Share via

Difference in evaluation of Audio in Cognitive Pronunciation Assessment tool and same python SDK

jayenthiran 0 Reputation points
2023-09-21T04:51:04.5833333+00:00

I encountered a technical issue while testing the Pronunciation Assessment Python SDK sample. I used an audio file (temple.wav) that contains the word "temple" pronounced in US English dialect.

When I tested the same audio in your Speech Studio Pronunciation Assessment tool, it provided a result of 100 marks in JSON format: [https://speech.microsoft.com/portal/pronunciationassessmenttool].

However, when I used your Pronunciation Assessment Python SDK with the same audio, it returned very low marks for the identical audio with the US dialect. You can find the SDK here: [https://github.com/Azure-Samples/Cognitive-Speech-TTS/tree/master/PronunciationAssessment/Python].

I have the following queries:

  1. Why is there a different evaluation result for the same audio in the same dialect (en-us)?

Please find the attached text document with both JSON results for your reference.ResultJson-Onlinetool.txt

Azure AI Language
Azure AI Language
An Azure service that provides natural language capabilities including sentiment analysis, entity extraction, and automated question answering.
523 questions

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.