The ways we experience media on the web tend to be designed only for the publisher's intended medium. You don't consume videos just for their sound, for example, and podcasts don't often come with word-for-word transcripts, but there are ways of enjoying those pieces of content with augmentation. Some of the toughest challenges in making text sites cross-consumable via dictation has been in naturalizing machine voices and translating stories rich in grammatical tapestry from other languages. Google Assistant is now bringing its answers to those challenges with new text-to-speech functionality available today.

Google previewed these so-called Read It features at CES this year, touting their voices' improved handling of syntax and expression as well as the Assistant's ability to translate between 42 languages.

Once you trigger Assistant on top of a webpage and tell them to "read it," a voice will start reading what's displayed, highlighting words as they're saying them. You'll have playback controls including skip and reverse toggles as well as different read speeds. Through the overflow menu, you'll also be able to change the voice, which language the voice will read, and send feedback.

App developers can enable Read It through Actions on Google while webmasters can block the features for pages unsuitable for dictation by raising the nopagereadaloud tag.

Source: Google