404 Voice ms-MY-Wavenet-D not found

[lweaver10]

I'm using the Google Cloud Text to Speech API, and the code works as expected for all languages I've tried so far. However, for this voice id: ms-MY-Wavenet-D, I'm getting this error in production:

{

response: { "error": { "code": 404, "message": "Voice ms-MY-Wavenet-D not found.", "status": "NOT_FOUND" } } ,

status: 404

}

The voice is listed on the website, and when I run the API call to fetch all voices, it's also listed:

{#7620
+"languageCodes": [
"ms-MY",
],
+"name": "ms-MY-Wavenet-D",
+"ssmlGender": "MALE",
+"naturalSampleRateHertz": 24000,
},

Has anyone else encountered this problem and is there anything that can be done?

[dawnberdan]

Hi @lweaver10,

Welcome to Google Cloud Community!

The error code you encounter is likely caused by a missing entity, such as a file or directory. To resolve this issue,  I recommend checking the affected file or its directory in your environment.

In addition, despite the code functioning for other voices, there could be a slight variation in how you're identifying the voice ID for ms-MY-Wavenet-D. Review your code for any typos or inconsistencies in constructing the request for this particular voice in comparison to others.

Additional Recommendations:

• Before generating your text-to-speech synthesis, it's advisable to utilize the voice endpoint of the API. This allows you to programmatically check if the desired voice is available in your current region at the time of execution
• Ensure you are using the latest version of the Text-to-Speech API client library. 
• Double-check your authentication credentials for accessing the Text-to-Speech service.

If the above options don't work, you can contact Google Cloud Support to further look into your issue.

I hope the above information is helpful.

[lweaver10]

Hi Dawn, sorry, but the above information is not particularly helpful, except for the part about that it seems to be a missing file or directory on the Google side. 

Any voice I try without a "Wavenet" type works fine, but it seems many of the Wavenet voices don't actually work. 

As I mentioned in the post, I go the list of voices straight from the API, and for example:

{"stdClass":{"languageCodes":["cmn-CN"],"name":"cmn-CN-Wavenet-D","ssmlGender":"FEMALE","naturalSampleRateHertz":24000}},

Is provided as an available voice. However, when I try to use that voice, I get:

\"code\": 404,

\"message\": \"Voice cmn-CN-Wavenet-A not found.\",

\"status\": \"NOT_FOUND\"

This implies to me that Google has deleted the voice. However, on the list of supported voices: https://cloud.google.com/text-to-speech/docs/voices

And as I mentioned, in the API get voices response, it's listed as available...

[lweaver10]

[cont] So are the voices unavailable but listed incorrectly? Or is there something special about how we need to call 'Wavenet' voices in the API?

Thanks for your response.

[dawnberdan]

Hi @lweaver10,

I understand that you're frustrated about the inconsistent Wavenet voices in the Text-to-Speech API. The information I provided didn't specifically address the Wavenet issue, but let's take a closer look at your points and see if we can find a solution together.

Potential Causes:

1. Temporary Outage:The Wavenet voices may be temporarily unavailable due to maintenance or unforeseen issues.
2. Regional Limitations:Despite being listed in the API, certain Wavenet voices might not be accessible in all regions. Verify your project's region and confirm that the desired voices are supported there.
3. API Discrepancy:There might be inconsistencies between the information provided by the get voices API call and the actual availability of the voices. This could be a bug on Google's side.

Recommendations:

1. Confirm Region Compatibility: Verify the region of your project and ensure that the Wavenet voices you require are supported in that specific location. Consult the Google Cloud Console for region information.
2. Test Voice Availability Dynamically: Before generating speech, utilize the voices endpoint of the Text-to-Speech API to programmatically check if the desired Wavenet voice is available in your region at runtime. This proactive step helps identify potential regional limitations.

If the above options don't work, you can contact Google Cloud Support to further look into your issue.

I hope the above information is helpful.