Docs / Providers
Microsoft Azure (Neural TTS)
In one sentence: the brand with the broadest multilingual coverage. Chinese isn't its best, but if you need to read Japanese, Korean, Spanish, Arabic… or a passage mixing several languages, this one is the least hassle.
In one sentence: the brand with the broadest multilingual coverage. Chinese isn't its best, but if you need to read Japanese, Korean, Spanish, Arabic… or a passage mixing several languages, this one is the least hassle.
What This Is
Neural TTS in Microsoft Azure Cognitive Services — one of the world's most widely used commercial speech-synthesis services. It displays as Microsoft Azure International in Soniva.
Pricing
| Model | Official upstream price | Credits / 1,000 chars |
|---|---|---|
Neural TTS (azure-neural-tts) | $15 / 1M chars | 17 |
Price source: https://azure.microsoft.com/en-us/pricing/details/speech/ (verified 2026-08-17)
There's only one model tier, so there's no choosing among models — just pick a voice.
At 17 credits/1,000 chars it sits in the site's lower-middle range, almost identical to Alibaba Cloud's default tier (16). Given the number of languages it covers, the price is a bargain.
Microsoft also has a pricier Neural HD ($22/M) tier, but Soniva currently hasn't integrated those voices, so you can't accidentally buy the more expensive tier.
The Biggest Advantage: Languages
Microsoft's voice library covers 100+ languages and regional variants, the most complete of the eight:
- Chinese includes not just Mandarin but Cantonese, Taiwan Mandarin, and various mainland dialect accents
- English has a dozen-plus accents — US/UK/AU/IN/CA/IE, etc.
- Japanese, Korean, French, German, Spanish, Portuguese, Russian, Arabic, Hindi… basically all there
- Same language further split into male/female/child voices and multiple speaking styles
If your content isn't Chinese, or isn't only Chinese, try this brand first.
Voice List Is Pulled Live
Microsoft's voice list is pulled live from the upstream interface by Soniva every time — when Microsoft adds voices, you see them in Soniva immediately, no need to wait for a release.
So this document doesn't state a specific voice count (it would go stale fast) — go by what you actually see in the product.
The free-preview representative voice is Xiaoxiao (zh-CN-XiaoxiaoNeural), Microsoft's most common Mandarin female voice.
When to Choose Microsoft
| Scenario | Recommended? |
|---|---|
| Japanese / Korean / European language content | ✅ First choice — the other seven've coverage of non-Chinese/English languages doesn't match it |
| Content mixing Chinese and English | ✅ The same voice switches languages naturally without garbling |
| Need a specific regional accent | ✅ The only brand offering this level of accent granularity |
| Pure Chinese long text, value-focused | ⚠️ Usable and not pricey, but Tencent Premium (5) / Baidu Basic (5) cost less |
| Want the most natural Chinese possible | ⚠️ Alibaba Cloud CosyVoice and Doubao Speech have an edge in Chinese |
Related Documents
Was this article helpful?
Anonymous feedback is enabled when you allow analytics and the service is available. You can also report an issue by email.
The email includes this article. You can also copy the details and send them to soniva.cde@gmail.com