Microsoft on Monday made available the first public preview of Skype Translator, its service for real-time language translation. The preview program will kick off with two spoken languages, Spanish and English, according to Skype.
Additionally, more than 40 instant messaging languages will be available to Skype customers who have signed-up via the Skype Translator sign-up page and are using Windows 8.1 on their desktops or mobile devices. It’s also available to users with Windows 10 Technical Preview. Microsoft is expected to make it ultimately available across a wide variety of devices and operating systems over time
Microsoft announced Skype Translator last May, mentioning at the time that a public preview would be available this year. Microsoft has been working to deliver real-time language translation for at least two years. The new service translates conversations both ways in near real-time. The service will display on-screen transcripts of calls, and will ultimately translate instant-message chats in more than 45 languages.
"Skype Translator is a great example of the benefit of Microsoft’s investment in research," Skype said in a blog post. "We’ve invested in speech recognition, automatic translation and machine learning technologies for more than a decade, and now they’re emerging as important components in this more personal computing era. Skype Translator is the most recent and visible example."
The translator application uses machine learning, which means the more the technology is used, the more it improves. Skype said it expects the quality of the English and Spanish translations to advance as more people use the Skype Translator preview with the two languages.
Skype Translator’s machine learning protocols train and optimize speech recognition and automatic machine translation tasks. Those elements transform the recognized text to facilitate translation. This process includes the removal of disfluencies ("ahs," umms," and so on), the division of the text into sentences, as well as the addition of punctuation and capitalization.
The training data for speech recognition and machine translation comes from several sources, including translated Web pages, videos with captions and previously translated and transcribed one-on-one conversations. Translator records conversations to analyze the scripts and to better train the system to learn each language.
Skype started letting users sign up for the preview last month. At that time, the company said the tool would at first only be available on Windows 8.1 computers and tablets, but it later added support for Windows 10 Technical Preview.
Early Monday, the tool’s signup page listed 10 other languages besides English and Spanish as choices for users: Arabic, Chinese (Cantonese), Chinese (Mandarin), French, German, Italian, Japanese, Korean, Portuguese and Russian. The remaining 10 languages presumably are near-future candidates to be added to Skype Translator.
"While this moment is a major milestone for our team, we see the preview as simply another step in creating the best translation experience we possibly can," stated Skype.