Windows 10: No speech recognition at all

(Feel free to skip the intro)
So I'm working on a c# app that uses Text To Speech and Automatic Speech Recognition, in Visual Studio 2015 on Windows 10.
I'm using System.Speech for Synthesis (TTS) and Microsoft.Speech for Recognition (ASR).
I can't use Microsoft for Synthesis because it will tell me there are no voices installed or available, and I can't use System for Recognition because, apparently, there is no Recognizer installed.

The first one (no voices with Microsoft) is just ridiculous, I have a couple of voices which all work fine using System.
But the latter might have a point:
TL;DR
I have no way of using the Windows Speech Recognition.
I've seen screenshots of people's control panel windows where there is a recognition tab next to the TTS tab. That is just completely gone for me:



I can't configure, run or condition Speech Recognition:




Speech recognition is not supported in the current language.
That is just BS, both de-DE and en-US are obviously supported.
Those are the language packs I have installed (MUI).

I've tried rebooting with English as display language, changing locale settings around, having the ugly language settings toolbar active and displayed, re-downloading the language packs, and a bunch of other things I can't remember just now.

Now, the settings work for my app as far as simple grammar goes, I can have predictive / pre-determined conversations / commands with command-style grammar.
However, to get free dictation to work, I need to use the System API for Recognition - and it looks like that won't work for the same reason the built-in Windows recognizer won't.
It's like that component simply isn't installed.

I couldn't find any disabled services that might be missing.

I've searched and tried for hours upon hours (after wasting hours and hours looking for a problem in Visual Studio)
So... Any help would be greatly appreciated!
What are the components that are involved in Speech Recognition (Windows 10)?
What might be the problem?
How might I go about fixing it?
Any ideas?

:)

 

I can't open speech referance card on WIN 10 can someone help me ?

When I open it from the System Tray icon running with Speech Recognition Desktop app, it does open

Windows Speech Recognition commands - Windows Help .

From the Speech Recognition Control Panel at Control Panel\All Control Panel Items\Speech Recognition it also opens from the link at bottom to same page.

Where are you finding it still doesn't open?

 

Speech recognition does not shut down

Hello



I have a bug with Windows Speech Recognition (SAPI) in Windows 10, version 1709, OS Build 16299.125. (issue was present in earlier builds of Windows 10 as well)



Speech Recognition will refuse to start again once closed if certain criteria are met.

I have placed a shortcut for starting Windows Speech Recognition on my Desktop.

Shortcut copied from C:\ProgramData\Microsoft\Windows\Start Menu\Programs\Accessibility\Windows Speech Recognition

- Click the shortcut to start Windows Speech Recognition

- Open Start

- Close Start

- Close Windows Speech Recognition

- Click the shortcut to Windows Speech Recognition



Windows Speech Recognition won’t open. Looking at Task Manager you’ll see Speech Recognition still running.

If you open Start, you’ll see Speech Recognition disappearing from Task Manager and you are able to open Windows Speech Recognition again. If you close Windows Speech Recognition without opening Start, you are able to start Windows Speech Recognition
again.

 

spent countless hours and still cant make it work. i have found that many users have the same problem, with little or no help available from microsoft. can anybody help?

 

Good to hear I'm not an isolated case. Be sure to come back here and share your findings if you found a solution; I'll do the same

 

thanx. will do

 

Hi, this may be of interest:
Installing additional language packs for Windows 10 speech recognition random ranting

This has downloads, + a complete list of languages availablefor Win 10:
https://www.microsoft.com/en-us/down....aspx?id=27224
(but note the date)

fwiw I just tried it- my locale, speech, display language etc are set to UK English. (Which took some doing this time, or something was slightly messed up).

If I type speech in Cortana's search box, say, I get Windows Speech Recognition as top search.

I use Classic Shell, with the Control Panel expanded, so in a second or less I can see 'Speech Recognition' listed there.

Similarly, if I type 'Speech' in Settings I can easily see Speech Recognition.
------------------------------------------------------------------------------------
With the speech recognition bar displayed, task manager shows:
SpeechRuntime.exe in C:\Windows\System32\Speech_OneCore\Common ..........

sapisvr.exe in C:\Windows\Speech\Common

running.

=================================================
Some (older) search results:

I came across this for Win 7
Speech Recognition - Language Configuration is Not Supported Error Fix - Windows 7 Help Forums
- maybe there's an updated one here (see Tutorials).

There are a couple of helpful links at the end of this (but Win 7 again)
http://answers.microsoft.com/en-us/w...c-cb0ae84c939d

 

Hi, thank you for the useful information.
Personally, I also use ClassicShell with CPanel enabled, and I do have the root "Speech Recognition" icon, but I can't open it, and I can't access the further settings (such as configuring ASR Language or training your mic) - and of course I've tried a million ways of accessing the settings *Wink
I did set my display language and my input language to en-US, as well as my locale, to no avail (and restarting didn't do the trick either) - and then back to de-DE (which is also supported) - the ASR language settings, however, remain elusive^^
I have also tried having ctfmon running and the input language bar displayed in various states

I have downloaded different language packs, but some of them would not install, the installer disappears after "gathering required information"

Thank you for sending the executable and service, that was exactly what i was looking for!
I'm not at home right now, I'll check in a couple of hours - but that looks immensely helpful already. I did find the speech folders in Sys32 and SysWOW myself before and tried changing access permissions, but to no avail.

I hope running the SAPI server will do the trick - I was expecting exactly something like this to be missing or blocked from/in my system

 

Hope you find a way, otherwise you're into standard 'fixes' like create new user (which you could do just to see if it works in a new user profile then delete the user if unwanted), sfc scannow etc etc.

However Win 10 can be very sensitive to the multiple steps needed to change language (curious- never had to worry about that in previous editions).....

 

It's solved... And pardon my ignorance.

So I went and took ownership and granted full access to everybody for all the speech related directories in %windir%, System32 and SysWOW64.

There are several speech libraries, at least in my case, natively in windows 10.
There is, for example, Speech, and then there is Speech_OneCore.
And then there are several synthesizers and recognizers available as well, some of them differ quite a bit.

Some of the versions are optimized for server-apps, some for client-apps.
Generally, the system.speech API is optimized for clients/Desktop, microsoft.speech is optimized for servers.
Then there are versions optimized for minimal use, streamlined for handhelds, etc.
And for everything, of course, you have to have the right language.

Now, after granting permissions didn't change anything, I went back to check options in the control panel, no luck.
I even turned on Cortana and configured her - of course that worked fine. Exactly what I don't need.

I then got annoyed by the language bar that had been hanging around since I added the en-US input language and wouldn't turn off even though I told it to.
So I checked the options, and they took me to that weird ugly "Settings" place that doesn't feel like it's part of Windows and like it just aggregates Control Panel functions and sometimes won't let you use your scroll wheel.
That one.
I've been there often before, on one occasion to specifically check for further downloads for my languages: Supposedly, you'll find input languages for download under System -> Apps and features -> manage additional features. Back then, there was only extended support for fonts available to me.
I figured "Well, I downloaded complete LIPs, the recognition package is probably installed already, that's why no further downloads show up". Little did I know that those are NOT part of a Language Interface Pack in windows 10 - and they might also not show up under additional features along with other language related downloads. Win10 just don't roll that way


TL;DR
Now, I clicked on one of the languages installed, and went to options.
Lo and behold, the following page finally offered a download link for voice recognition.
Time and Language -> Region & Language: under Languages, select your language -> Options: There should be a download button.

At first, I thought "Ah, well, that's the 100th thing I've downloaded now. Probably won't change a thing.
But, at long last, it let me access the Recognition settings and finally also use system.speech.

I'd just been neglecting the Settings panel the whole time, because usually there's nothing particularly useful / unique: Most settings are accessible through the good old Control Panel, where you can use your mouse wheel (on a big chunky desktop computer, like the cave man I am) - and regular language settings and downloads ARE available from there...

The problem for me, after my own ignorance, lay in a) the many many version of different engines; b) segmentation of what might be considered an "input language"; and c) lack of transparency and consistency within Windows 10: Which settings are there - and where on earth is "there", and misleading error messages and behavior.

Now, I'm sure not many people will run into the same problem as me - but for the chosen few fellow grumpy cave men: This here might help.
@chuckbowers: Are you one of them, too? *Smile

Good luck - and thank you everybody for your input.

 

Oh, PS:
For me it totally didn't matter what display language I had active, the recognition settings are handled separately, contrary to what the error message told me in the beginning "...not available for your display language", yeah right