Google Upgrades Text-to-Speech Voices on Android
Google is rolling out an enhanced set of models for its Speech Services by Google to make the Android app with text-to-speech voices easier to understand and more human-like. The improved clarity and more qualitative “human” improvements can be heard in the comparison below.
|Sample Current Speaker||Sample Upgraded Speaker|
The new models will be used for all 421 voices in 67 languages offered through Speech Services by Google. The upgrade will happen on the backend, so developers using text-to-speech and other voice services won’t have to change anything. Google will include the new voices automatically when users download the latest update for any 64-bit Android device from the Google Play Store. Plenty of native Google apps will deploy the new voices too, including Google Maps and Google Translate.
“We are upgrading the Speech Services by Google speech engine in a big way, providing clearer, more natural voice,” Google staff software engineer Rakesh Iyer and group product manager Leland Rechis explained in a blog post. “We’ve seen a significant side by side quality increase with this change, particularly in respects to clarity and naturalness. With this upgrade we will also be changing the default voice in en-US to one that is built using fresher speaker data, which alongside our new stack, results in a drastic improvement.”
Google’s ongoing improvements to its speech synthesis models are crucial as more and more companies look for advanced human speech synthesis as part of their software for both consumers and enterprise clients. Companies are eager to boast of improvements and new features, as exemplified recently by Nvidia’s upgraded Riva synthetic speech engine, WellSaid Labs’ upgraded voice models, and Neosapience rolling out an AI-powered tool allowing users to write out the emotion they want virtual actors to use when speaking. Reading websites, directions, or novels all require realistic sounding voices, and Google isn’t likely to forget that as it pushes out similar upgrades to its Speech Services in the future.
Follow @voicebotaiFollow @erichschwartz
WellSaid Labs Teaches Synthetic Text-to-Speech AI Proper Pronunciation
New Neosapience Tool Synthesizes Any Text into Emotion for Virtual Actor Speeches – Exclusive
Nvidia Riva Synthetic Speech AI Adds New Languages and Voice Recognition