Unveiling Bilingual Speech Modeling for Seamless Multilingual Conversations


By Jacky Kang, Eric Sun, Jing Pan, Yuhui Wang, Mark Hillebrand, Padma Varadharajan, Fei Zuo

We're thrilled to announce a groundbreaking feature in our real-time speech modeling—Bilingual Speech Modeling. This innovation is a game-changer, as our speech model now seamlessly supports bilingual language pairs, such as English and Spanish, as well as English and French. Imagine effortlessly switching between these languages during your real-time interactions—it's now a reality!

Why Bilingual Support Matters:

Our latest release is tailor-made for bilingual speakers who fluidly transition between English and Spanish. Whether you're at work, home, or engaging in community settings, this feature is designed to enhance your communication experience.

Elevate Your User Experience:

Bilingual speakers stand to benefit significantly from this feature, thanks to the model's remarkable ability to understand and respond in both languages in real time. The result? A more effective and fluid communication experience across various contexts.

Make the Most of It:

Choosing between languages has never been easier. Starting from 11/30, when you call the Speech Service API or explore Speech Studio, simply opt for es-US (Spanish and English) or fr-CA (French and English). Feel free to speak either language or mix them together—the model dynamically adapts to provide accurate and context-aware responses in both languages. It's a new era of seamless, multilingual communication!


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