Quote:
Originally Posted by kovidgoyal
Released version 7.17.101. With the major feature being a new text to speech engine for the viewer Read Aloud feature that uses a neural network for more realistic sounding speech. The neural network is run locally on your computer no cloud services are used.
Additionally, the content server viewer's read aloud feature now support word tracking on some browser/OS combinations, such as chrome/firefox on windows, safari on macOS, firefox on android, and probably (not tested) safari on iOS.
|
Wow, Brilliant!
I did some tests on Viewer and Content Server, everything working fine overall. Pausing doesn't seem to work all the time though, I'd have to restart the viewer to get TTS running again.
On the Content Server I see the word tracking feature working precisely and pausing works fine too. (perhaps because it's using built-in system voices)
I like the wide range of English voices and they're incredibly realistic.
I've also experienced some delay on word tracking in the Viewer, either in flow or paged mode, after "turning pages" sometimes.
I did the attached video captures, one of which captured the delay I mentioned, though I should have left the recording go on for a few more seconds.