Properties

confidencelevel

Specifies the speech recognition confidence level.  Values range from 0.0 (minimum confidence) to 1.0 (maximum confidence).  Recognition results are rejected (a nomatch event is thrown) if the confidence level of the results is below this threshold.

0.5

sensitivity

Specifies the level of sensitivity to speech.  Values range from 0.0 (least sensitive to noise) to 1.0 (highly sensitive to quiet input).

0.5

speedvsaccuracy

A hint specifying the desired balance between speed vs. accuracy when processing a given utterance.  Values range from 0.0 (fastest recognition) to 1.0 (best accuracy).

Note: The Nuance MRCP engine uses the value of the speedvsaccuracy property to set its proprietary rec.Pruning parameter, using the following algorithm:

If x is the speedvsaccuracy value, and x <= 0.5 then
   rec.Pruning = (x * 400) + 600
else
   rec.Pruning = (x * 800) + 400

0.5

completetimeout

The length of silence required following user speech before the speech recognizer finalizes a result (either accepting it or throwing a nomatch event).  The completetimeout is used when the speech is a complete match of an active grammar and no further words can be spoken.

1s

incompletetimeout

The length of silence required following user speech before the speech recognizer finalizes a result (either accepting it or throwing a nomatch event).  In contrast to completetimeout, the incompletetimeout is used when the speech is an incomplete match to an active grammar, or when the speech is a match but it is possible to speak further.

1s

maxspeechtimeout

The maximum duration of user speech.  If this time elapsed before the user stops speaking, the maxspeechtimeout event is thrown.  Note that with Nuance 8.0, maxspeechtimeout must be more than incompletetimeout.  Otherwise, maxspeechtimeout will be disabled.

• SpeechWorks allows values in {1ms,99999ms}
• Nuance allows values in {0s,600s}
• For Watson, 0 means no limit

60s

maxnbest

Maximum number of results returned by the recognizer.  Also represents the maximum size of the application.lastresult$ array.

1

interdigittimeout

The timeout period allowed between each digit when recognizing DTMF input.

3s

termtimeout

The terminating timeout to use when recognizing DTMF input.

0s

termchar

The terminating DTMF character for DTMF input recognition.

#

inputmodes

Determines which input methods to use. Value is a space separated list of input methods:

• dtmf - allows DTMF sequences as input
• voice - allows voice as input

dtmf voice

timeout

Once the prompt has finished playing, the length of time to wait, if no speech or dtmf input occurs, before throwing a noinput event.

10s

universals

Specifies universal command grammars to activate.  Value is a space-separated list of all or fewer of the following command grammars:

• cancel - If this grammar is activated, and the caller says "cancel" (or equivalent phrase configured for another language), the cancel event is thrown.
• exit - If this grammar is activated, and the caller says "exit" (or equivalent phrase configured for another language), the exit event is thrown.
• help - If this grammar is activated, and the caller says "help" (or equivalent phrase configured for another language), the help event is thrown.

none

asrengine

(VoiceGenie extension)

Specifies the name of the ASR (Automatic Speech Recognition) engine to use.  The name is configured by the ASR Manager or SRM Server.

It is valid to specify a particular engine only if that engine is installed for the platform running the application.  Otherwise, an error.asr.unknownengine event will be thrown.

Note: The configured name for SpeechWorks OSR must be "SPEECHWORKS", otherwise a recognition error will occur.

platform-specific

ttsengine

(VoiceGenie extension)

Specifies the name of the TTS (Text-to-Speech) engine to use (i.e. the voice).  The name is configured by the TTS Manager or SRM Server.

It is valid to specify a particular engine only if that engine is installed for the platform running the application.  Otherwise, an error.tts.unknownengine event will be thrown.

platform-specific

com.voicegenie.grammartype.<asrengine>

For example:

• com.voicegenie.grammartype.speechworks
• com.voicegenie.grammartype.nuance
• com.voicegenie.grammartype.watson

(VoiceGenie extension)

Changes the grammar type to be used as the inline default when using <asrengine>, so that the specified grammar type can be used without including the type attribute.  It is valid to specify a particular grammar type only if the <asrengine> supports that grammar type.

platform-specific

com.voicegenie.endbeep

(VoiceGenie extension)

Specifies whether or not a beep should be played at the end of prompts in fields, when bargein is disabled. When bargein is enabled, this attribute has no effect (there is never a beep). Platform owners can access the audio file at /usr/local/phoneweb/audio/effects/ endofprompt.vox or C:\VoiceGenie\mp\audio\ effects\endofprompt.vox (the location is specified by the beepaudio parameter in the VoiceXML Interpreter configuration).

false

com.voicegenie.utterancedest

(VoiceGenie extension)

Specifies the path of the directory to use for saved utterance audio files. The value will be resolved relative to the /usr/local/phoneweb/utterance/ or C:\VoiceGenie\mp\utterance\ directory. Can be used with either the com.voicegenie.saveutterance or recordutterance property.

Note that if you specify the utterancedest AND enable the "savetmpfiles" property, the utterance will ONLY be saved under the utterancedest path; it will not also be saved with the other tmp files.

files are written to the tmp directory (may or may not be saved, depending on whether the "savetmpfiles" property is enabled)

recordutterance

(VoiceXML 2.1 feature)

This property tells the platform to enable recording while simultaneously gathering input from the user.  Set to true to enable user utterance to be recorded.  Set to false otherwise.  Upon completion of user input, the recording shadow variable will be set.  Note that the <vxml> version attribute must be specified as 2.1 (or higher) to use this property.

false

recordutterancetype

(VoiceXML 2.1 feature)

This property specifies the audio format to use for recording utterances.  Only used with the recordutterance property.  VoiceGenie currently supports the following types:

• audio/basic - Raw (headerless) 8kHz 8-bit mono mu-law [PCM] single channel. (G.711)
• audio/x-alaw-basic - Raw (headerless) 8kHz 8-bit mono A-law [PCM] single channel. (G.711)
• audio/x-wav - WAV (RIFF header) 8kHz 8-bit mono mu-law [PCM] single channel.
• audio/x-wav - WAV (RIFF header) 8kHz 8-bit mono A-law [PCM] single channel.

audio/basic

com.voicegenie.asr.get_swi_literaltimings

(VoiceGenie extension)

Set to true to allow the special OSR variable, SWI_literalTimings, to be accessed through the application.lastresult$ variable.  Requires com.voicegenie.fieldobject to be set to true.  Available with SpeechWorks ASR only.

false

COM.VOICEGENIE.ASR.<vendor_specific_name>

(VoiceGenie extension)

When using VoiceGenie's MRCP direct integration with an ASR engine, the VoiceXML application can use this property format to specify arbitrary vendor-specific parameters to be sent to the ASR engine.

In the property name, <vendor_specific_name> is replaced with the actual vendor-specific parameter name; and the value of the property must be a valid value for that vendor-specific parameter.  For example, to set Nuance's rec.GrammarWeight parameter to 10:

<property
name="COM.VOICEGENIE.ASR.rec.GrammarWeight"
value="10"/>

Notes:

• Vendor parameter names and values could be case-sensitive. Refer to the vendor documentation to ensure you are using valid names/values.
• You can only set a vendor parameter via <property> if the parameter can be set by the ASR engine at runtime. Refer to the vendor documentation to confirm which parameters are runtime-settable.
• Once a vendor parameter is set via <property>, the setting will stay in effect for the remainder of the call, unless it is set again later in the VoiceXML application.

parameter-specific

swiep_*/swirec_*

(VoiceGenie extension)

Many of OSR's swiep_*/swirec_* configuration parameters can also be set as VoiceXML properties.

To find out whether a particular parameter can be set as a property, look it up in the OSR Reference Manual.  If the line under the parameter name includes "API" (and if the description mentions SWIepSetParameter() or SWIrecRecognizerSetParameter()), then it can be set as a property.

Some of the parameters that are commonly used are:

• swirec_suppress_event_logging
• swirec_suppress_waveform_logging
• swirec_audio_environment (OSR 2.0+ only)
• swirec_backward_compatible_confidence_scores (OSR 2.0+ only)

See the OSR Reference Manual for details about the values/usage for each parameter.

These properties are specific to Nuance OSR, and are only supported in VoiceGenie's MRCP native integration with OSR.  (They are not supported in VoiceGenie's MRCP direct integration with OSR, via SWMS.)

parameter-specific

com.voicegenie.logtoasr

(VoiceGenie extension)

If set to true, this will enable VoiceGenie to log data directly to the ASR engine's log.

true

com.att.asr.useActive

(VoiceGenie extension)

Set to false to support large directory/grammar enhancement to the Watson engine.  (Available with Watson ASR only.)

false

com.att.asr.*

(VoiceGenie extension)

When using the Watson ASR engine, if there are VoiceXML properties specified that start with the prefix com.att.asr, VoiceGenie will pass those properties as configuration parameters to the Watson engine, along with the corresponding values specified in the properties.

(Available with Watson ASR only.)

parameter-specific

bargein

Controls whether or not user input can be collected before prompts have finished playing:

• true - Any user input can barge in during prompts.
• false - No user input can barge in during prompts.

true

bargeintype

Specifies the bargein type:

• speech - Any user utterance can barge in the prompt.
• hotword (equivalent to recognition) - Only user input that matches a grammar can barge in on the prompt.

Note: Not all bargeintypes are supported with all ASR engines.

speech

com.voicegenie.recognitionbargeinminimum

(VoiceGenie extension)

Minumum duration of the bargein phrase if bargeintype is set to recognition or hotword. (Recommended range is 0s-2s.)

Note: This property is not supported with all of the ASR engines that support hotword bargein.

0.3s

com.voicegenie.recognitionbargeinmaximum

(VoiceGenie extension)

Maximum duration of the bargein phrase if bargeintype is set to recognition or hotword. (Recommended range is 0.5s-10s.)

Note: This property is not supported with all of the ASR engines that support hotword bargein.

1.7s

com.voicegenie.wakeupword

(VoiceGenie extension)

Specifies whether or not Wakeup Word Spotting should be used for input in fields, menus, and initials.  If set to true, recognition is only performed if input length is between a minimum and maximum length, and (only with Nuance OSR 2.0+) if input matches a grammar.

false

com.voicegenie.wakeupwordminimum

(VoiceGenie extension)

If com.voicegenie.wakeupword is set to true, this specifies the minimum length that input must be in order for recognition to be performed.

com.voicegenie.wakeupwordmaximum

(VoiceGenie extension)

If com.voicegenie.wakeupword is set to true, this specifies the maximum length that input may be in order for recognition to be performed.

COM.VOICEGENIE.ASR.Recognition-Mode

(VoiceGenie extension)

Set to 'hotword' to enable the OSR selective barge-in or magic word recognition mode:

• Selective Barge-in: Only user input that matches a grammar can barge in on the prompt.  (This mode is enabled if COM.VOICEGENIE.ASR.Hotword-Max-Duration is set to 0.) 
• Magic Word: Only user input that matches a grammar, and whose duration is between a minimum and maximum length, can barge in on the prompt.  (The minimum and maximum utterance lengths are specified by COM.VOICEGENIE.ASR.Hotword-Min-Duration and COM.VOICEGENIE.ASR.Hotword-Max-Duration.)

For example:

<property
name="COM.VOICEGENIE.ASR.Recognition-Mode"
value="&quot;hotword&quot;"/>

Note: After setting this property, the specified mode will remain in effect for all subsequent recognitions (even if the property is not set in subsequent input fields), unless a new mode is explicitly set. So, to switch back to normal recognition mode after using one of the above hotword modes, the application must explicitly set this property back to 'normal' (and not set any of the three related properties listed below). For example:

<property
name="COM.VOICEGENIE.ASR.Recognition-Mode"
value="&quot;normal&quot;"/>

(Available with Nuance SWMS 3.1.4+ only.)

COM.VOICEGENIE.ASR.Hotword-Min-Duration

(VoiceGenie extension)

If COM.VOICEGENIE.ASR.Recognition-Mode is set to hotword, this specifies the minimum length (in ms) that input must be in order for recognition to be performed.  For example:

<property
name="COM.VOICEGENIE.ASR.Hotword-Min-Duration" 
value="&quot;50&quot;"/>

COM.VOICEGENIE.ASR.Hotword-Max-Duration

(VoiceGenie extension)

If COM.VOICEGENIE.ASR.Recognition-Mode is set to hotword, this specifies the maximum length (in ms) that input may be in order for recognition to be performed.  For example:

<property
name="COM.VOICEGENIE.ASR.Hotword-Max-Duration" 
value="&quot;2000&quot;"/>

COM.VOICEGENIE.ASR.Hotword-Confidence-Threshold

(VoiceGenie extension)

If COM.VOICEGENIE.ASR.Recognition-Mode is set to hotword, this specifies the speech recognition confidence level that should be used. Values range from 0 (minimum confidence) to 1000 (maximum confidence). Recognition results are rejected (a nomatch event is thrown) if the confidence level of the results is below this threshold.

For this property to take effect, you must also set the standard confidencelevel property to an equivalent decimal percentage. For example:

<property
name="COM.VOICEGENIE.ASR.Hotword-Confidence-Threshold" 
value="&quot;100&quot;"/>
<property name="confidencelevel" value="0.1"/>

audiofetchhint

Defines when audio files can be fetched:

• prefetch - audio file may be downloaded when the page is loaded
• safe - only load the audio file when needed

prefetch

audiomaxage

Defines maximum acceptable age, in seconds, of cached audio resources.

undefined

audiomaxstale

Defines maximum staleness, in seconds, of expired cached audio resources.

undefined

datafetchhint

Defines when XML data files can be fetched:

• safe - only load the XML data file when needed

safe

datamaxage

Defines maximum acceptable age, in seconds, of cached XML resources.

undefined

datamaxstale

Defines maximum staleness, in seconds, of expired cached XML resources.

undefined

documentfetchhint

Defines when next document can be fetched:

• safe - only load the next document when needed

safe

documentmaxage

Defines maximum acceptable age, in seconds, of cached documents.

undefined

documentmaxstale

Defines maximum staleness, in seconds, of expired cached documents.

undefined

grammarfetchhint

Defines when grammar files can be fetched:

• prefetch - grammar file may be downloaded when the page is loaded
• safe - only load the grammar file when needed

prefetch

grammarmaxage

Defines maximum acceptable age, in seconds, of cached grammar resources.

SpeechWorks OSR 1.x does not support this.

undefined

grammarmaxstale

Defines maximum staleness, in seconds, of expired cached grammar resources.

SpeechWorks OSR 1.x does not support this.

undefined

objectfetchhint

Defines when objects can be fetched:

• prefetch - object may be downloaded when the page is loaded
• safe - only load the object when needed

prefetch

objectmaxage

Defines maximum acceptable age, in seconds, of cached object resources.

undefined

objectmaxstale

Defines maximum staleness, in seconds, of expired cached object resources.

undefined

scriptfetchhint

Defines when scripts can be fetched:

• prefetch - script may be downloaded when the page is loaded
• safe - only load the script when needed

prefetch

scriptmaxage

Defines maximum acceptable age, in seconds, of cached script resources.

undefined

scriptmaxstale

Defines maximum staleness, in seconds, of expired cached script resources.

undefined

fetchaudio

The URI of audio to play while waiting for documents to be fetched.

builtin:background_audio.wav

fetchaudiodelay

The length of time to wait at the start of a fetch delay before playing fetchaudio.

1s

fetchaudiominimum

The minimum length of time to play fetchaudio, once started, even if the fetch result arrives in the meantime.

0s

fetchtimeout

Timeout for fetches.  This is not supported when using Nuance(MRCP).  An error.badfetch is thrown when a fetch duration exceeds fetchtimeout.

30s

com.voicegenie.noaudiocontrol

(VoiceGenie extension)

If this property is set (to any value), the com.voicegenie.audiocontrol property is disabled.

undefined

com.voicegenie.audiocontrol

(VoiceGenie extension)

(Only used if com.voicegenie.noaudiocontrol is undefined.) Set to true to enable Audio Control during playing of audio. Set to false to disable the feature.

true

com.voicegenie.audio.skipduration

(VoiceGenie extension)

Sets the duration of audio to be skipped when using the skipahead/skipback features.  The default unit is milliseconds.

6000ms

com.voicegenie.audio.skipahead

(VoiceGenie extension)

Sets the DTMF button for skipping ahead in the audio file/TTS. The duration skipped depends on the value of the com.voicegenie.audio.skipduration property. If set to "-" or undefined, this feature is disabled.

undefined

com.voicegenie.audio.skipback

(VoiceGenie extension)

Sets the DTMF button for rewinding the audio file/TTS. The duration rewound depends on the value of the com.voicegenie.audio.skipduration property. If set to "-" or undefined, this feature is disabled.

undefined

com.voicegenie.audio.louder

(VoiceGenie extension)

Sets the DTMF button for turning volume up. If set to "-" or undefined, this feature is disabled.

This is not supported with VoIP.

undefined

com.voicegenie.audio.softer

(VoiceGenie extension)

Sets the DTMF button for turning volume down. If set to "-" or undefined, this feature is disabled.

This is not supported with VoIP.

undefined

com.voicegenie.audio.pause

(VoiceGenie extension)

Sets the DTMF button for pausing playback temporarily, until the pause button is pressed a second time. If set to "-" or undefined, this feature is disabled.

undefined

com.voicegenie.audio.stop

(VoiceGenie extension)

Sets the DTMF button for stopping all queued audio playback. If set to "-" or undefined, this feature is disabled.

undefined

com.voicegenie.audio.next

(VoiceGenie extension)

Sets the DTMF button for interrupting the current audio playback, and starting the next audio playback in the queue. If set to "-" or undefined, this feature is disabled.

undefined

com.voicegenie.audio.faster

(VoiceGenie extension)

Sets the DTMF button for increasing the rate of audio playback. If set to "-" or undefined, this feature is disabled.

This is not supported with VoIP.

undefined

com.voicegenie.audio.slower

(VoiceGenie extension)

Sets the DTMF button for decreasing the rate of audio playback. If set to "-" or undefined, this feature is disabled.

This is not supported with VoIP.

undefined

com.voicegenie.loglevel

(VoiceGenie extension)

The loglevel limits execution of <log> elements to the ones whose level attribute have a value up to (including) the loglevel value.

1

com.voicegenie.metricslevel

Controls the amount of detail logged to the metrics file (pw_metricsfile) on the platform.  Here's the general type of metrics information at each level:

• 0 - inbound/outbound call states, and <log> content (see note below)
• 1 - level 0 detail + page compilation and application name
• 2 - level 1 detail + application end, prompt ends, and errors
• 3 - level 2 detail + document fetching, application traversal/states/events, prompts, and warnings
• 4 - level 3 detail + resource fetching (other than VoiceXML pages)
• 5 - level 4 detail + variable/script/expression evaluation

See the System Reference Guide for complete details.  Note that if <log>s are executed (which depend on the loglevel), their content is only written to the metrics file if specified by their dest attribute.

1

com.voicegenie.maintainer.sendwhen

This property indicates if the maintainer email message should be sent.  Valid values are: always, never, and on_message.

on_message

com.voicegenie.savetmpfiles

The value is interpreted as a string with a list of words.  The words may be: all, none, prompts, inputs, pages, recordings.  When a list of keywords is specified, the superset of all the keywords are saved.  In particular, this means if someone specifies <property name= "com.voicegenie.savetmpfiles"  value="none inputs" /> it is equivalent to specifying <property name= "com.voicegenie.savetmpfiles"  value="inputs"/>.  Refer to the Debugging Tutorial for more information.

none

com.voicegenie.savetmpfilesmode

This property can take on one of two values: immediate or delayed.  This property only takes effect when com.voicegenie.savetmpfiles is enabled.  If set to immediate the files are written to disk immediately.  If set to delayed the files are stored in memory.  Refer to the Debugging Tutorial for more information.

immediate

com.voicegenie.onexit.keeptmpfiles

This property specifies whether or not keep temp files around after the VoiceXML session has ended.  This property will only have meaning if at least one temp files has been saved.  If this value is false, all temp files on the disk will be erased, and any files in memory will be discarded.  If this value is true, all temp files on disk will be kept, and files in memory will be flushed to disk.  Refer to the Debugging Tutorial for more information.

true

com.voicegenie.maxrecordtime

Defines the default (also the upper limit) for the maxtime attribute of the <record> element.

10 minutes