The speech engine Speech Services by Google is being upgraded to improve clarity and make text-to-speech voices in Android apps sound more natural. You can hear the difference between the old voices and the updated voices for yourself through prerecorded snippets on the Android Developers Blog.
Frankly, while the voices do sound clearer, I’m skeptical about the claims it sounds more natural. It’s also still difficult to ascertain what the first sentence in these US English-language recordings actually says — is this my gun? Is this my god? Apparently it says “is this mic on?” but that was lost on me.
All 421 voices in 67 languages within the system are getting a new voice model and synthesizer. The current default voice in “English-US” is changing to one built using “fresher speaker data,” which, alongside other updates, results in a recognizable improvement from the previous default voice. You can also listen to how the updated voices sound in languages such as “Spanish-US” and “Brazilian-Portuguese.”
The update announcement says that folks already using text-to-speech tech don’t need to do anything to receive the new voices, as “everything will happen behind the scenes” with the updates being downloaded automatically. Its listing on the Google Play Store states that the service is already used by a variety of native applications such as Google Maps, Google Translate, and the Android Recorder app, so chances are that if you use an Android device, you probably already use the Speech Services by Google speech engine, even if you don’t know it. The update is rolling out over the next few weeks to all 64-bit Android devices via the Google Play Store.