The web was all abuzz in December when a scoop by Android Police showed screenshots taken of a new version of the Google Translate app that integrated Word Lens’s image translation technology. No longer a scoop, it’s here — the details and demonstrations show that, happy hype or not, the Google Translate app with Word Lens functionality was well worth the wait.
Fame Monster App
Word Lens, a formerly independent “reality translation” app, joined the Google family of services after Google’s acquisition of Quest Visual in May in 2014. Quest Visual, the makers behind Word Lens, first launched the app in December 2010 to widespread praise. Besides gaining and holding the number one slot in Apple’s iTunes App Store for a number of days after its release, it also garnered major mainstream press coverage from the New York Times to the Wall Street Journal, and made it onto several “best of” lists in the technology space. Early in 2014, both Google and Apple had included the app in their own marketing and promotion efforts: it was demonstrated along with Google Glass (now dead or not dead, depending on your source) and was featured in Apple’s iPhone 5s television commercial and microsite.
While a translation application, Word Lens’s most compelling feature was not the translation content but the visual presentation. Clearly printed source text was viewed through a smartphone’s camera lens, optically recognized, translated, and supplanted in the image by the translated text — all in real time and in the same font and sizing as the source. Google Translate had a camera translation feature itself as early as August 2012 but the translation was not visually presented in context.
Its Google Translate integration shows that the strength of that core functionality remains. While at least some have noted that the translations show that we are still far from universal translator of the Star Trek universe, it is still right for the diners, travelers, and other casual users who will rely on its on-the-fly translation support. Otavio Good, Word Lens’s developer and the founder of Quest Visual, was quoted saying to Forbes in 2010, “It’s not perfect, but perfect was not the goal.”
Google’s acquisition may change that. We may tease about its limitations, but Google Translate does a lot of things right. It is the go-to tool for some 500,000 users each month and processes more than 1 billion translations per day, according to the Official Google Translate Blog. As of January 2015, it supports 90 languages, including the 10 minority languages that were added from India, Indonesia, Kazakhstan, Malawi, Madagascar, Myanmar, Sri Lanka, South Africa, Tajikistan, and Uzbekistan in late 2014, covering hundreds of thousands of native speakers of these languages.
Some Restrictions May Apply
The comprehensive language support that is available from Google Translate via the desktop browser is not however available in its app version. Instead of all 90 languages, only 36 are available to the visual translation feature. While impressive, the voice translation feature is similarly limited. Voice translation has been available since 2013, but performance enhancements make for a smoother experience. Previously, as the Official Google Translate Blog describes it, each user would tap the app’s microphone icon to speak into the smartphone to receive the translation, before handing it over to the other person for a reply. Now, the language is set once and then the app “listens” for the source and target languages.
In the video below, a journalist with the Guardian conducts on-the-street tests in New York City’s Times Square for both the visual and spoken translation features.
With news of Microsoft’s beta rollout of its Skype instant translation feature and Facebook’s acquisition of Wit.ai, a natural language recognition API, there is plenty of exciting activity in the language tech space. Who knows what we will be treated to next!