Text-to-Speech, Windows 10, "Microsoft James" Voice

[LinuxDevice]

I use English USA on my Windows 10. However, there are a number of optional voices available directly from Windows. I installed some interesting voices from Australia, Ireland, and England. I even set my default text-to-speech voice in Windows 10 to "Microsoft James" (reminds me of the AI in Iron Man's armor). I can preview that voice, and I can use it in Windows.

However, VoiceAttack does not see this voice. It is puzzling though because VoiceAttack does see a couple of voices which were not present before this:

• Microsoft David (the original)
  Microsoft Hazel
  Microsoft Zira

Why are the other voices not visible? FYI, I'm setting that via the "Voice" dropdown menu when editing the specific line of text-to-speech.

I also installed the Microsoft Speech Platform for 64-bit described at:
https://voiceattack.com/alternate.aspx

If I go into the actual toolbox for VoiceAttack, in the System/Advanced tab, I now have a radio button selector which can choose:
"Use Installed Speech Platform Text-To-Speech Synthesizers"
...if I choose this, it picks one voice (somewhat randomly) among the voices. This is not the Windows default. There are no choices.

If I click on the radio button "Use Installed Speech Platform Speech Engines", then I see the volume indicator showing the microphone is working, but nothing will trigger. Everything goes inert. The microphone is not showing as blocked, it simply does not work ("the lights are on, but nobody is home").

Is there a trick to getting access to the Windows 10 "Microsoft James"?

[Pfeil]

Microsoft chose not to make those voices SAPI compliant. This is not VoiceAttack-specific.

There is a method the should enable their use with SAPI (they are theoretically SAPI compatible), though it is not officially supported (credit to Exergist for the link).


Speech Platform 11 requires a language to be installed as well as the framework itself, and you need to make sure the correct engine is selected in the "Speech Engine" dropdown on the "Recognition" tab of the VoiceAttack options window. The "lightweight" engine is not actually usable.

Speech Platform 11 is wholly separate from the speech facilities built-into Windows.