Rate this page:

How to Use Google v1p1beta1 Features

Since Google gives developers access to its v1p1beta1 of their Speech API, we've implemented its support as well.

New ASR Parameters

Currently, Voxengine ASR functionality supports the following features:

  • enableSeparateRecognitionPerChannel
  • alternativeLanguageCodes
  • enableWordTimeOffsets
  • enableWordConfidence
  • enableAutomaticPunctuation
  • diarizationConfig
  • metadata

To use these features, you have to set the beta parameter to true when creating an ASR instance, like this:

ASR Instance with v1p1beta1

ASR Instance with v1p1beta1

See these parameters' description in the reference.

And here is how the ASR result for our sample will look like:

2020-08-12 08:16:55 Sent event to JS onPhoneEvent with params[{
    id = fidawBgKTWujC7UbO2RqNB9Jsbx36k8DmPoXGzxu7wM;
    name = ASR.Result;
    requestId = ;
    response = {
        alternatives = [{
            confidence = 0.954676688;
            transcript = Hello internet, how are you ? ;
            words = [{
                confidence = 0.954676688;
                endTime = 1 s;
                speakerTag = 1;
                startTime = 0.100 s;
                word = Hello;
            }; {
                confidence = 0.954676688;
                endTime = 1.700 s;
                speakerTag = 1;
                startTime = 1 s;
                word = internet, ;
            }; {
                confidence = 0.954676688;
                endTime = 2.600 s;
                speakerTag = 1;
                startTime = 1.700 s;
                word = how;
            }; {
                confidence = 0.954676688;
                endTime = 2.700 s;
                speakerTag = 1;
                startTime = 2.600 s;
                word = are;
            }; {
                confidence = 0.954676688;
                endTime = 3 s;
                speakerTag = 1;
                startTime = 2.700 s;
                word = you ? ;
            };];
        };];isFinal = true;
        languageCode = en-us;
        resultEndTime = 4.450 s;
    };
    source = KdlKuFjhRUCr4F4Js7sTBU8Ysw1bmEMUib79EzYuKd0;
};]

New Result Properties

The ASREvents.result event has been updated with the resultEndTime, channelTag, and languageCode properties. You can see them in your session logs.