Unveiling Breakthroughs in Real-Time Translation with Skype Translator

Earlier this evening Microsoft CEO Satya Nadella in Rancho Palos Verdes, Calif., in  talk during the Code Conference, unveiled an early look at the Skype Translator app. This app represents a breakthrough in language translation jointly developed by Microsoft researchers and Skype engineers, bridging geographic and language barriers through the use of real-time speech-to-speech translation. The functionality combines Skype voice and instant messaging, Microsoft Translator and machine-learning based technologies for speech recognition that are used in Windows and Windows Phone Translation applications today.

During Nadella’s conversation with Kara Swisher and Walt Mossberg of the Re/code tech website relating to a new era of personal computing, he asked Gurdeep Pall, Microsoft Corporate Vice President for Lync and Skype, to join him on stage. While on stage, Pall demonstrated for the first time publicly the Skype Translator app, with Pall conversing in English with German-speaking Microsoft employee Diana Heinrichs.

Watch the Demo

Speech has been a natural evolution of the translation work that Microsoft has been delivering to consumers and businesses across a broad number of products and solutions. The work represents over a decade of work within Microsoft Research that has become a reality through a series of remarkable research advances in translation, speech recognition, and language processing. This demonstration is the next step in delivering the real time speech translation experience to users that Rick Rashid, then the worldwide head of Microsoft Research, demonstrated a year and a half ago.

The Skype Translator app will available first on Windows 8 later this year as a limited beta.

It has been an exciting day as we unveil this remarkable technology advancement that brings people one step closer to removing barriers of communication regardless of language or location!

Learn More about Skype Translator

Comments (4)

  1. Manuel Herranz says:

    This is a massive breakthrough in combining speech and MT technologies. As a developer myself and with years of machine translation customization behind me at Pangeanic, I can only wonder how massive the corpus and combination of syntax and statistical resources required only for fast MT.

    I would like to try the system as usually very specific speech may fail in very specific domains, and hard accents are a known problem.  I'm curious about online training and the ability to improve with user feedback

  2. Ari says:

    I would like to suggest another application of this technology, which is closed captioning of a video call. My wife is deaf but speaks and this would enable us to communicate via Skype instead of text messages.

  3. @Ari: this is a great idea and is actually here already in this version. If you look the video carefully, you will see that both the spoken text recognized by the speech-to-text engine and the translation are showed at the bottom of the video.

  4. NN says:

    I have good idea – why not introduce Firefox add-on Bing translator? I tried many and all of them insist on using Google Translae or just don't work even when they promote using Microsoft service, so in the end I can't avoid Google even if I want to use Bing Translator and I am for sure not switching to IE.

    Btw. nice job with redesign of Translator webpage but you still don't display pinyin transcription of Chinese characters so can't really compete with Google Translate even here. 🙁

Skip to main content