Weather Talk Female Voice Gone?

I just upgraded to 10.37 S 23. I generate a wav file for my Asterisk PBX with a weather report every 15 minutes. The voice used to be female. Since upgrading the voice is now male.

When I look at Control Panel/Weather Talk, the Speech Driver window is blank. If I “Select Speech Driver”, David and Zira display. I have selected Zira and saved settings but Zira must be a male like David.

I am running Windows 10 with the “Anniversary Update”.

Update: Could this actually be an artifact of the Anniversary Update? I went into the Windows 10 Control Panel speech to text settings and made Zira the default voice. Now thiings are back to normal. So it needs to be determined why Weather Display can no longer change the voice driver from within the program.

Brian or anyone else, any thoughts on this yet?

I went into the Windows 10 Control Panel speech to text settings and made Zira the default voice. Now thiings are back to normal.
so you now have a female voice for WD, or?

Yes I do. However, it is no longer selectable or viewable from within WD. No voices are displayed at all in the select speech driver window. So you have to select the system default via the Windows 10 control panel.

This issue seems to have returned. I just upgraded to Win 10 ver 1709 (Fall Creators). I am running WD 10.37S62.

As I did last time this came up in 2016, I set the system default voice to Zira (the female voice). However, WD is stubbornly using David (male), disregarding the selection made in the Control Panel/Weather Talk section as well as the system defaults.

I am using Weather Talk to generate a custom wav file that gets sent over to my PBX for dialup weather reports.

Update: Windows 10 Fall Creators seems to have broken the text to speech settings. You can get to them via the Windows 10 settings or via the legacy control panel. I found that while I had set Zira as the system default voice for Windows 10, the legacy control panel thought it was still David. This is definitely a Microsoft Windows bug.

Now I get the correct voice. The voice selection within WD seems to be ignored, but it could all relate back to the finding above.