SpeechTech MRCP server product allows the use of modules of speech recognition (ASR) and synthesis (TTS) through a universal protocol MRCP, which is becoming almost a standard for uniform access to multimedia content and services in a networked environment. Through the MRCP protocol is possible to use voice services such as ASR and TTS of most of commercial IVR systems of reputable companies. Using our MRCP Server the customer can integrate our ASR and TTS modules with many IVR systems in different environments and configurations.
SpeechTech Text-To-Speech MRCP server
- Available TTS cores: SpeechTech TTS 2.6 a SpeechTech TTS 2.9
- Excellent voice quality. Available voices:
- TTS 2.6: Jan CZ, Alena CZ, Tomáš CZ, Beáta SK
- TTS 2.9: Jan CZ, Alena CZ, Beáta CZ
- Ability to scale for TTS version 2.9
- Extreme Computing Performance and modesty in version 2.6 TTS
- Based on open standards - to protect your investment
- Support for Media Resource Control Protocol - MRCPv1 a MRCPv2
- Connects to VOIP infrastructure using SIP
- can be connected to conventional telephone infrastructure
- Collaborates with other major manufacturers IVR systems (GVP, IBM …)
- Supports 32-bit and 64-bit platforms: Windows XP / Server 2003 / 7 Pro / Server 2008
- Supports SSML protocol for entering text
- Allows developers to easily and immediately use TTS in the IVR system
- Allows sharing of power with SpeechTech ASR MRCP server - better use of your resources
SpeechTech ASR MRCP server
- ASR core suitable for speech recognition using grammars or dictionaries
- Excellent acoustic recognition models prepared for the telephone domain
- Quality acoustic models eliminates the problem of environmental noise
- ASR Standard variant
- recognizes the utterances of size of vocabulary up to 1000 words
- Appropriate for recognition of short lists of dozens of items such as numbers 0-9, a list of commands, list of services, list of products, shorter list of municipalities, the list of countries, etc.
- Preparation of recognition grammars is immediate - in real time
- ASR Large variant
- recognizes the utterances of size of vocabulary up to 100 000 words connected to the potential millions of individual utterances
- Suitable for large lists - such as full name, complete list of municipalities in the Czech R ./ Slovak R. sentences containing more simple inputs (eg, questions such as: product and enter your phone number)
- Preparation of recognition grammar is time consuming (hours). It is needed such as automated offline preparation every 24 hours
- Very demanding of hardware - upon complexity of the recognition model (grammar)
- Based on open standards - to protect your investment
- Supports Media Resource Control Protocol - MRCPv1 a MRCPv2
- Supports for input and output protocols ABNF, SRGS a NLSML
- Optionaly supports Precompilation of complex recognition grammars into binary images
- cts to VOIP infrastructure using SIP
- can be connected to conventional telephone infrastructure
- Collaborates with other major manufacturers IVR systems (GVP, IBM …)
- Supports 32-bit and 64-bit platforms: Windows XP/server 2003/7 PRO/server 2008
- Allows sharing of power with SpeechTech ASR MRCP server - better use of your resources