What servers allow Long Audio API for text-to-speech with Aria vocaloid on Azure?

DDS 1

DDS 1 Reputation point

2020-10-07T21:57:07.94+00:00

I am getting the following message:

{"code":"InvalidRequest","details":null,"message":"The voice synthesis API is not enabled in this region.","target":null,"innerError":{"code":"InvalidPayload","details":null,"message":"The voice synthesis API is not enabled in this region.","target":null,"innerError":null}}

wheh I run my voiceclient.py code from the following Github:

https://github.com/Azure-Samples/Cognitive-Speech-TTS/blob/master/CustomVoice-API-Samples/Python/voiceclient.py
GiftA-MSFT 11,151 Reputation points

2020-10-07T23:00:55.317+00:00

Hi, thanks for reaching out. Based on the error message, you need to change the speech subscription key with a supported region. Please check out this document for Long Audio API supported regions. The Long audio API is available in multiple regions with unique endpoints. Hope this helps!
DDS 1 Reputation point

2020-10-08T22:29:40.223+00:00

I tried the correct parameters for the voiceclient.py but it doesn't work. What is going on? Could you please spell out each parameter in the program so I don't have to debug the code? I am prepared to do that but this could be avoided.
DDS 1 Reputation point

2020-10-08T22:45:07.213+00:00

Here's the code example for voiceclient.py:

https://github.com/Azure-Samples/Cognitive-Speech-TTS/blob/master/CustomVoice-API-Samples/Python/voiceclient.py
GiftA-MSFT 11,151 Reputation points

2020-10-09T20:54:07.977+00:00

What is the region of your resource?
DDS 1 Reputation point

2020-10-10T06:24:18.683+00:00

I picked one of them and it still doesn't work.
DDS 1 Reputation point

2020-10-10T06:25:20.137+00:00

East US https://eastus.customvoice.api.speech.microsoft.com
South Central US https://southcentralus.customvoice.api.speech.microsoft.com
West US 2 https://westus2.customvoice.api.speech.microsoft.com

I tried all of these and they didn't work with the right voice IDs. What is going on with MSFT?
GiftA-MSFT 11,151 Reputation points

2020-10-12T21:25:21.42+00:00

Sorry for the inconvenience, we are looking into this and will get back to you shortly. Thanks.

1 answer

GiftA-MSFT 11,151 Reputation points

2020-10-13T17:53:01.217+00:00
It seems to be working on my end. Please follow these steps:

These are the supported regions

In this example, I will be using eastus.

Run the following command to get available voice list

python voiceclient.py --voices -region eastus -key xxxxxxxxxx

I will be using one of the voices returned from previous command

Name: zh-CN-XiaoyouNeural, Description: {"Version":"20200520111712"}, Id: 89262f4e-76b6-41db-96fc-f9151e0954f6, Locale: zh-CN, Gender: Male, PublicVoice: True, Created: 2020-05-20T11:43:09Z

Then submit synthesis request:

python voiceclient.py --submit -region eastus -key xxxxxxxxxx -file en-US.txt -locale zh-CN -voiceId 89262f4e-76b6-41db-96fc-f9151e0954f6 -format riff-16khz-16bit-mono-pcm --concatenateResult

Also, please refer to the Long Audio API documentation and sample code usage details. Hope this helps!
Please sign in to rate this answer.

0 comments No comments
Sign in to comment

Share via

What servers allow Long Audio API for text-to-speech with Aria vocaloid on Azure?

1 answer