Azure Speech / Voice Live – Missing DragonHDOmniLatestNeural voices (no German available)

Question

Azure Speech / Voice Live – Missing DragonHDOmniLatestNeural voices (no German available)

Tristan Rehm 20

Background

I am building a production voice assistant using Azure Voice Live (eastus2 region). The assistant serves German-speaking users and relies on DragonHDOmniLatestNeural voices for natural multilingual pronunciation, expressiveness, and conversational quality.

Currently, German Omni voices are not available — not through the API, not in the Voice Gallery, and not in Voice Live. I have verified this across multiple regions.

Current Behavior

Calling:

GET https://{region}.tts.speech.microsoft.com/cognitiveservices/voices/list

Returns:

~646 total voices
~10 DragonHDOmniLatestNeural voices only
- en-US (~8)
- zh-CN (~2)
0 German Omni voices

Tested across:

eastus2
germanywestcentral
swedencentral

This is consistent with the Voice Gallery and the Foundry UI.

Expected Behavior

The official GitHub sample lists ~700 Omni voices including German:

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

I would expect these voices to be available via the API and usable in Voice Live.

Questions

Is there a timeline for German Omni voice availability?
Is the GitHub voice list (700+ voices) a planned roadmap or outdated?
Which source should be trusted for actual Omni availability: API, Voice Gallery, or docs?

Thank you for any clarification.

0 comments

Answer accepted by question author

0 additional answers

Your answer

Answer 1

SRILAKSHMI C 18,745 Microsoft External Staff Moderator

Hello @Tristan Rehm

Thank you for the detailed information and for sharing the validation steps you performed.

Based on the behavior you described, your observations are consistent with the current public availability of DragonHDOmniLatestNeural voices in Azure Speech / Voice Live services.

At present, DragonHDOmniLatestNeural voices are currently available only for a limited set of locales, primarily:

• en-US

• selected zh-CN voices

German Omni voices are not currently exposed through:

• the Speech voices API (/cognitiveservices/voices/list)

• Voice Gallery

• Azure AI Foundry / Voice Live or Speech Studio

Your testing across multiple regions including eastus2, germanywestcentral, swedencentral aligns with the currently deployed backend availability.

For German language scenarios, standard Dragon HD voices are currently available, including:

• de-DE-Seraphina:DragonHDLatestNeural

• de-DE-Florian:DragonHDLatestNeural

However, the corresponding multilingual “Omni” variants for German do not appear to be publicly available at this time.

Regarding the GitHub sample listing ~700 Omni voices:

The GitHub sample JSON appears to represent a broader reference/superset catalog that may include planned voices, staged rollout entries, preview/internal definitions, or future roadmap items.

Because of this, the GitHub repository should not be considered the authoritative source for currently deployable production voices.

At this time, the most reliable sources for actual voice availability are the live GET /cognitiveservices/voices/list API response, Voice Gallery, Azure AI Foundry / Speech Studio UI

If a voice does not appear in these sources, it is generally not yet available for public deployment or Voice Live usage.

Regarding your specific questions:

• There is currently no publicly announced ETA for German DragonHDOmniLatestNeural voice availability.

• The GitHub voice list should be treated as a broader sample/reference catalog rather than confirmation of live service availability.

• The live API and Voice Gallery are the authoritative sources for currently supported Omni voices.

If early access to German Omni voices is important for your production scenario, we recommend engaging your Microsoft account team or opening an Azure support request to inquire about potential preview availability for de-DE Omni voices.

In the meantime, the currently available German DragonHDLatestNeural voices remain the supported GA option for German conversational synthesis workloads.

Please refer this

High-Definition voices in Azure Speech (includes HD & Omni naming convention) https://learn.microsoft.com/azure/ai-services/speech-service/high-definition-voices?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#supported-azure-speech-hd-voices

Language and voice support for Azure Speech (German HD voices table) https://learn.microsoft.com/azure/ai-services/speech-service/language-support?wt.mc_id=knowledgesearch_inproduct_azure-cxp-community-insider#text-to-speech-voices

I Hope this helps. Do let me know if you have any further queries.

If this answers your query, please do click Accept Answer and Yes for was this answer helpful.

Thank you!

Tristan Rehm 20 Reputation points

2026-05-20T13:56:24.3066667+00:00
Thank you for the detailed response and clarification.

I appreciate confirming that the live API is the authoritative source for available voices.

However, I'd like to follow up on one point: The official Microsoft Tech Community announcement for Dragon HD Omni explicitly states "700+ voices" as a key feature — not as a roadmap, but as a product capability: https://techcommunity.microsoft.com/blog/azure-ai-foundry-blog/introducing-dragon-hd-omni-azure-speech-new-voice-type-now-in-preview-via-micros/4481288

The article says:

"delivering over 700 high-quality voices"

"700+ prebuilt voices"

"Dragon HD Omni merges a wide range of prebuilt voices into one"

The official HD voices documentation also states:

"DragonHDOmni: 700+ voices (all released voices + new AI-generated)"

Source: https://learn.microsoft.com/en-us/azure/ai-services/speech-service/high-definition-voices

If only ~10 of these 700+ voices are actually available, should the announcement and documentation be updated to reflect this? Currently they give the impression that 700+ voices are usable, which is misleading for developers building on this platform.

Is there any timeline for when the remaining voices (especially German) will actually become available?

Thanks again for your help.
SRILAKSHMI C 18,745 Reputation points Microsoft External Staff Moderator

2026-05-21T16:15:45.32+00:00

Hi **Tristan Rehm,**Thank you for the additional clarification and for sharing the official Tech Community announcement and documentation references. I completely understand your concern here, especially from a production-planning perspective for multilingual Voice Live deployments.You are correct that the current public messaging and documentation around Dragon HD Omni can create the impression that all 700+ voices are immediately available for deployment and use through the Speech API, Voice Gallery, and Voice Live services.

Based on the current service behavior, the “700+ voices” wording appears to refer to the broader Dragon HD Omni voice architecture/catalog rather than the subset of voices that are presently enabled for public preview consumption across all regions and APIs.

At this time, the live service endpoints remain the authoritative source for actual deployable availability. As you observed, the currently exposed DragonHDOmniLatestNeural voices are limited primarily to:

• en-US voices

• selected zh-CN voices and German Omni voices are not yet publicly surfaced through:

• GET /cognitiveservices/voices/list, Voice Gallery, Azure AI Foundry / Voice Live, Speech Studio

I do agree that the distinction between:

• “supported by the Omni architecture/catalog” versus

• “currently deployable in public preview” is not clearly communicated in the current public documentation and announcement wording.

Unfortunately, Support does not currently have visibility into an official ETA or rollout timeline for additional Omni locales such as de-DE. Voice availability is typically rolled out progressively based on:

regional backend enablement, preview expansion stages, GPU/capacity readiness, and service validation milestones.

At the moment, there has not yet been a publicly communicated timeline for German DragonHDOmniLatestNeural availability.

For production workloads today, the recommended and fully supported German options remain the currently available: • de-DE-Seraphina:DragonHDLatestNeural • de-DE-Florian:DragonHDLatestNeural

while monitoring the live Voices List API and release notes for future Omni locale expansion.

Could you please take a moment to retake the survey on the above response? Your feedback is greatly appreciated.

Thank you!
SRILAKSHMI C 18,745 Reputation points Microsoft External Staff Moderator

2026-05-25T16:10:28.34+00:00

Hi Tristan Rehm,

Did you get any chance to review the above response. Do let me know if you have any further queries.

Thank you!
SRILAKSHMI C 18,745 Reputation points Microsoft External Staff Moderator

2026-05-26T10:53:27.3633333+00:00

Hi Tristan Rehm,

Did you get any chance to review the above response. Do let me know if you have any further queries.

Could you please take a moment to retake the survey on the above response? Your feedback is greatly appreciated.

Thank you!

Share via

Azure Speech / Voice Live – Missing DragonHDOmniLatestNeural voices (no German available)

Background

Current Behavior

Expected Behavior

Questions

0 additional answers

Your answer