Article: AudioCodes - Configuring continuousASRTimeoutInMS for Handling User Input Pauses in Voice Responses

This Article provides information about the call control parameter continuousASRTimeoutInMS and it’s usage.

  • The Parameter “continuousASRTimeoutInMS” is applicable when the Continuous ASR feature is enabled.

  • This is applicable for only Audio Codes.

  • This Specifies the time limit for automatic speech recognition (ASR) in milliseconds and the timer activates upon receiving a recognition from the speech-to-text service.

  • Voice AI Connect will then collect all the speech-to-text recognitions and, if a period of user silence lasting as configured by this parameter is detected, it will bundle them into a single textual message sent to the bot.

  • The acceptable value range spans from 500 (equivalent to 0.5 seconds) to 60000 (equivalent to 1 minute).
    The default setting is 3000.

Note:
The value of this parameter must not exceed or be equal to the value of the “userNoInputTimeoutMS” parameter.

Example:
If the user responds in voice, the bot will acknowledge only a portion of their response and disregard the rest if there is less than a 1- or 2-second pause between the user’s sentences.

Bot: What is your problem?

User: Recently, I'm experiencing headache. <less than 1 or 2 second silence from user>. 
The symptoms are getting increased day by day.

If the user’s response contains a pause of less than 1 or 2 seconds, the bot will only take into account “Recently, I’m experiencing a headache” and exclude “The symptoms are getting increased day by day.”
Subsequently, the bot will proceed to ask the next question.

Solution:
To resolve these type of scenario’s where the bot acknowledges the part of the user input and disregards the other part of the user input we can configure the continuousASRTimeoutInMS call control parameters as shown below**.**

continuousASRTimeoutInMS - 6000

Reference Screenshot:

Please Click Here, for the list Audio Codes Call control Parameters.

Hello Srujan
How are you?
i have question in this case, how to make this pramiter used in all node that take input from user like (confirm node and entity node).
i got issue with punctuation mark so i used this ( sttDisablePunctuation ) but my problem is i need to put it in all Call Control Parameters in every single node Is there a way to use it by default?

thank you for your response

Hi @r.hassan ,

I’m doing well, and I hope you are too.
We’re currently in the process of verifying this requirement with the relevant team, and we’ll get back to you with the details as soon as we can.

Thank you,
Srujan Madderla
Kore.ai Community Team

@r.hassan,

Having consulted with our team regarding your requirement, we confirm that, by default, this will not be achievable, and you will need to incorporate in all the nodes.

Thank you,
Srujan Madderla
Kore.ai Community Team

@srujan.madderla Do we have an equivalent of this for SAVG? Please share more information on using continuousASRTimeout and userNoInputTimeout in SAVG

Hi @srinath ,

Welcome to Kore.ai Community.
We would like to inform you that continuousASRTimeoutInMS works in SAVG, we don’t need to add the continuousASR = true in the call control parameters.

userNoInputTimeout also works in SAVG, but the time we will configure for this parameter needs to be equal or greater than continuousASRTimeoutInMS.

Thank you,
Srujan Madderla
Kore.ai Community Team

@srujan.madderla I was not able to get the userNoInputTimeoutMS work. The maximum time i was able to have the bot wait in SAVG is 60 secs which is the max we can select in voice call properties. I want to extent it to 3 minutes using call control parameters. Tried using NoInputTimeout in combination with other properties line continuousASR, NoInputSpeech, NoInputRetries and nothing worked. The value in voice call properties → timeout prompts seems to be taking precedence always