Google Cloud is extending its speech-to-text API to an additional 200 million people around the world today with support for Albanian, Burmese, Estonian, Macedonian, Mongolian, Punjabi, and Uzbek. Including today’s language upgrades, the API now supports more than 70 languages.
Speech-to-text is a form of natural language processing (NLP) that converts speech into words and is often applied to things like video subtitles or transcription. As part of today’s news, a version of the Google Cloud Speech-to-Text API that delivers 60% better performance than the base model when transcribing telephone audio is now available in U.S. Spanish, U.K. English, and Russian.
The API is also getting an improved ability to attribute words to specific speakers in transcripts and to automatically punctuate languages like German, French, Japanese, and Swedish. This is the latest expansion of Google’s language offerings. Last week, Google Translate added support for new languages like Uighur and Turkmen, and on Wednesday Google Assistant gained reading and translation of long-form content with a more expressive voice. Google Assistant speaks twice as many languages today as Amazon’s Alexa, and more than any other AI assistant on the market.
In other recent Google Cloud news, due to the COVID-19 coronavirus, the Next conference was converted into an online-only event.