Like the Babel Fish, Google has managed to do the seemingly impossible. Announced at their October 4th hardware event, their new Pixel Buds are capable of translating 40 different languages in near real-time. They’re fast enough to decode what another person is saying, so much so that one can carry on a conversation with ease.
Taking the stage at their San Francisco event, the earbuds were able to translate Swedish to English on the fly. This was however an admittedly controlled event, so only time will tell how well the device works in the real world where traffic, accents, and whatnot could confuse the software. Still though, it was an exciting glimpse into how Google sees our future.
To use the device, you will need a data connection though. Due to the amount of processing power needed for the translation, the audio is processed via Google’s AI-focused data centers. The audio is first converted to text, translated into another language, and then turned back into synthesized speech which is relayed to the listener.
Normally, the last part of this process is done by stitching pre-recorded words and word fragments together. However, Google has recently revealed that through its AI research they’ve been able to develop a system called WaveNet, which is used to generate human sounding voices. Even better, WaveNet is now part of Google Assistant which is found in the latest Android phones and Google hardware devices. This means that, possibly, your phone will be generating the synthesized speech for you, instead of a Google data center. Eventually, we may have devices capable of doing the translation without the need for an internet connection whatsoever.
Retailing for $150, the Google Pixel Buds provide roughly 5 hours of battery life before they can be recharged from a battery pack in their carrying case. So while they may not last you through an entire day in a foreign country, they’re certainly the start of something exciting.