SOURCE: Cepstral

August 20, 2007 09:11 ET

Cepstral Donates Open Source MRCP Stack to Telephony Industry

PITTSBURGH, PA--(Marketwire - August 20, 2007) - Cepstral LLC is proud to announce that it is providing the first open source Media Resource Control Protocol (MRCP) library, OpenMRCP, at no charge to telephony application developers today. Cepstral, a pioneer in Text-to-Speech (TTS) voice software, is donating OpenMRCP to the developer community in order to spur the development of telephony applications that interact with distributed speech resources such as TTS and Automatic Speech Recognition (ASR). MRCP is a standards-based communications protocol that allows telephony applications to communicate with speech resources, similar to the way TCP/IP allows browsers to communicate with web sites.

MRCP has been widely adopted by many of the largest telephony platforms and speech technology vendors. However, from its 2002 inception until now, sample MRCP source code has never been publicly published. Cepstral sponsored OpenMRCP in an effort to lower entrance barriers for developers and to focus efforts on the creation of meaningful applications. "The previous requirement for each company to re-invent the protocol layer was inefficient. By seeding the community with a freely distributed protocol stack, Cepstral hopes OpenMRCP will accelerate the development of speech applications," said Cepstral CEO, Craig Campbell.

OpenMRCP has already saved at least one "man-year" of effort for two telephony providers. FreeSWITCH, an open source telephony platform, has added MRCP v1 and v2 client support for ASR and TTS using the OpenMRCP library as part of its Beta 1 release. UK-based Cereproc is adding MRCP v1 and v2 server support for its Text-to-Speech voices. "OpenMRCP levels the playing field. MRCP compliance will allow customers to test and deploy our new high-quality TTS voices with little or no engineering effort," said Cereproc CEO Paul Welham.

MRCP has been adopted by many levels of the telephony industry. "OpenMRCP is our gift to the developer community, and we hope it will encourage the deployment of a broad range of speech applications. Speech technology is one of the keys to Web 3.0, and OpenMRCP brings us one step closer," said Campbell. The open source MRCP library is available for download at

About Cepstral LLC

Cepstral is a speech technology company based in Pittsburgh, PA, USA, which provides Text-to-Speech voices for the spoken delivery of information. We build high quality, natural sounding voices for hand-held, desktop, and server applications. Cepstral: We Build Voices™.

Contact Information