What you need to know
- New voices powered by deep neural networks are now available in Microsoft Edge.
- The voices are meant to sound much more natural and less robotic.
- You can try the new voices now in the Edge Dev and Canary channels.
Microsoft today launched a new set of voices in the Microsoft Edge Dev and Canary channels that should make the "read aloud" feature sound much more natural. Responding to feedback that the current voices sound too robotic and it was onerous to install language packs to read other languages, the new voices are powered by deep neural networks in the cloud.
The voices are broken down into two categories that you can choose from, according to Microsoft:
- Neural voices – Powered by deep neural networks, these voices are the most natural sounding voices available today.
- Standard voices – These voices are the standard online voices offered by Microsoft Cognitive Services. Voices with "24kbps" in their title will sound clearer compared to other standard voices due to their improved audio bitrate.
You can distinguish these voices from the others by their names. The new voices are labeled as "Microsoft
[[ voicename ]] Online," Microsoft says.
You can try the new voices out for ourself by selecting text from a webpage, right clicking, and selecting the "read aloud" option from the context menu. From there, you'll have access to a read aloud menu bar where you can choose different voices and the speed at which they speak.
These voices are still in testing as part of the new Chromium-powered Edge, so there will likely be some hiccups. However, hearing more natural, hman-sounding voices should make the read aloud feature much more pleasant to use. Further, Microsoft has opened these voices up to developers as an API so they can use them in their own web-based text-to-speech apps.
If you have yet to give the new Chromium-based Edge a shot, you can download the Dev and Canary channels from the Edge Insider website to get started.
