SpeechToText - Objective-C - Can I get the Word level pronunciation assessment result

[wing199901]

Is your feature request related to a problem? Please describe.
I am not finding a way to get word level pronunciation assessment result with the Objective-C SDK, and it doesn't appear to be covered in the sample.

[pankopon]

Hi, please take a look at the how to use pronunciation assessment page. Parameter granularity:SPXPronunciationAssessmentGranularity_Phoneme in the example should already produce scores on the word level, though you can also use granularity:SPXPronunciationAssessmentGranularity_Word. (See configuration parameters for details.)

[wing199901]

Thanks very much for the answer. I want to call the Word level pronunciation assessment result like this (https://docs.microsoft.com/en-us/dotnet/api/microsoft.cognitiveservices.speech.pronunciationassessment.pronunciationassessmentresult.words?view=azure-dotnet#microsoft-cognitiveservices-speech-pronunciationassessment-pronunciationassessmentresult-words).
How can I do this in Objective-C?

[pankopon]

Right - based on https://docs.microsoft.com/objectivec/cognitive-services/speech/spxpronunciationassessmentresult SPXPronunciationAssessmentResult seems to lack a property for words, while it exists in C#. So there's just the JSON result with word level scores.
@yulin-li Can you please confirm the above & is there a plan to add the Words property to SPXPronunciationAssessmentResult in Obj-C?

[pankopon]

@wing199901 Meanwhile, you may read the JSON result string like

    [speechRecognizer addRecognizedEventHandler: ^ (SPXSpeechRecognizer *recognizer, SPXSpeechRecognitionEventArgs *eventArgs) {
        NSString *jsonResult = [eventArgs.result.properties getPropertyById:SPXSpeechServiceResponseJsonResult];
        NSLog(@"Received JSON: %@", jsonResult);
        // insert JSON parsing as appropriate...
    }];

The structure of the JSON result is as in the sample response.

[yulin-li]

Hi @wing199901 thx for using Pronunciation Assessment and reaching out. Currently out objc SDK doesn't parse the json to object. I think you can parse it as @pankopon suggested.
We heard your requirement and added this work to backlog. I will let you know when it finished.

[yulin-li]

Hi @wing199901 We will implement this feature and release it with next version (1.21).

We tracked this internally (id 4094454), so I will close the issue here. Fell free to reopen it for any concerns

cc @pankopon

[dargilco]

Speech SDK 1.21 was release. See release notes: https://docs.microsoft.com/azure/cognitive-services/speech-service/releasenotes?tabs=speech-sdk#speech-sdk-1210-april-2022-release