Hey guys,

Do you have any plans on adding support to more languages to the Translation API in Cognitive Services?

Old version of your service (Bing Translator API) used to be able to detect many more languages than it can now (I already reported gaps in the documentation here: https://github.com/MicrosoftDocs/azure-docs/issues/29223). Specifically, I'd be interested in seeing the support for Albanian language being added.

We had to drop support of this language in our client's product due to the migration to the Azure Translation API V3.

Do you have any timeline for re-adding Albanian?

Best,
Kamil

Come on, I know you've got it. It would really do me a solid, help a brother out if you could include the individual phrase elements with timings and perplexity measurements inside the results json for the speech.phrase socket event. Confusion network data would also be fantastic to have.

{
"RecognitionStatus": 1,
"RecognizedPhrases": [{
"DisplayText": "Where are the speech phrase elements?",
"LexicalForm": "where are the speech phrase elements",
"SREngineConfidence": 0.894183,
"PhraseElements": [],
"Locale": "en-us",
"AudioTimeOffset": "00:00:00.7200000",
"AudioTimeLength": "00:00:02.2900000",
"InverseTextNormalizationResults": ["where are the speech phrase elements"],
"ProfanityTags": null,
"MaskedInverseTextNormalizationResults": ["where are the speech phrase elements"]
},
"ConfusionNetworkData": null
}

Currently there doesn't appear to be a good way to give feedback for any of the speech or vision APIs

Right now there's no easy way to export test cases with their steps. Our clients need test cases with steps in printable and readable format. Exporting from query after adding a column for steps adds way too many html tags in between the steps and the format is not at all readable.

Like the Bing Spech API I want an endpoint on WebSocket.
Example:
wss: //eastasia.stt.speech.microsoft.com/speech/recognition/conversation/cognitiveservices/v1

The Time Series Insights API is really powerfull but documentation (query API, syntax) contain some documentation mistakes regarding aggregates. As a developer I need a lot of time to analyze the queries and responses to get the rights fields. As a developer I would like to use a swagger api which allows me to test queries and generate clients.

Add an API that can transcribe any audio or video file input. E.g. A podcast mp3 or a video interview