I’m new to using Azure Cognitive Services (and programming in general).
The Azure Text to Speech service is provided through two interfaces: the Speech SDK and the REST API. I was wondering when one should choose to use the SDK or REST API? What are the advantages of the SDK over the API, and vice versa?
The documentation of the Azure Speech SDK is quite vague on this. It says:
In some cases, you can't or shouldn't use the Speech SDK. In those cases, you can use REST APIs to access the Speech service. For example, use the Speech-to-text REST API v3.0 for batch transcription and custom speech.
For context, I’m looking to implement an text-to-speech application that takes a JSON of conversation strings, synthesizes each string to a speech file and saves it to cloud storage. So far I've been using the Speech SDK locally and have managed to save speech files to local storage, but I need to do this in the cloud.
Let me know if I should provide more details and thanks in advance for any insight!
Related
I would like to connect the Document Service from Java (Spring Boot application deployed in SAP BTP Cloud Foundry environment with multitenancy) and I am wondering about the intended way to connect the document service in a multitenant szenario.
Does SAP cloud sdk helps with connectivity?
If we're talking about consuming the service on the SAP Business Technology platform - then yes. Check the multi-tenancy docs.
The SAP Cloud SDK is a framework for consuming services and usually not the best choice for publishing them. You can use it as a proxy or data adapter, of course, if that's what you need.
For service publishing with multi-tenancy, I'd recommend looking into the direction of the Cloud Application Programming model (CAP)
Are you building OData, OpenAPI or generic REST enabled service? The SAP Cloud SDK can help to consume your service with multi-tenancy out of the box using our OData or OpenAPI code generators and type-safe clients. You can also use the SDK for testing. Please, check the details in our docs.
I'd also recommend checking if the service you're building already exists or you can adapt some other service with help of the SDK.
Is there a possible way to integrate the speech services with bot? So i would like to know what is the process that goes in the integration of Speech Services with Bot?
How is it possible to do the integration with bot through the key that is obtained from the Azure Speech Service created?
And how to check whether the created Speech Services is working?
Is there any common portal for the Speech Services to be tested like the portal for custom speech(https://westus.cris.ai/Home/CustomSpeech)?
And What are the different ways for integrating?
To test the speech service whether works, you can use the code sample provided by Microsoft without write addition code. Refer to the Next steps section of Text to Speech
Or you can make the REST API using POST man according the documentation.
And to integrate the Speech Service with Bot, you can refer the figure below:
For IVR application, I have integrated the google speech API. Its working fine.
But When I deploy the same in client location, there is no internet facility available.
So is it possible to use the google speech API offline in linux machines
It is not possible to use Google cloud speech api offline. Google cloud means that your app is interacting with online algorithms to process image, speech etc.
How to integrate Watson Discovery with Watson Virtual Agent? Can anyone give hints, sample code and provide documentation link? https://github.com/watson-virtual-agents/chat-widget does not contain information regarding this? Thank you very much!
Basically, you need to use a custom code with 2 APIs endpoint: Discovery and Virtual agent/ Conversation.
For now, have one example with Conversation Service and Discovery Service in Java, from IBM Developers, you can see in this link.
And a few weeks ago I built some simple example using the project conversation-simple for integrating Discovery, using the highlight and another custom params, using Node.js, you can see in this link
After creates one IBM Cloud (Bluemix), go to Catalog and Select the Watson services above, pay attention to the Services credentials when you click in your Watson Service for use each service, as you can see in my .env and IBM Professionals file, for all works, you need to paste the values into the Enviroment Variables or replace the values with the respective variables in your code.
See the Official Documentation for Using Watson Virtual Agent instead of Conversation.
See the Official Documentation and API Reference for use Watson Discovery.
See the Official Documentation and API Reference for use Watson Conversation.
is there any way to publish my API in azure mobile service API instead of using its own. I want just to know it, because assume I have a backup server in amazon and as you know I don't have access to azure sdk in there and it means technically I cant use the API anywhere outside the azure.
Azure mobile services is actually based on the .Net web api in their C# flavor and on node.js on the Javascript side. You could surely write an API using those technology and have the same behavior. You will gain the fact of being more portable, however you will lose some of the preconfigured stuff from Mobile services.
If you are using .NET, please check this to see how to build a Web api! http://www.asp.net/web-api.
If you are using Javascript, here is where you should start : https://nodejs.org/
Both tech are quite easy to learn and super powerfull, have fun! When your API is built, you could just publish them as an azure Web APP (http://azure.microsoft.com/en-us/services/app-service/web/) or an Azure API (https://azure.microsoft.com/en-us/services/app-service/api/) instead of the mobile App
Hope it helps, if you have more questions, please ask!
You can create and publish your own custom APIs in azure mobile service. You could even access in via azure mobile service sdk in client by using "InvokeApiAsync<>()" method.
https://msdn.microsoft.com/en-us/library/windows/apps/xaml/dn614130.aspx
If you want to access the api via fidder/httpclient than AZM SDK, pass the mobile service key in HTTP header as Name:'x-zumo-application' value:'application key from portal'