Plum IVR Support

We have a plum voice server. It seems like it has voice detection problem. It can't even recognize the word 'YES' for input.
And also it is taking a long pause after taking the input.
Do we need to do anything with the server like changing some configuration setup?
If we need to change the configuration, I can send the email with my CustomerID to plum support.

here is the sample code:
<field name="form1">

<grammar type="application/x-jsgf" mode="voice">
^(yes|no|yeah|sure|S)$
</grammar>

<prompt>

<sentence>You entered some number</sentence>

<sentence>To confirm this , say yes. </sentence>
<sentence>To cancel this , say No.</sentence>

</prompt>

</field>

Thanks in advance.

Hi,

We've just tested your sample IVR code and it works fine for us:
Can you post an IVR call log where this IVR issue occurs and provide us with more details on how to produce this IVR error?

Regards,
Plum Support

Hello,

It can recognize the voice only if we have the receiver near by our mouth.
If we talk on the speaker phone or having receiver far from our voice or talking loud and also in the noisy environment it can not get the input. sometimes we have to say that two or three times to get it work.

I have
<property name="confidencelevel" value="0.5"/> <property name="incompletetimeout" value="2s"/> <property name="recordcall" value="true"/>
<property name="sensitivity" value="0.5"/>
<property name="interdigittimeout" value="300ms"/>
<property name="timeout" value="3s"/>
<property name="inputmodes" value="dtmf voice"/>

Do these setup to do anything with that?

I provide a part of the call log when it occurs.Please let me know if
this is not enough to debug.


VXI::queue_prompts()
bargein set to true
INPUTMODES set to "DTMF VOICE"
Audio segment added to prompt queue from TTS application/synthesis+ssml for:
---------
<?xml version='1.0'?><speak><voice name="Donna" gender="female"><s><speak> You are paying <say-as type="currency"> $400.00</say-as></speak></s><s><speak> with Checking Account ending in 56</speak></s>
on <say-as type="date"> 12/02/2008 .</say-as><s><speak> To confirm this payment, say yes. </speak></s><s><speak> To cancel this payment, say No.</speak></s></voice></speak>
---------
VXI::field_element - activating grammars for form = 'getConfirm' formitem = 'confirm'
VXI::do_recognition()
PromptManager::Play()
Newly queued prompts are now being played


Tue 12 Feb 2008 10:41:22 AM EST:

Found grammar match
hypothesis #0: yes (0.8647)
hypothesis #1: S (0.4496)
bargein set to true
INPUTMODES set to "DTMF VOICE"
Audio segment added to prompt queue from TTS application/synthesis+ssml for:
---------
<?xml version='1.0'?><speak><voice name="Donna" gender="female">Please wait while we process your request.</voice></speak>
---------
Newly queued prompts are now being played

Thank you.

Hi,

You could try lowering the sensitivity to a value like 0.3 to account for the noisy environment. You could also make all of your prompts non-bargeable to prevent <nomatch> messages from occurring.
Hope this IVR code helps.

Regards,
Plum Support

Thank you for the quick response. I will try that.