Comment 1 by camlorn on 2013-05-27 01:16
These voices are only available on the Android unless I am very much mistaken; no windows version exists. It is possible to gain access to them via an unsanctioned hack as part of the Google translate interface, but this would require downloading a sound every time something is to be said. Ignoring potential legalities, this would lead, in the best case, to a delay before we could start streaming and, in the worst case, a few seconds because your internet connection is too slow for that approach. This, of course, adds the complexity of determining if your internet connection is fast enough in the first place, or limiting such a feature to those with good internet connections; in this latter case, good internet connection must be determined by the user, not NVDA.
I'm not 100% on my facts so am leaving this open, but suspect that this would be a huge undertaking; I looked into doing this for my own projects once, and this is what I came up with. I hope the above information is useful.

Downloading speech synthesis data from Google Translate simply for the sake of using Google text to speech is definitely not something meant for NVDA core. Regardless, this may also not be legal. @jcsteh Could NV Access's contacts at Google be notified about the interest of NVDA users in porting Google TTS to Windows?

We actually had some discussions with Google about this a few years ago, but despite initial hopes, this turned out to be infeasible for technical reasons. The technology they use to compile the TTS engine for Chrome for Windows is Chrome specific. Closing as can't fix.