Microsoft Launches Live Captions Real-Time Translation Feature
- Business
- January 6, 2025
Microsoft declared on December 18, 2024, that it was testing real-time translation for its live captioning function on Copilot+ PCs with AMD and Intel processors. For use cases including “real-time video calls, recordings, and streamed content,” the roll-out will progressively include the capability to translate live captions from over 44 source languages into English, according to a Windows blog post.
Additionally, the upgrade improves real-time translation for Copilot+ PCs with Snapdragon processors, enabling users to translate between 26 languages in real-time to Simplified Chinese.
Even though the service is called “live captions,” the new real-time translation capability effectively transforms it into a subtitling tool: Whether a video is recorded or live, live subtitles will “instantly” translate it “in any app or video platform.” In an adjustable overlay window, users can select whether to see the captions “above screen,” “below screen,” or “floating on screen.”
The name “captions” appears to be a reflection of Microsoft’s branding of the function under the accessibility umbrella, which is frequently used to distinguish between users who are deaf and hard of hearing and those who are accessing content in a language they do not understand. In a Microsoft 365 Support page, Microsoft advises users to “use live captions to better understand audio,” expanding the term’s usage.
“To make more content accessible to more people, live captions now has the ability to provide translations and will turn any audio that passes through your PC into a single English-language caption experience,” the page explains.
After requesting permission to process voice input on the device, live captions ask users to download language files for on-device speech recognition from one or more of the 44 available source languages. Users can utilize their PC’s microphone to caption their own voice and add additional languages by downloading the relevant files later.
One of the biggest tech firms in the world, Microsoft, made a number of its advancements in language technology public throughout 2024. These included releasing a multilingual AI translation model, previewing an AI speech translation app, and revising a patent file for automatic dubbing.
However, Microsoft warned users about “known issues to be fixed in future updates to Windows Insiders” (along with recommended patches), pointing out that live captions may crash during the initial launch and when switching languages while the microphone is on or the audio is playing.