iFLY Mobile Speech Platform: Enabling Voice Recognition and Synthesis

Speech recognition (ASR) and speech synthesis (TTS) convert between natural language and binary symbols, enabling voice input to replace keyboards and shaping daily life. Apple’s Siri exemplifies this shift.

The iFLY Mobile Speech Platform (MSP) provides a cloud‑based architecture with load balancing, parallel computing, and data storage, supporting Android, iOS and other terminals.

ASR lets machines “listen” and extract textual information from spoken language, serving call centers, telecom services, and enterprise systems, and is considered a revolutionary human‑machine interface technology.

TTS automatically transforms any text into natural, continuous speech, meeting the demand for real‑time, personalized audio services.

MSP’s main goals are to deliver multi‑channel voice synthesis, recognition, and dictation over 2G/3G and Internet networks, and to offer a unified development interface for voice applications on mobile and desktop platforms.

The platform follows a C/S model with four logical layers: Speech Programming Interface (SPI), Mobile Speech Client (MSC), Mobile Speech Server (MSS), and MSP Infrastructure. These layers together provide APIs, network communication, audio codec, voice activity detection, protocol parsing, server‑side recognition engines, management tools, load balancing, parallel processing, and data storage.

MSC supports various client environments, including Android, iOS, Symbian, Windows Mobile/CE, and MTK.