Google plans to add a real-time translation feature to its Google Translate app for Android sometime in the future. This feature will allow users to record audio in one language and then present it in another language in real time, according to The Verge.
Although the feature is still in the prototype stage, Google demonstrated the technology on Tuesday in a series of artificial intelligence demonstrations in its San Francisco office, the report said.
Google said that the feature may require an Internet connection at launch, meaning it will be different from the current AI and device-based translation features in Google Translate.
The company said this is because real-time multilingual transcription is much more complicated than simply translating written text from one language to another or typing a single phonetic sentence and translating it into another language.
At least initially it won't support audio files. It has to record audio in real time through the smartphone microphone, but Google says that users can play the recorded audio through the speaker and record it this way.
Google said that because audio is constantly in progress, its transcription function constantly evaluates entire sentences. In addition, it adds punctuation and corrects certain word choices based on the context of the sentence. It also attempts to correct accents and local dialects.
Unfortunately, the company did not give a specific timetable for when Google will launch this real-time transcription and translation function, but only said that it would be launched at a "future" point in time.
