Streaming media properties (MRCP)

The following parameters are part of the MRCP specification, and can be controlled via the mediaserver.conf file. It is located by default in mrcp-api/docker/lumenvox/media_server.con

They may also be set with the appropriate header as part of a RECOGNITION or SET-PARAMS method. Refer to the MRCPv2 draft, section 9.4, for more information about changing the parameters using recognizer headers. A majority of these parameters are hidden from view in the configuration file since we recommend usage of the headers in SET-PARAMS or RECOGNITION to set these values. However, entering any of these values in the [MRCP] section of the media server config file should work.

See Configuring Parameters for more information about changing various MRCP parameters.

nbest_length

When the recognizer matches audio with a grammar, it may come up with more than one alternative match because of confidence levels in certain words or conversation paths. By default, the Speech Engine will only return the best match above the confidence threshold. The nbest_length parameter determines how many different results to return. All alternatives must still be above the confidence-threshold. A value greater than one does not guarantee that the recognizer will send the requested number of alternatives. The minimum value for this field is 1.

Default: 1

confidence_thrsld

When a recognition resource recognizes an utterance with some portion of the grammar, it associates a confidence level with that conclusion. The confidence-threshold parameter tells the recognizer resource what confidence level should be considered a successful match. This is a number from 0-1 indicating the recognizer's confidence in the recognition. If the recognizer determines that its confidence in all its recognition results is less than the confidence threshold, then it returns no-match as the recognition result.

Default: 0.45

sensitivity_lvl

Controls the sensitivity level for barge-in. Tthe higher the value, the more easily the system will barge-in. This parameter currently maps to the volume sensitivity in the LumenVox API, however, in the LumenVox API, a higher value refers to lower sensitivity. Another similar value to control sensitivity is available in our Vendor-Specific Recognizer Properties as snr_sensitivity.

Range: 0 - 100 (MRCP v1) or 0.0 - 1.0 (MRCP v2)

Default: 50 (MRCP v1) or 0.5 (MRCP v2)

speed_vs_accuracy

Specifies how much accuracy is valued from an SRE decode, at the cost of speed. In MRCP v1, a value of 0 is for low speed and higher accuracy. In MRCP v2 and the LumenVox API, a value of 0 is for high speed and lower accuracy.

Range: 0 - 100 (MRCP v1) or 0.0 - 1.0 (MRCP v2)

Default: 50 (MRCP v1) or 0.5 (MRCP v2)

speech_incomplete_timeout

At this time, this parameter is unused.

no_input_timeout

After recognition has started, if there is no speech detected for a certain period of time, the recognizer will send a RECOGNITION-COMPLETE event to the client and terminate the recognition operation. The no-input-timeout parameter controls the length of this time period. The value is in milliseconds.

Default: 20000

dtmf_termination_timeout

The terminating timeout to use when recognizing DTMF input. When no DTMF input has been detected for the length of time set by this parameter, the system will stop waiting for DTMF input. The value is in milliseconds. The value for this field ranges from 0 to MAXTIMEOUT, where MAXTIMEOUT is platform specific. Refer to the VoiceXML 2.0 specification, Appendix D for detailed information.

Default: 10000

dtmf_inter_digit_timeout

This parameter specifies the inter-digit timeout value to use when recognizing DTMF input. The value is in milliseconds. The value for this field ranges from 0 to MAXTIMEOUT, where MAXTIMEOUT is platform specific. Refer to the VoiceXML 2.0 specification, Appendix D for detailed information.

Default: 5000

recognizer_start_timers

If set to false (the default), the recognizer will not start the timers until the client sends a RECOGNITION-START-TIMERS method to the recognizer.

Default: false

recognition_timeout

After recognition has started and there is no match for a certain period of time, the recognizer can send a RECOGNITION-COMPLETE event to the client and terminate the recognition operation. The recognition-timeout parameter field sets this timeout value. The value is in milliseconds. The value for this field ranges from 0 to MAXTIMEOUT, where MAXTIMEOUT is platform specific.

Default: 10000

dtmf_inter_digit_timeout

Default: 5000

save_waveform

The MRCP server can save the audio stream that was used for a recognition if the client sets this parameter to true. The server records the audio and makes it available to the client in the form of a URI returned in the waveform-uri header field in the RECOGNITION-COMPLETE event. If there is an error in recording the stream, the recognizer returns a blank waveform-uri header field.

Default: false

waveform_url_location

This parameter is used in conjunction with the save-waveform parameter. This tells the MRCP server where to put the recorded audio-stream file on the local filesystem. It is only used if the save_waveform parameter is set to true.

Default: "" (empty string)

waveform_url_prefix

Note: This option was added to Legacy Version 10.4 onwards of the LumenVox Media Server.

Specifies prefix of returned wavefile URLs. This prefix will be returned with the filename appended. The default waveform_url_location will be used if this value is not set.

Using this option may be useful when exposing these files via some http/https server, allowing you to specify a different returned URL in place of the actual location of the audio file(s) on the file system.

Default: "" (empty string)

remove_waveform_files

If enabled (1), any waveform files generated on the local filesystem will be erased at the end of the session.

This option extends the MRCP specification, which indicates that any generated waveform files should be erased.

Disabling this option (0) allows these files to persist beyond the lifespan of the session, however in this situation, the Media Server will never clean up these files, and it is therefore the responsibility of some external mechanism to maintain these files if this option is selected.

Possible Values:

0 (disabled) - files will not be erased

1 (enabled) - files will be erased at session end

Default: 1 (enabled)

Was this article helpful?