Although Google Translate and similar services already offer offline text translation, offline voice translation, which requires the extra step of automatic speech recognition (ASR), is the more challenging proposition — as revealed by Google’s own efforts in rolling out Translatotron.
With Tensor, Pixel 6 can operate Google’s Live Caption and Interpreter Mode features simultaneously, something not possible with previous chips.
Slator 2021 Language Industry Market Report
80-pages. Market Size by Vertical, Geo, Intention. Expert-in-Loop Model. M&A. Frontier Tech. Hybrid Future. Outlook 2021-2025.
“Speech recognition is another foundational technology where you will see a huge improvement in #Pixel6,” according to an August 2, 2021 tweet from Google’s product-focused Twitter account, @MadebyGoogle. “Google Tensor allows us to make big leaps in: voice commands, translation, captioning, & dictation,” claimed a follow-on tweet.
Osterloh has already demonstrated Pixel 6’s “nearly instantaneous” talk-to-type function. Users can continue to speak while editing the resulting text by using their keyboard. In another demo, Osterloh turned on both live captions and live translation, converting a French recording into English.
Fans from various countries are already clamoring for the Pixel 6 and Pixel 6 Pro. Playing coy, Google said they will share more details in the fall. The tech behemoth has also yet to specify which languages (or even how many) will be covered by the live-translated-caption feature.
Pixel 6’s capabilities represent just one possible productization of Google’s recent research into speech-to-speech translation. It will be interesting to see how far Google can go after quintessential smartphone maker Apple struggled to gain a following for the iPhone’s Translate app since its September 2020 release.