![]() |
|
|
However, traditional approaches to deploying speech enabled services burden general purpose servers with both speech and non-speech functions. For example, a typical solution requires VoiceXML script interpretation, DTMF processing, announcement playback, and telephony interfaces in addition to speech synthesis and speech recognition. When the processing needs of all these capabilities are combined to deliver a complete speech solution, the number of concurrent speech channels supported by even the most powerful general purpose servers rapidly declines and the price per channel climbs. Convedias approach, however, overcomes these limitations by optimally dividing the various tasks between general purpose servers, dedicated for the speech synthesis and speech recognition function, and Convedias highly optimized IP Media Servers dedicated to performing all the other supporting tasks. Communication between Convedias Media Servers and the speech processing servers is accomplished through a new protocol known as Media Resource Control Protocol (MRCP) and is fully compatible with ASR and TTS engines from leaders such as Nuance and SpeechWorks. "By using a combination of workstations and IP Media Servers, service providers can realize improved scalability and lower costs," said Grant Henderson, Convedias Executive Vice-President of Marketing and Strategy. "With the addition of MRCP to our long list of open media control protocols, Convedia reinforces the value and critical importance of an open, shared resource media processing platform in a next generation service architecture." Convedia is demonstrating its speech capabilities
and a number of enhanced service solutions based on its family of Media
Server products this week at SUPERCOMM 2003 in Atlanta.
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||