enotani | 2024-12-04 13:47:07 UTC | #1
Hello,
Our customer's organization has installed the integration (Azure Speech to Text) and is using custom STT (Speech to Text). The end user's speech is requested from the flow to the custom STT, and upon receiving a response json (properties, text) from the custom STT, it ultimately replies to the end user using Genesys Enhanced TTS. The response from the custom STT contains several candidates, each consisting of the following properties and values (text) as a set.
"lexical" : text1 "display" : text2 "itn" : text3 "maskedItn": text4
Our customer tuned the custom STT's dictionary to set the converted text in the "display" property. This follows the description found at the following URL. https://learn.microsoft.com/en-us/azure/ai-services/speech-service/batch-transcription-get?pivots=rest-api "display" - "The display form of the recognized text. Added punctuation and capitalization are included." The result did not meet our customer's expectations, and the text in the "display" property was not processed on the Genesys Cloud side. It seems that the text from a different property was applied instead. Could you let us know which property of the custom STT response JSON is targeted on the Genesys Cloud side?
Best regards, Hirofumi Enotani
system | 2025-01-04 13:41:55 UTC | #2
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.
This post was migrated from the old Developer Forum.
ref: 30695