Microsoft: Win10/11 to deprecate speech recognition?

[JackoJ]

At the end of next year, many functions are to be removed from Win10 and Win11 through an update, including "Windows Speech Recognition".
See https://learn.microsoft.com/en-us/windows/whats-new/deprecated-features

What does this mean for VoiceAttack?

[SemlerPDX]

Just read the link you shared, not sure where you got the impression that being deprecated and replaced means that it is being removed.

Windows speech recognition is deprecated and is no longer being developed. This feature is being replaced with voice access. Voice access is available for Windows 11, version 22H2, or later devices. Currently, voice access supports five English locales: English - US, English - UK, English - India, English - New Zealand, English - Canada, and English - Australia. For more information, see Setup voice access.

You can learn more about feature lifecycles and these terms here:
https://learn.microsoft.com/en-us/windows/whats-new/feature-lifecycle

You can find features that have actually been removed here:
https://learn.microsoft.com/en-us/windows/whats-new/removed-features

[Gary]

As SemlerPDX mentioned, the built-in speech recognition for Windows is being deprecated - that means that they are no longer going to be developing anything new for that feature going forward.  To be straight up, I don't believe anything has actually been done for those features in probably several years anyway.  Since the built-in speech engines are 'fine' as they are and do everything that they need to do for VoiceAttack's voice features, I don't believe that to be problematic (although updates are always really nice).

There is no mention of removal of the built-in speech services at this time, which is great for the folks that continue to use the built-in speech engines.  That said, if Microsoft decided to suddenly abandon everyone and completely pull the built-in speech engines, the installable speech engines will always be available for VoiceAttack users (https://voiceattack.com/alternate).  I don't get the feeling that Microsoft is going to do this, as System.Speech is still very much available in .NET 8 (current, long-term support framework).

On a side note, I do have roadmap items in the coming months to integrate some form(s) of online voice recognition into VA, starting with dictation services, and maybe some AI fun to boot.  I've dragged my feet over the last few years in regards to integrating online/server speech recognition as I have found it relatively slow compared to the local installed speech engines (by relatively, I mean excruciatingly).  However, offering an online alternative will allow for additional functionality, with the hope of speed increasing over time.

Hope that answers your question!

[Gary]

Also, where are we getting, 'End of 2024' from?  I didn't see it.




EDIT - I changed the title of this thread to reduce confusion.

[JackoJ]

I got the information from this German article
https://www.computerbase.de/2023-12/rueckzug-aus-der-vr-microsoft-stellt-windows-mixed-reality-ein/
End of Windows Mixed Reality
"Microsoft quietly announces the end in the "deprecated features" overview in the developer portal. In addition to Windows Mixed Reality (WMR), the following other Windows 10 and 11 functions are also about to be discontinued:

WMR for Steam VR
WMR app for Windows 10 & 11
Microsoft Defender Application Guard for Edge
Legacy console mode
Windows speech recognition

The functions will be removed in a "future version of Windows", the company writes. Microsoft does not specify a date, so the withdrawal could also take place with Windows 12 next year."

Other sources that do not so clearly address the end of speech recognition but only the obsolescence:
https://learn.microsoft.com/en-us/windows/whats-new/deprecated-features
https://learn.microsoft.com/en-us/windows/whats-new/removed-features
https://learn.microsoft.com/en-us/windows/whats-new/feature-lifecycle
https://learn.microsoft.com/en-us/windows/whats-new/feature-lifecycle#features-removed
https://www.uploadvr.com/windows-mixed-reality-headset-support-end-date/

As I own the WMR HP Reverb G2 VR glasses, these are also articles specifically for Windows Mixed Reality.
The first stage for speech recognition could also be deprecated, the next the removal of these components.
See "Terminilogy": https://learn.microsoft.com/en-us/windows/whats-new/feature-lifecycle#features-removed

[Gary]

Thank you for providing that.  Those are good articles, but mostly deal with WMR.  I see the one line in the first regarding speech recognition, but it doesn't relay anything more than what Microsoft had indicated in their article about deprecation.

Again, worst case scenario - if online speech engine options end up being the only speech engines available out of the box, VoiceAttack will be adapted.  Speech Platform 11 engines will still continue to be available as a local alternative if speed is an issue. 

The bottom line is that no matter what Microsoft decides to do, VoiceAttack will continue to evolve as it's just as much my application as it is everybody else's (I am a user the product that I am building).  Hope that helps!

[Xileh Nori]

I just got a pop-up message in Windows 11 that Windows Speech Recognition would be uninstalled in a future update, a link to a URL stating that Windows Speech Recognition would be EOL Dec 2024 and that I should use Voice Access

https://support.microsoft.com/en-us/topic/use-voice-access-to-control-your-pc-author-text-with-your-voice-4dcd23ee-f1b9-4fd1-bacc-862ab611f55d

Will this break Voice Attack? as I just purchased it on STEAM

[Gary]

Thank you for the link.

As I was saying earlier, VoiceAttack will adapt to what Microsoft puts out as a solution, with the worst case scenario of requiring users to install a Speech Platform 11 engine as a local speech recognition utility.  What that would entail would be two small additional files, which are currently already available here:  https://voiceattack.com/alternate

[Gary]

I received confirmation from my contact at Microsoft that only the Windows Speech Recognition application itself will be replaced, and not the underlying APIs that handle speech recognition. 

I do want to say that the wording could use a little bit of tweaking o_O

Hope that helps!

[Cliff3D]

Dragon Naturally Speaking, by a company called "Nuance", has long been considered the best voice recognition software available.

With that in mind I thought the following link might be of interest:

https://news.microsoft.com/2022/03/04/microsoft-completes-acquisition-of-nuance-ushering-in-new-era-of-outcomes-based-ai/

With Microsoft now owning Nuance they might be considering switching to a new version of Dragon in place of their existing engine "at some stage". Other interpretations are available.