This browser is no longer supported.
Upgrade to Microsoft Edge to take advantage of the latest features, security updates, and technical support.
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(297.6, 97%, 38%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3EVS%3C/text%3E%3C/svg%3E)
RealtimeResponse - In the response.done event - while listening to server event - I was expecting metadata field to be present as given in OPENAI realtime documentation to created out of band response. https://platform.openai.com/docs/api-reference/realtime-server-events/response/done - here there is a field called metadata where we can get metadata - when we send the client event - response.create with the metadata field. Why in Azure OpenAI - its not available?
Am using this model version -
"2024-10-01-preview"
Could you help give permalink of the specific section you are talking about? Because I trigger response.create with metadatafield, I was expecting response.done having a metadata field - but I dont see here - https://platform.openai.com/docs/api-reference/realtime-server-events/response/done
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(281.6, 88%, 37%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ESA%3C/text%3E%3C/svg%3E)
We support only queries on Azure OpenAI here not OpenAI, I can find mention of using metadata in Azure OpenAI here. Azure OpenAI .
please find the screenshot:
Could you check in OpenAI forums if it is specific to OpenAI.
Thank You.
Got it understood! Just a quick question - "input_audio_transcription": { "model": "whisper-1" }
In OpenAI documentation - language field is there - but in Azure OpenAI documentation - its not there!
![](data:image/svg+xml, %3Csvg xmlns='http://www.w3.org/2000/svg' height='64' class='font-weight-bold' style='font: 600 30.11764705882353px "SegoeUI", Arial' width='64'%3E%3Ccircle fill='hsl(166.4, 22%, 26%)' cx='32' cy='32' r='32' /%3E%3Ctext x='50%25' y='55%25' dominant-baseline='middle' text-anchor='middle' fill='%23FFF' %3ESC%3C/text%3E%3C/svg%3E)
In OpenAI's Whisper API, the language parameter allows specifying the input language for better accuracy. However, Azure OpenAI's Whisper model does not document this feature, likely relying on automatic language detection. If manual language selection is essential, consider using OpenAI's API directly.
Thank you!
Thank you for getting back to me! Does it mean to turn off input_audio_transcription - and instead connect with Azure open ai whisper model to be deployed on resource?
Yes, that's correct, if you stop using input_audio_transcription, you can switch to the Azure OpenAI Whisper model. It can convert spoken words into text and works in many languages.
Kindly refer link: whisper-overview
Thank You.
We haven’t heard from you on the last response and was just checking back to see if you have a resolution yet.
Thank You.
Welcome to Microsoft Q&A Forum, thank you for posting your query here!
Yes, this format is supported in Azure OpenAI real time tool, Please see this section of response.done from text in, audio out section.
kindly refer below link: Azure OpenAI |
Thank You.