A new iPhone translation app, called Word Lens, is turning heads by showing how the power of augmented reality can be applied to real-time translation. The app allows users to point their iPhone camera at Spanish or English text and see an almost instant translation on their screen with words appearing in a similar font to the original. The app is limited to those two languages for now, though other languages are in the works.
But doesn’t Google do this with Goggles? Yes, and no. It uses a phone camera and optical character recognition to pull off translations, but it doesn’t do it in real time or lay digital images over your camera view, which is what augmented reality does. So which is better? I took them both for a quick spin.
Word Lens. First, Word Lens, which is listed as a free app, is not is cheap. You have to pay $4.99 for a language pack to translate anything. Sadly, you have to buy one pack for Spanish to English and another pack for English to Spanish. After choosing between language, the app brings up a camera view. As you line up text, it starts to spit out translations.
The translations are hit and miss. Sometimes, it nails the meaning of a sentence, but the translation can be very literal. And the words being translated keep jumping back and forth between the original word and words that might serve as good translations. For example, I looked at business headline in my local Spanish newspaper and got, “More workers pass now to the sector private.” It gets you the idea, but not the way you’d actually translate it. It isn’t nearly as smooth as a demonstration video that developer Questvisual put out.
The fun is in actually seeing the words transform right before your eyes. It can pull this off in one second, although, again, the words will often keep changing. This feeds my impatient side and really shows you what augmented reality can do. When done right, AR can help bring a foreign world into focus. This implementation, while still a work in progress, demonstrates one of the best uses of AR and may help bring the technology into the mainstream consciousness. For it to really happen, it should fix some of its flaws and improve its accuracy. A price cut would also help a lot.
Google Goggles. Google has created a powerful tool in Goggles, a mobile app for Android and iOS which can read book titles, contact info and text and can recognize landmarks, logos and wine labels. Earlier this year, Google enabled translations of text, first between German and English and later, a host of languages. Unlike AR, Googles doesn’t process and present things in real time. It snaps a photo, then sends it to Google’s servers, which do recognition and translation before sending the results back.
What you get is a slower process, but one that’s more accurate. First, you often have to use a cropping tool to limit the amount of words being translated. Then, after you snap a picture of a sentence, it can often take at least 20 seconds to be converted into digital text. After a press of the translate button, you get back the meaning, which is definitely more accurate and nuanced than Word Lens’s attempts. That same Spanish newspaper headline came back as “More workers now go to private sector,” pretty much the way a normal person would translate it.
This has the ability to wow as well, especially with the more than 40 languages that can be translated, and it’s all free. Howeer, it’s still too slow for my tastes. Overall, neither service is extremely compelling at this point. I don’t like the limited language selection, price and inaccuracy of Word Lens, and I need better immediacy from Google Goggles. What we really need is the quick response of Word Lens with the accuracy and broad powers of Google Goggles. I imagine we’ll get that over time. Then, we’ll really be seeing some jaw-dropping stuff. It’s not that far off, so stay tuned.
Related research from GigaOM Pro (subscription req’d):