
Seiko Epson Corp. has released a new synthesis chip for embedded applications featuring Fonix text-to-speech (TTS) function.
S1V30120 is a companion chip that converts text-based characters into a synthetic speech signal, which is output via the integrated DAC. Since it uses a message protocol based on a clock-synchronous serial interface, the chip can be easily integrated with various host devices and MCUs.
S1V30120 also supports decoding of ADPCM data, which allows mixtures of pre-recorded speech and synthetic speech to be combined as required. For example, fixed phrases could be implemented using pre-recorded signals, while variable phrases would be synthesized.
Target applications include systems that require voice guidance such as those embedded in home appliances, buildings and office/industrial equipment.
The product features a text-to-speech synthesis function which supports the Fonix DECtalk v5, fully parametric speech synthesis, as well as U.S. English, Castilian Spanish and Latin American Spanish languages. The function also features nine pre-defined voices and has a sampling rate of 11.025kHz.
The chip offers audio playback (ADPCM decoding, in Epson's original format) with 80-, 64-, 48-, 32- and 24Kbps bit rates and sampling rates of 16kHz and 8kHz.
Featuring a digital volume control and a synchronous serial host interface, the chip offers a 16bit audio data output with a high quality, low power consumption DAC. Sampling frequencies include 16kHz, 11.025 kHz and 8kHz.
The chip has a single clock input of 32 kHz. Its supply voltage for core is 1.8V (typ.) and I/O is 3.3V (typ.). The part is available in a 64-pin TQFP package (10-by-10-by-1.2mm, 0.5mm pitch).