rtsp://[2002:914:fc12:195:0000:8a2e:0370:7334]/media/recognizerThis applies to all protocols.
For an MRCP configuration the value of InitSessionString should include the URI of the speech server’s recognition engine. You can also specify an additional URI for a back-up speech server’s recognition engine in an MRCP configuration, for example:
RecoService=Reco_GB PlugInClass=com.ibm.telephony.directtalk.mrcp.MRCPReco InitSessionString=URI=rtsp://myasrserver.example.com:554/media/recognizer InitSessionString=URI=rtsp://mybackupasrserver.example.com:554/media/recognizer RecoType=Recoen_GB ;
The first InitSessionString entry must reference the primary server, and the second entry must reference the backup server.
To specify speech recognition using dynamic engine allocation, so that speech recognition engines are not assigned for the duration of call, but instead are allocated only for the duration of each period of speech recognition, append ,keepsessionforcall=no to the end of the InitSessionString URI value. The default setting is keepsessionforcall=yes. See Allocating speech recognition and TTS engines for more information. The vendor of your speech product may not support the use of dynamic engine allocation.
The RecoType keyword is a platform-independent label for the RecoService entry. RecoService entries for the same technology on different platforms must therefore share the same RecoType.