Change Speech Recognition Profile

[Exergist]

Is it possible to switch between Recognition Profiles within VA? This would be handy especially for setups with multiple microphone configurations. So you'd train a Recognition Profile for each mic and then have VA switch to the appropriate profile depending on the recording device selected.

Thoughts?

[Gary]

I'm going to be bold and say that 99.999% of VA users have only one microphone that they use for speech recognition, and when they do switch devices for speech recognition it's probably to replace the previous device.  Also going to be bold and say that almost all VA users are using the default recording device.  The ability to set the device in VA is a shortcut to the Windows setting... it's also there to provide a way to quickly show the end user what device Windows is using.

It's a good suggestion, and I'll keep it in mind.  There's a lot of other stuff going on with the Windows speech profiles that would require VA to keep track of multiple things to keep it all straight.  If practically nobody is going to use it, it would be a lot of work for little if any use.

[Technical Ben]

I assume Windows has a separate trained Recognition profile for each Windows User login? So for multiple users, they could log in again. Can the OP, Exergist, use that option as a current work around?

[Exergist]

That would certainly be a possibility, although still inconvenient.

I use a wired over-ear headset and microphone at my PC but intend to use in-ear headphones (much less bulky) with inline mic connected to my XBOne controller when I stream games to my TV via SteamLink (effectively creating a semi-wireless controller/mic/headset configuration). Using this TV setup with VA the first time blew my mind 

Anyway it makes sense that these different mics would produce better results if each has its own recognition profile. Admittedly my use case seems like an edge case, but I figured I'd ask

[DruiD]

Hi,
Also looking for a way to "quick switch" the Windows voice Recog profiles.  I actually have a few, as I found mixing microphones under the same 'profile' disastrous to recognition quality.

1. Default [Profile] (wish I could rename it to Astro A50)
- Astro A50 headset - triple monitor gaming.    Voice trained at least 10 times, most voice packs (HCS) get a confidence level over 93% for nearly all commands - flawless victory    Gaming includes triple screen elite dangerous etc.

2. RIFT HMD Mic [Profile]
- Oculus RIFT HMD built in mic - VR gaming, Elite dangerous etc - but VA fails to work with profile #1,  thus need to retrain under the new different RIFT Mic - trained 4-5 times as a trial, Confidence level around 70-85,  a few 91/92 scores. So, if I VR game with Voice Attack - I need switch to recog profile #2, before loading VA or gaming in VR.   

3.  QCY Q29 [Profile]
- not really for gaming - small Bluetooth headset\mics- great for walking around my house dictating things \ controlling home automation through VA.   Might eventually replace with some Google home devices.

I also use a program called Sound Switcher to 'quick switch' audio devices, but I also get away with setting the RIFT as default mic\audio devices (when plugged in) - and then unplugging it .. and it drops back to Astro A50 as default mic\sound. etc

So trying to get a way for VA to switch recog profiles as I cannot find a quick key combination in Windows or other method of doing it yet. 

Workaround #1  - dont use RIFT headphones - remove them and put my A50 over my RIFT - yeah ok .. but not ideal.
Workaround #2 - manually modify windows control panel \ sound etc  each and every time - not ideal as often I quickly drop out of VR and back into it etc.

Any thoughts \ takers?

[Exergist]

FYI I'm working on a profile that contains commands (which leverage a C# inline function) for changing the WSR profile based on a variable as well as voice command. I need to do some more testing and consult with Gary though before I publish it.

[TheThingIs]

Gary, I'd actually argue that point with you. Sure most people do have only 1 mic but more and more people have VR now which tend to have their own built-in mics plus a normal headset and mic for non-vr games. I personally have to have a profile for my VR and my headset so easy switch would be great.

[Exergist]

In particular when it comes to VR I think more folks are likely to have multiple users, which again makes easy profile switching useful.

[Exergist]

So I believe I've gone as far as I can with testing my WSR profile change commands in Windows 7 and Windows 10. So far it appears as though the commands successfully change the selected WSR profile on both platforms. Now all that remains is for Gary to find some time to check out my code and see if he can tell from the VoiceAttack side of things whether or not VoiceAttack recognizes and properly acts upon the WSR profile change.

I'll keep my fingers crossed!

[Exergist]

Update!

I overhauled some of my code based on Gary's feedback so now I'm just waiting on additional thoughts from him on the (hopefully) nearly complete command set. Everything appears to be working, but Gary indicated that there is a small catch. In order for VA to "recognize" the WSR profile change you have to (at minimum) perform a VA profile switch after performing the WSR profile change. The ability to "reset VA profile" natively is on Gary's to-do list, but it's probably not a high priority. Another option might be to force VA to restart after performing the WSR profile switch.

Assuming everything checks out on Gary's end I'll upload the completed profile after some final adjustments.

[Exergist]

Finally released v2.1.0 of my VoiceAttack profile for changing the speech recognition profile! Feel free to check out the post here. So in the end, I guess I answered my own question