https://github.com/Azure-Samples/Cognitive-Speech-TTS/blob/master/Blog-Samples/Introducing-Dragon-HD-Omni/dragonhdomni_voice_list.json is broken

Question

https://github.com/Azure-Samples/Cognitive-Speech-TTS/blob/master/Blog-Samples/Introducing-Dragon-HD-Omni/dragonhdomni_voice_list.json is broken

Ming-Li Lin 20

There are multiple issues.

The json file is broken.
All the AI generated voices were removed. Seriously? You removed half of the Omni voices without any announcement first? I'm seriously considering moving my backbone TTS service from Azure to other platforms. I tested one voice name that's removed: zh-cn-amberopal:DragonHDOmniLatestNeural it can still work querying tts.speech.microsoft.com/cognitiveservices/v1. So what's your plan now? Can I expect them to come back? Or those voices will be gone eventually?

Vinodh247-1375 43,746 Reputation points Volunteer Moderator

2026-07-29T04:32:26.5433333+00:00
Please confirm if you had a chance to try any of the suggestions provided for your question.

Please 'Upvote'(Thumbs-up) and 'Accept' as answer if the reply was helpful.

If in case you have any resolution please share the same here with the community as it can be helpful to others. Else, please respond with more details and we will try to see if we can help further.

1 answer

Your answer

Vinodh247-1375 43,746 Reputation points Volunteer Moderator

2026-07-29T04:32:26.5433333+00:00

Please confirm if you had a chance to try any of the suggestions provided for your question.

Please 'Upvote'(Thumbs-up) and 'Accept' as answer if the reply was helpful.

If in case you have any resolution please share the same here with the community as it can be helpful to others. Else, please respond with more details and we will try to see if we can help further.

Answer 1

Hi ,

Thanks for reaching out to Microsoft Q&A.

this looks like a docs/sample repo drift + preview churn, not a cleanly communicated deprecation. Treat Omni voices as non-stable contracts right now, lock down the voices you depend on, add fallback mapping, and plan for churn until GA

What is happening:

Dragon HD Omni is still in preview and evolving, and Microsoft has been actively reshaping the voice catalogue (it originally introduced hundreds of AI generated voices)
The GitHub JSON is not a source of truth, it is just a sample artifact and can lag or break (which is what you are seeing).
The actual source of truth is the /voices/list API, and as you noticed, voices like zh-cn-amberopal:DragonHDOmniLatestNeural can still work if the backend model is not yet retired.
There are already signs of inconsistent availability across regions and surfaces (voices missing from gallery/API/UI depending on region or feature surface)

What this likely means:

No clear public signal that all removed voices are coming back.
Expect consolidation toward fewer “stable” personas + dynamic Omni behaviour, rather than maintaining hundreds of static AI-generated voices.
If a voice is no longer returned by /voices/list, assume it is effectively deprecated, even if it still works temporarily via direct calls.

Please 'Upvote'(Thumbs-up) and 'Accept' as answer if the reply was helpful. This will be benefitting other community members who face the same issue.