Introducing more multilingual AI voices optimized for conversations

By Gang Wang, Lihui Wang, Peter Pan, Xi Wang, Shaofei Zhang, Zhihang Xu, Wenning Wei, Jingzhou Yang, Zheng Niu, Melissa Ma, Nick Zhao, Qinying Liao, Garfield He, Jun-Wei Gan, Lei He, Binggong Ding, and Sheng Zhao

In our interconnected world, effective communication often faces the significant challenge of language barriers. Fortunately, rapid advancements in and natural language processing are bringing about revolutionary solutions. Microsoft's cloud-based service, Azure AI Speech text to speech, stands at the forefront of this transformation. Notably, the introduction of JennyMultilingualV2 and RyanMultilingual in August has fluently expanded language capabilities to 41 locales. This enhancement empowers users to engage with a more global audience while maintaining a consistent persona across diverse markets.

The JennyMultilingual voice has progressed significantly from its initial support for 14 languages. In the August update, we expanded language capabilities to encompass 41 languages/accents, introducing auto language detection for the primary locale. Furthermore, users now have the flexibility to utilize SSML to set preferred accents in such as opting for a British accent (en-GB) for English.

Presently, we are unveiling new multilingual voices characterized by a more conversational style, stemming from diverse primary languages. This expansion broadens our capacity to articulate content across a total of 91 languages and their variants. This advancement underscores our steadfast commitment to surmounting language barriers and promoting a more inclusive and accessible global communication landscape.

List of supported language/variants

Locale Languages Locale Languages Locale Languages Locale Languages
af-ZA Afrikaans (South Africa) jv-ID Javanese (Latin, Indonesia) es-ES Spanish (Spain) ru-RU Russian (Russia)
am-ET Amharic (Ethiopia) ka-GE Georgian (Georgia) es-MX Spanish (Mexico) si-LK Sinhala (Sri Lanka)
ar-EG Arabic (Egypt) kk-KZ Kazakh (Kazakhstan) et-EE Estonian (Estonia) sk-SK Slovak (Slovakia)
ar-SA Arabic (Saudi Arabia) km-KH Khmer (Cambodia) eu-ES Basque sl-SI Slovenian (Slovenia)
az-AZ Azerbaijani (Latin, Azerbaijan) kn-IN Kannada (India) fa-IR Persian (Iran) so-SO Somali (Somalia)
bg-BG Bulgarian (Bulgaria) ko-KR Korean (Korea) fi-FI Finnish (Finland) sq-AL Albanian (Albania)
bn-BD Bangla (Bangladesh) lo-LA Lao (Laos) fil-PH Filipino (Philippines) sr-RS Serbian (Cyrillic, Serbia)
bn-IN Bengali (India) lt-LT Lithuanian (Lithuania) fr-BE French (Belgium) su-ID Sundanese (Indonesia)
bs-BA Bosnian (Bosnia and Herzegovina) lv-LV Latvian (Latvia) fr-CA French (Canada) sv-SE Swedish (Sweden)
ca-ES Catalan (Spain) mk-MK Macedonian (North Macedonia) fr-CH French (Switzerland) sw-KE Swahili (Kenya)
cs-CZ Czech (Czechia) ml-IN Malayalam (India) fr-FR French (France) ta-IN Tamil (India)
cy-GB Welsh (United Kingdom) mn-MN Mongolian (Mongolia) ga-IE Irish (Ireland) te-IN Telugu (India)
da-DK Danish (Denmark) ms-MY Malay (Malaysia) gl-ES Galician th-TH Thai (Thailand)
de-AT German (Austria) mt-MT Maltese (Malta) he-IL Hebrew (Israel) tr-TR Turkish (Turkey)
de-CH German (Switzerland) my-MM Burmese (Myanmar) hi-IN Hindi (India) uk-UA Ukrainian (Ukraine)
de-DE German (Germany) nb-NO Norwegian Bokmål (Norway) hr-HR Croatian (Croatia) ur-PK Urdu (Pakistan)
el-GR Greek (Greece) ne-NP Nepali (Nepal) hu-HU Hungarian (Hungary) uz-UZ Uzbek (Latin, Uzbekistan)
en-AU English (Australia) nl-BE Dutch (Belgium) hy-AM Armenian (Armenia) vi-VN Vietnamese (Vietnam)
en-CA English (Canada) nl-NL Dutch (Netherlands) id-ID Indonesian (Indonesia) zh-CN Chinese (Mandarin, Simplified)
en-GB English (United Kingdom) pl-PL Polish (Poland) is-IS Icelandic (Iceland) zh-HK Chinese (Cantonese, Traditional)
en-IE English (Ireland) ps-AF Pashto (Afghanistan) it-IT Italian (Italy) zh-TW Chinese (Taiwanese Mandarin, Traditional)
en-IN English (India) pt-BR Portuguese (Brazil) ja-JP Japanese (Japan) zu-ZA Zulu (South Africa)
en-US English (United States) pt-PT Portuguese (Portugal) ro-RO Romanian (Romania)

 

Examples of new multilingual voices in some languages

Please note that the provided samples showcase only a selection of languages. Feel free to experiment with your own texts using the Audio content creator tool in Speech Studio.

  

Voice Gender Primary language Sample Text Sample locale (s) Sample Audio
en-US-AvaMultilingualNeural Female English (United States) The world is full of beauty when your heart is full of love.

Le monde est plein de beauté quand votre cœur est plein d'amour.

Il mondo è pieno di bellezza quando il tuo cuore è pieno d'amore.

English – French – Italian
en-US-EmmaMultilingualNeural Female English (United States) Well, John, that's very kind of you to invite me to your quarters. I'm flattered that you want to spend more time with me.

No coração da cidade, onde as luzes cintilantes se encontram com o céu estrelado, há uma praça tranquila.

我独自走在这个小镇的夜晚,仿佛可以听到岁月的脚步,感受到历史的气息。

English – Portuguese – Chinese
en-US-AndrewMultilingualNeural Male English (United States) Success is not the key to happiness, but happiness is the key to success.

Erfolg ist nicht der Schlüssel zum Glück, aber Glück ist der Schlüssel zum Erfolg.

El éxito no es la clave de la felicidad, pero la felicidad es la clave del éxito.

English – German – Spainish
en-US-BrianMultilingualNeural Male English (United States) In this moment, I realized that amid the chaos of life, tranquility and peace can always be found.

人生は、一度きりのチャンスです。失敗しても、次に向けて立ち上がりましょう。

강물은 은하수처럼 반짝이며, 바람이 나뭇잎을 부드럽게 스치면서 향기로운 공기를 흔들었습니다.

English – Japanese  – Korean
de-DE-SeraphinaMultilingualNeural Female German (Germany) Die Einführung nachhaltiger Praktiken in unserem täglichen Leben, wie z. B. die Einsparung von Wasser und Energie, die Auswahl umweltfreundlicher Produkte und die Reduzierung unseres CO2-Fußabdrucks, kann erhebliche positive Auswirkungen auf die Umwelt haben.

Conserving water resources through efficient usage and implementing responsible water management practices is crucial, especially in regions prone to drought and water scarcity.

鼓励负责任的旅行选择、支持当地社区,并推广环保住宿,有助于保护脆弱的生态系统并保护文化遗产。

German – English – Chinese
de-DE-FlorianMultilingualNeural Male German (Germany) Umwelterziehung sollte in den Schulcurricula integriert werden, damit zukünftige Generationen die Bedeutung des Umweltschutzes verstehen können.

I media svolgono un ruolo fondamentale nel plasmare l'opinione pubblica e sensibilizzare sulle questioni ambientali.

환경 교육은 미래 세대가 환경 보호의 중요성을 이해할 수 있도록 학교 커리큘럼에 통합되어야 합니다.

German – Italian – Korean
fr-FR-VivienMultilingualNeural Female French (France) Des efforts de collaboration entre les pays sont nécessaires pour lutter contre le changement climatique, protéger les océans et préserver les écosystèmes fragiles.

Al implementar programas integrales de reciclaje y fomentar la reutilización de materiales, podemos minimizar la cantidad de desechos que terminan en vertederos o contaminan nuestros océanos.

責任あるジャーナリズムは、環境問題に焦点を当て、成功した保全活動を紹介し、行動を促すことができます。

French – Spanish – Japanese
fr-FR-RemyMultilingualNeural Male French (France) L'éducation environnementale doit être intégrée dans les programmes scolaires afin que les générations futures puissent comprendre l'importance de la protection de l'environnement.

Proteger o meio ambiente não é apenas preservar os recursos naturais; trata-se também de priorizar o bem-estar humano.

Het aanmoedigen van verantwoorde reiskeuzes, het ondersteunen van lokale gemeenschappen en het promoten van milieuvriendelijke accommodaties helpt kwetsbare ecosystemen te beschermen en cultureel erfgoed te behouden.

French – Portuguese – Dutch
zh-CN-XiaoxiaoMultilingualNeural Female Zh-CN Chinese (Mandarin, Simplified) 嗯,我能理解你的担忧,社会对于身材的期望和审美标准可能会给人带来压力和焦虑。但是呢,我们每个人的身体都是独一无二的,健康与美丽并不仅仅取决于体重或外表。重要的是关注自己的整体健康和幸福感。

The garden was Alice's sanctuary, where flowers danced and whispered. “Do you speak?” she asked a blooming rose, her tone filled with wonder. “In the language of fragrance,” it seemed to reply, a sweet scent enveloping her in a conversation as old as nature itself.

Wenn du durch die Straßen gehst, hörst du oft seine Melodien aus den Cafés und sogar von Straßenkünstlern. Es ist, als ob Beethoven immer noch hier ist und die Menschen mit seiner Musik verbindet.

Chinese – English – German
zh-CN-XiaochenMultilingualNeural Female Zh-CN Chinese (Mandarin, Simplified) 嗯,如果要早起的话,我觉得你可以先设定一个闹钟,闹钟响了之后啊,你就离床远一点儿。培养规律,再就是早餐,早餐是很重要的,一定要吃。然后呢,就是锻炼,你可以培养决心和意志。希望你可以早起成功!

¿Por qué a los jugadores de fútbol mexicanos les gusta tanto el café? Porque son expertos en tomar “penales” dentro del área.

한국 치킨은 다양한 맛의 소스로 튀겨지거나 튀겨진 후 소스를 발라 먹는 경우가 많아요. 특히 후라이드 치킨, 양념 치킨, 간장 치킨 등이 인기 있어요. 튀김 과정에서 특유의 바삭한 피부와 쫄깃한 살이 만들어지는 것이 특징이에요. 이로 인해 한 입 베어 물 때의 식감이 독특하게 느껴져요.

Chinese – Spanish – Korean
zh-CN-YunyiMultilingualNeural Male Zh-CN Chinese (Mandarin, Simplified) 呃,我是知道阿西莫夫的机器人三大定律的,那这些定律啊在科幻文学中是被用来规范机器人的行为的,嗯,它们是有助于探讨与机器人和人工智能、嗯,人工智能相关的伦理问题。但是其实啊,这些定律它并不完美,嗯,比如说,第一定律要求机器人不得伤害人类,但是他没有明确地说什么叫做伤害,那这样的话,它的模糊性可能就会导致不确定和不一致的解释。

『NARUTO』は、岸本斉史によって作られたマンガで、若い忍者の成長と冒険の物語を描いています。物語の主人公はうずまきナルトで、彼は意気込みがあり、情熱的な忍者ですが、同時に村の人々から孤立している存在でもあります。なぜなら、彼の誕生時に封じ込められた妖怪キツネが原因で、村人たちから嫌われています。

Joyeux Noël et une nouvelle année remplie de bonheur, d'amour et de réussite. Que la magie de Noël illumine votre cœur et apporte la chaleur de l'amitié dans votre foyer. Paix et joie pour vous et vos proches pendant cette saison festive et tout au long de l'année à venir. Joyeux Noël

Chinese – Japanese – French

Additional Updates

  • Xiaoxiao is supporting a new dialect: Minnan “nan-CN”, as well as new Mandarin accents from more provinces, such as: Anhui, Hunan, Gansu
  • Xiaoxiao is also able to speak multiple Chinese dialects/accents as below:

Examples

Voicename Secondary Locale Language variant Text Audio
XiaoxiaoDialects zh-CN-shaanxi Chinese (Zhongyuan Mandarin Shaanxi, Simplified) 事情就是这样的,你现在听清楚了么?
zh-CN-sichuan Chinese (Southwestern Mandarin, Simplified) 村子里上了些岁数的人都认得我。
zh-CN-shanxi Chinese (Shanxi Accent Mandarin, Simplified) 你说电视台让你失望,你给电视台什么希望没有?
nan-CN Chinese (Southern Min, Simplified) 其实人生,本没那么简单,也没那么复杂。
zh-CN-anhui Chinese (Jianghuai Mandarin Anhui, Simplified) 有问题了是封不住的,就算封住了对我们有什么好处呢?
zh-CN-hunan Chinese (Hunan Accent Mandarin, Simplified) 所以趁着天还亮,赶紧生火,准备烧烤。
zh-CN-gansu Chinese (Lanyin Mandarin Gansu, Simplified) 有了明确的任务,孩子学习时就有了动力,才能保持紧张状态。
zh-CN-shandong Chinese (Jilu Mandarin, Simplified) 与外迁相应的是技术升级,淘汰落后,脱胎换骨。
zh-CN-henan Chinese (Zhongyuan Mandarin Henan, Simplified) 爱是风儿,是阳光,是雨露,是绿叶,是花香。
zh-CN-liaoning Chinese (Northeastern Mandarin, Simplified) 这么顺利?俺们这咋一点儿进展没有呢。
zh-TW Chinese (Taiwanese Mandarin, Traditional) 他們也非常瞭解要如何在工作與休閒之間取得平衡,同時很善於說故事。

These new updates above will be in public preview in three regions: East US, West Europe, and Southeast Asia. 

Get started

Microsoft offers over 400 neural voices covering more than 140 languages and locales. With these text to speech voices, you can quickly add read-aloud functionality for a more accessible app design or give a voice to chatbots to provide a richer conversational experience to your users. In addition, with the Custom Neural Voice capability, you can easily create a brand voice for your business with professional voice clone.

You now have the capability to activate the Speech feature in ChatGPT Playground. This enhancement facilitates seamless audio testing, offering a refined and more authentic conversational experience for integration into your bot application.

GarfieldHe_0-1702980951597.png

For more information

 

This article was originally published by Microsoft's Azure AI Services Blog. You can find the original article here.