Speech Technology Magazine

 

SpeechWorks Introduces OpenSpeech MediaServer; SpeechWorks Expands Voice Offerings for Speechify

BOCA RATON, FL - SpeechWorks International Inc. announced the availability of its OpenSpeech MediaServer, expanding the range of deployment options for its OpenSpeech Recognizer speech recognition engine.
Posted Oct 1, 2002
Page1 of 1
Bookmark and Share
BOCA RATON, FL - SpeechWorks International Inc. (Nasdaq: SPWX) announced the availability of its OpenSpeech MediaServer, expanding the range of deployment options for its OpenSpeech Recognizer speech recognition engine. OpenSpeech MediaServer supports the Media Resource Control Protocol (MRCP), an emerging standard for managing speech recognition servers in distributed system architectures. OpenSpeech MediaServer is built using OpenSpeech Recognizer to provide performance, accuracy and scalability optimized for VoiceXML applications. MRCP builds upon VoIP protocols that transport speech across packet-switched data networks. "MRCP is an important advance in speech system architecture because it clearly delineates platform software from the underlying speech technologies, expanding the configuration choices available to customers by leveraging the flexibility of VoIP routing," said Bill Meisel, president of TMA Associates and publisher of Speech Recognition Update. "MRCP complements VoIP and VoiceXML to create a complete, distributed speech application environment built wholly on open standards." SpeechWorks, Cisco Systems Inc. and other speech technology and VoIP providers helped develop MRCP to accelerate the adoption of speech applications. By implementing a common protocol for communicating with speech servers, MRCP eliminates much of the effort applied to integrating speech recognition products from multiple vendors and allows new technologies to be delivered to customers sooner. In addition, MRCP allows speech data to be routed directly to the media server, improving system scalability by avoiding the extra bandwidth required for "double trunking" data through the application platform on its way to be recognized. The MRCP specification has been submitted to the Internet Engineering Task Force (IETF) for consideration, with the expectation that it will eventually become a formal international standard. A draft of the MRCP specification is available on the IETF website at http://www.ietf.org. "SpeechWorks is committed to establishing, embracing, and promoting standards that accelerate the adoption of speech services," said Steve Chambers, chief marketing officer at SpeechWorks. "OpenSpeech MediaServer is the latest in a series of open system initiatives designed to make it easier for our partners and customers to work with SpeechWorks products."
SpeechWorks Expands Voice Offerings for Speechify
BOCA RATON, FL - SpeechWorks International Inc. (Nasdaq: SPWX) introduced Speechify Tom, a synthesized U.S. English male voice for its Speechify 2.1 text-to-speech (TTS) software. Speechify 2.1 software offers TTS for network-based applications. In addition, Speechify Tom integrates with the recorded prompts of a speech service to create a dialog with callers. Speechify Tom is the latest voice offering for SpeechWorks Speechify 2.1 TTS software, which converts text into spoken language for delivery over any phone and communications device. "Speechify Tom provides best-in-class voice quality that really shines through its delivery of information in a smooth, natural sounding manner," said Jonathan McIntosh, Senior Vice-President, Operations and Technology at West Corporation, a leading provider of integrated, outsourced customer care solutions. "We expect customers will find Speechify Tom to be approachable and easy to interact with, creating a pleasant experience for callers of our speech services." Organizations can combine Speechify Tom - or a uniquely created custom voice based on Speechify 2.1 software - with recorded prompts from an automated speech recognition (ASR) application. This capability - allowing the same voice talent to be used for both TTS and ASR- integrates TTS and real speech into one application, enabling a more natural dialog to take place with callers. It also adds flexibility to applications because edits and updates can be made instantly, and lowers the cost of speech services because less studio recording time is required. "Speechify Tom further demonstrates that voices of extremely high quality can be developed for organizations that want to use a unique, branded voice in their speech services," said Steve Chambers, chief marketing officer at SpeechWorks. "This capability will accelerate the role of TTS throughout the enterprise by delivering speech services that reinforce a corporate identity across multiple customer touch-points." Speechify uses concatenated speech synthesis where segments of recorded human voice are used to speak out the text. The product also uses advanced unit selection algorithms to model voice quality and overall rhythm and intonation.
Page1 of 1