I don’t believe speech and photo capture is supported in emulation for the editor. There is a plan to to get this supported in the editor, unsure of the dates for this work.
Hmm, the Windows.Speech libraries do work for the most part. Just the DictationRecognizer fails.
Used to work in the past, so I’m unsure what the delta is.
It does seem to run for about 50 update loops before it fails.
I’ve even tried initializing and starting it well after the scene has started using a button.
Interesting. This has been a reported issue that we’re still working to resolve with Microsoft, but only affecting 64-bit builds of Unity. If you’re seeing a failure with 32-bit builds (I’m assuming so, since you’re posting here on the HoloLens sub-forum), this might be a new issue we haven’t seen reported.
Hey Wesley, changing the timeout didn’t seem to fix this issue.
Pfreese, no worries, I’m currently using the 64 bit editor. Sounds like a library didn’t get updated properly or something. If it’s being compiled from C++ don’t see why x86 would work and x64 fail, unless there’s some dependency on an x86 process.
I got a better description of the Error.
Error Description ERROR: StartAsync on the continuous recognition session failed unexpectedly! [Operation has failed with error 0x80045509: (null)]
[Edit]
Forgot to change windows settings.
If someone is getting this error be sure to check out this link.
After updating my windows settings I’m just getting a failed status from the dictation recognizer but no error codes.
Unity 5.5.1p3 both 64 & 32 editor builds fail.
Running into this same issue, and from what I’ve read, it sounds like a fix is all up to Microsoft. Can you guys verify that? Also, any advice on alternate speech recognition tools for the time being?
This error seems to be fixed on the new Windows 10 update: windows 10 creator. (usually I like to wait before I update windows, but unfortunately had to update manually so I could continue working on my project)
I had to dig through some bug case links, but it appears this is a problem with user settings in the OS that we aren’t reporting in a helpful way.
The following was written up in a comment on a bug case internally when he looked into this issue:
“This issue is caused by Speech Dictation not being permitted in the users privacy settings. My guess as to why this regression occurred is that at some point, the default permissions in Windows 10 for speech dictation changed from enabled to disabled. Users can check their settings with Settings->Privacy->Speech, inking & typing. If the button text is ‘Get to know me’, dictation is not enabled. When dictation is enabled, the button text will changed to ‘Stop getting to know me’.”
Looking at the OS settings on my RS2 Windows 10 machine right now, it seems that the “Get to know me” and “Stop getting to know me” buttons have changed to “Turn on/off speech services and typing suggestions” instead, but I would assume it’s still the same issue.
I’ll ask QA to file a bug against this to make sure we report this in a far more helpful way than an general failure message.