P2A with Azure speech possible ?

[Aikadil 1]

Hi, on the topic on amazon AWS Polly, do you think this is possible with Microsoft azure neural speech service?

the topic: https://www.avsim.com/forums/topic/605552-amazon-polly-voices-with-p2a-issues/

reading part of the documentation, I understand that the equivalent of the polly plugin for windows is missing.

the doc: https://docs.microsoft.com/en-us/azure/cognitive-services/Speech-Service/

But it seems possible but requires knowledge that I do not have at all....

if a member of the community has some resource, it might work better, especially since azure is also able to transcribe voice to text and also to make real time translations... enough to push P2A in the future....

[krakosia 0]

Very good question, if possible it would be a great option

[neilhewitt 493]

The 'right' way to do it would be to integrate direct support for the Azure TTS API into P2ATC, rather than using what amounts to a proxy to use Azure TTS as if it were a SAPI voice. That's what the Polly for Windows plugin does. But that would be a lot of work for the dev team when the entire program is built to use SAPI voices and always has been. 

Could someone in the community write an Azure TTS-to-SAPI plugin? Yes, certainly. But then you have the same support issues that we do with Polly. The P2ATC team couldn't be responsible for that plugin and if it became unsupported as the AWS plugin seems to have become, bugs will not get fixed. The strife trying to use Polly with P2ATC at the moment is painful for many (oddly, mine has become much more stable of late). 

And of course, like Polly, use of Azure TTS is not free. End-users would have to pay for it. It's not hugely costly but it's a non-zero cost to continue using your ATC software. I guess there would need to be a huge groundswell of P2ATC users asking for Azure TTS to be added to the software to make it worth the effort.