[Azure Cognitive Services Connector] Bing Speech API - connection time out

[Azure Cognitive Services Connector] Bing Speech API - connection time out

  
Forge Component
(13)
Published on 26 Sep (3 weeks ago) by Pedro Oliveira
13 votes
Published on 26 Sep (3 weeks ago) by Pedro Oliveira

Hi guys, 

Im using bing speech api of AzureCognitiveServiceConnector to convert audio to text. while using the 'ConversionToText' Action, i'm getting 408 request time out Error(>14000 ms). Please help me here

Hi Jesudoss,

Unfortunately that API doesn't work. It requires files to be uploaded in chunked mode and we haven't solved that problem yet.

Sorry,

Daniel

Hi Daniel,

          1. could you please explain the problem , uploading the files in chunked mode? 

          2. how can i upload my file in chunked mode?

             2.1 Will it work if i give "chunked" as value for the input parameter "TransferEncoding" of                                  "ConversionRecognition"  Restful API.


Thanks

Jesudoss J

Hi Jesudoss,

The API documentation specifies the speech to text service requires HTTP chunked-transfer encoding to send the audio data.

At the time I tried to solve the problem by manipulating the request headers but I had no success.

A colleague suggested I write an extension to set the HttpWebRequest.SendChunked property before sending the request. I've been meaning to try this but I haven't been able to do it yet.

If you'd like to try this approach and don't know how, check the Advanced Request Customizations documentation.

Hope this helps.

Daniel

Hi Daniel,

 I have written an extension with HttpWebRequest.SendChunked property as "True". I'm getting "Chunked encoding must be set via the SendChunked Property. Parameter name: Value".


I have attached the screenshot of the error.Can you please help here.....


Regards

Jesudoss

Hi Jesudoss,

Unfortunately I can't gain much information from the error text.

Like I said, we abandoned the implementation of that API at the time and haven't had a chance to get back to it since then, so you're in uncharted territory.

If you share your code I may be able to spot something but again, I haven't tried this approach myself.

Cheers

Hi Jesudoss and Daniel: we have an updated version of the Azure cognitive services connector launching soon --- with a working sample of of the Bing speech to text and text to speech APIs. You can try it out here: https://events-dev.outsystemsenterprise.com/AzureCognitiveServicesSample/ (which may change soon).

Andy Pemberton wrote:

Hi Jesudoss and Daniel: we have an updated version of the Azure cognitive services connector launching soon --- with a working sample of of the Bing speech to text and text to speech APIs. You can try it out here: https://events-dev.outsystemsenterprise.com/AzureCognitiveServicesSample/ (which may change soon).

New URL is: https://wwdtp10-dev.outsystemsenterprise.com/AzureCognitiveServicesSample/