Pronunciation assessment : the phoneme level feedback needed in the IPA representation

Question

Pronunciation assessment : the phoneme level feedback needed in the IPA representation

Anju 31

I am using the detailed level feedback in using pronunciation Assessment .

The phoneme symbols in the response is in "American English phonemes", Is there a possibility to get the symbols as IPA : International phonetic alphabet ?

Thanks,

santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-22T11:12:48.8866667+00:00

Hi @Anju ,

Thank you for reaching out to Microsoft Q&A forum!

Pronunciation assessment is a part of Speech Service. It is possible to get the phoneme symbols in the IPA. Phonetic alphabets are used with the Speech Synthesis Markup Language (SSML) to improve the pronunciation of text to speech voices. Speech service supports the International Phonetic Alphabet (IPA)

For more info: SSML phonetic alphabets & Structured text phonetic pronunciation data

Hope this helps! Thank you.
santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-23T10:13:05.5566667+00:00

Hi @Anju ,

We haven’t heard from you on the last response and was just checking back to see if the given resolution was helpful. Thank you.
santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-24T10:36:51.2166667+00:00

Hi @Anju ,

Following up to see if the given information was helpful. Thank you!
Anju 31 Reputation points

2023-11-24T17:25:05.1466667+00:00

Hi Santosh,

Thanks for the quick response.

Got a little overwhelmed with so many links. I am not training any model, just using the REST APIs to get the pronunciation assessment of a sentence.

Structured text phonetic pronunciation data showed that we are getting response in Universal Phone Set (UPS). As per this documentation, there is UPS phoneme and its equivalent IPA representation.

Is there a way to set the IPA as default in this curl command call:

curl -X POST "https://centralindia.stt.speech.microsoft.com/speech/recognition/conversation/cognitiveservices/v1?language=en-US&format=detailed" -H "accept: application/json" -H "Ocp-Apim-Subscription-Key: <hidden>" -H "Content-Type: audio/wav; codecs=audio/pcm; samplerate=16000" -H "Pronunciation-Assessment: ewogICJSZWZlcmVuY2VUZXh0IjogIkhvdyBkbyBJIHJ1biB0aGlzIHByb2dyYW0uIiwKICAiR3JhZGluZ1N5c3RlbSI6ICJIdW5kcmVkTWFyayIsCiAgIkdyYW51bGFyaXR5IjogIkZ1bGxUZXh0IiwKICAiRGltZW5zaW9uIjogIkNvbXByZWhlbnNpdmUiCn0=" --data-binary @./en-US_0.wav

santoshkc 14,500 Microsoft External Staff Moderator

Hi @Anju ,

Thank you for your response.

I've included the below code snippet based on available information; I mention that I don't possess extensive knowledge about using the curl command.

curl -X POST "https://centralindia.stt.speech.microsoft.com/speech/recognition/conversation/cognitiveservices/v1?language=en-US&format=detailed" \ 
-H "accept: application/json" \ 
-H "Ocp-Apim-Subscription-Key: <hidden>" \ 
-H "Content-Type: audio/wav; codecs=audio/pcm; samplerate=16000" \ 
-H "Pronunciation-Assessment: ewogICJSZWZlcmVuY2VUZXh0IjogIkhvdyBkbyBJIHJ1biB0aGlzIHByb2dyYW0uIiwKICAiR3JhZGluZ1N5c3RlbSI6ICJIdW5kcmVkTWFyayIsCiAgIkdyYW51bGFyaXR5IjogIkZ1bGxUZXh0IiwKICAiRGltZW5zaW9uIjogIkNvbXByZWhlbnNpdmUiCn0=" \ 
-H "phonemeAlphabet: IPA" \ 
--data-binary @./en-US_0.wav

I hope this helps! Thank you.

Anju 31

Hi Santosh,

I used the parameter -H "phonemeAlphabet: IPA", but I did not get the IPA representation here is the sample output

{
          "Word": "this",
          "Offset": 22000000,
          "Duration": 6100000,
          "Confidence": 0.0,
          "AccuracyScore": 98.0,
          "ErrorType": "None",
          "Syllables": [
            {
              "Syllable": "dhihs",
              "Offset": 22000000,
              "Duration": 6100000,
              "AccuracyScore": 69.0
            }
          ],
          "Phonemes": [
            {
              "Phoneme": "dh",
              "Offset": 22000000,
              "Duration": 2100000,
              "AccuracyScore": 82.0
            },
            {
              "Phoneme": "ih",
              "Offset": 24200000,
              "Duration": 700000,
              "AccuracyScore": 100.0
            },
            {
              "Phoneme": "s",
              "Offset": 25000000,
              "Duration": 3100000,
              "AccuracyScore": 52.0
            }
          ]
        },

the Phoneme dh is "ð" in IPA ih is "ɪ" in IPA. So I guess it is still giving me the old phoneme representation. Thanks, Anju

santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-28T05:14:28.1+00:00

Hi @Anju ,

Sorry for the trouble you are facing. I request you to raise a support case through Azure portal.

I hope you understand. Thank you!
Anju 31 Reputation points

2023-11-28T14:05:13.2+00:00

Ok done,

Thanks for your support.

Best Wishes,

Anju

Your answer

santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-22T11:12:48.8866667+00:00

Hi @Anju ,

Thank you for reaching out to Microsoft Q&A forum!

Pronunciation assessment is a part of Speech Service. It is possible to get the phoneme symbols in the IPA. Phonetic alphabets are used with the Speech Synthesis Markup Language (SSML) to improve the pronunciation of text to speech voices. Speech service supports the International Phonetic Alphabet (IPA)

For more info: SSML phonetic alphabets & Structured text phonetic pronunciation data

Hope this helps! Thank you.
santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-23T10:13:05.5566667+00:00

Hi @Anju ,

We haven’t heard from you on the last response and was just checking back to see if the given resolution was helpful. Thank you.
santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-24T10:36:51.2166667+00:00

Hi @Anju ,

Following up to see if the given information was helpful. Thank you!
Anju 31 Reputation points

2023-11-24T17:25:05.1466667+00:00

Hi Santosh,

Thanks for the quick response.

Got a little overwhelmed with so many links. I am not training any model, just using the REST APIs to get the pronunciation assessment of a sentence.

Structured text phonetic pronunciation data showed that we are getting response in Universal Phone Set (UPS). As per this documentation, there is UPS phoneme and its equivalent IPA representation.

Is there a way to set the IPA as default in this curl command call:

curl -X POST "https://centralindia.stt.speech.microsoft.com/speech/recognition/conversation/cognitiveservices/v1?language=en-US&format=detailed" -H "accept: application/json" -H "Ocp-Apim-Subscription-Key: <hidden>" -H "Content-Type: audio/wav; codecs=audio/pcm; samplerate=16000" -H "Pronunciation-Assessment: ewogICJSZWZlcmVuY2VUZXh0IjogIkhvdyBkbyBJIHJ1biB0aGlzIHByb2dyYW0uIiwKICAiR3JhZGluZ1N5c3RlbSI6ICJIdW5kcmVkTWFyayIsCiAgIkdyYW51bGFyaXR5IjogIkZ1bGxUZXh0IiwKICAiRGltZW5zaW9uIjogIkNvbXByZWhlbnNpdmUiCn0=" --data-binary @./en-US_0.wav
santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-25T16:15:38.2366667+00:00

Hi @Anju ,

Thank you for your response.

I've included the below code snippet based on available information; I mention that I don't possess extensive knowledge about using the curl command.

curl -X POST "https://centralindia.stt.speech.microsoft.com/speech/recognition/conversation/cognitiveservices/v1?language=en-US&format=detailed" \ -H "accept: application/json" \ -H "Ocp-Apim-Subscription-Key: <hidden>" \ -H "Content-Type: audio/wav; codecs=audio/pcm; samplerate=16000" \ -H "Pronunciation-Assessment: ewogICJSZWZlcmVuY2VUZXh0IjogIkhvdyBkbyBJIHJ1biB0aGlzIHByb2dyYW0uIiwKICAiR3JhZGluZ1N5c3RlbSI6ICJIdW5kcmVkTWFyayIsCiAgIkdyYW51bGFyaXR5IjogIkZ1bGxUZXh0IiwKICAiRGltZW5zaW9uIjogIkNvbXByZWhlbnNpdmUiCn0=" \ -H "phonemeAlphabet: IPA" \ --data-binary @./en-US_0.wav

I hope this helps! Thank you.
Anju 31 Reputation points

2023-11-27T11:54:36.5233333+00:00

Hi Santosh,

I used the parameter -H "phonemeAlphabet: IPA", but I did not get the IPA representation here is the sample output

{ "Word": "this", "Offset": 22000000, "Duration": 6100000, "Confidence": 0.0, "AccuracyScore": 98.0, "ErrorType": "None", "Syllables": [ { "Syllable": "dhihs", "Offset": 22000000, "Duration": 6100000, "AccuracyScore": 69.0 } ], "Phonemes": [ { "Phoneme": "dh", "Offset": 22000000, "Duration": 2100000, "AccuracyScore": 82.0 }, { "Phoneme": "ih", "Offset": 24200000, "Duration": 700000, "AccuracyScore": 100.0 }, { "Phoneme": "s", "Offset": 25000000, "Duration": 3100000, "AccuracyScore": 52.0 } ] },

the Phoneme dh is "ð" in IPA ih is "ɪ" in IPA. So I guess it is still giving me the old phoneme representation. Thanks, Anju
santoshkc 14,500 Reputation points Microsoft External Staff Moderator

2023-11-28T05:14:28.1+00:00

Hi @Anju ,

Sorry for the trouble you are facing. I request you to raise a support case through Azure portal.

I hope you understand. Thank you!
Anju 31 Reputation points

2023-11-28T14:05:13.2+00:00

Ok done,

Thanks for your support.

Best Wishes,

Anju

Share via

Pronunciation assessment : the phoneme level feedback needed in the IPA representation

Your answer