AVCS4 Immersive Voice Control Radios for VoiceAttack
-
@SemlerPDX I’ve re-traced the steps to add/edit this voice command and still having this issue
-
@Atlas said in AVCS4 Immersive Voice Control Radios for VoiceAttack:
@SemlerPDX I’ve done that training more than 5 times now and my command is recognised each time I say it, just with a low confidence level. I don’t want to lower the confidence level of the recognition because it may then cause issues later on with other commands getting mixed up. Is there a way to improve confidence level for these phrases?
FTR, my good friend from Australia had to run through WSR training a dozen times before he found that “sweet spot”, of course this is due to his regional accent. I don’t want to assume, but if you have any regional accent, it may be helpful to train more. I happen to have a “Standard North American Accent”, and so was able to get by with just 3 training sessions.
Here are options you can try (in no particular order):
-Lower the (global) Min. Confidence Level to 30
-Train your WSR profile more times
-Edit WSR Dictionary, add custom pronunciation for “overhead” and “break”
-Erase this particular QCC command, open the profile, add a New Command following the same concept - to execute another command (by name), and adjust the Min. Confidence Level within this command only (to around 30)Using the final option would be the least desirable. The main benefit of QCC is that users can create new/alternate voice commands for this profile without having to edit the actual profile, and therefore this save file & voice commands can persist between different physical versions of AVCS4 BMS where users must download and import a new profile into VoiceAttack. This would otherwise lose any commands created manually within that AVCS4 BMS profile, and would have to be manually re-created or copied to the newer version of the AVCS4 BMS profile.
You can save and backup your WSR profile, and/or try training a new WSR profile from scratch, it is possible for a WSR profile to become messed up for one reason or another, though rare, and having a backup to restore from can really save time from not having to run through the training again.
Windows Speech Profile Backup Tool (on my Dropbox)
-
@Atlas said in AVCS4 Immersive Voice Control Radios for VoiceAttack:
@SemlerPDX I’ve re-traced the steps to add/edit this voice command and still having this issue
These commands are not “full commands”, they are part of the prefix/suffix set which are on the left side of the command reference radio menus infographic. All these commands follow a format of “[who] [what to do]”, and you’re having that QCC command try to execute this command only using the “[what to do]” portion and so it has no way to know “who” this is for.
These must be directed at a particular flight member (or the entire flight). They must be preceded by something like “Two”, or “Element”, or “Flight”, for example.
If you want to create a QCC command that is a wildcard for anything that ends with “rejoin formation”, then AVCS4 cannot tell which flight member you are talking to. You could add that to your ‘execute command (by name)’ if you want, but of course, then this alternate syntax wildcard command you’ve created would only ever apply to the Wingman, and never Element leader or the entire flight.
NOW - this profile IS totally customizable for users who want even more realism, so you can absolutely edit those config files to make them work in ways that I had not designed, but the tradeoff is that these would be moved to a “rollback” folder (named “config”) if I put out a patch update. You’d have to repeat this process after any patch updates (or tell AVCS4 BMS to use rollback configs, therefore losing any benefit of that latest patch update).
To accomplish this sort of customization, you would open the #1 config and the #3 config file(s) in something like Notepad++ (run as Admin). See image below - I have searched the #1 config file, the one for keypresses, and added this new phrase variant - this must be done for ALL the rejoins, so, for element & flight as well:
…and of course, in the #3 config (the one for command phrases), you would search for and edit the “Rejoin” command to add another option after the semicolon for “Rejoin Formation” - note that you must leave the original Rejoin, as shown in the pic below:
-
@SemlerPDX Thanks for your help! I don’t want to go too deep or get too complicated so I’ll try the simpler solutions first. Thanks for explaining the [who][what to do] bit as I looked at the commands and did not realise I was missing the [who] bit.
-
@SemlerPDX a while back, I came to you with a problem re: speech recognition and you said I can enter it as a phrase in my speech recognition software so that it gets recognised better. I need to scroll back quite a bit to get that info. Just wondering if you have that info on a page somewhere I can bookmark for easier access?
Just lazy me asking a question for now. Can’t wait to get this sorted this weekend for some flying and learning time next week. Current issue is I can’t talk to #2 as “two, rejoin” is recognised as “to rejoin” and therefore not issued as a command.
Thanks for any help.
-
-
-
@Atlas Thanks for the suggestion! I’ve gone ahead and added that content to the AVCS Wiki page here:
https://veterans-gaming.com/wiki/vg-labs-projects/avcs-wiki/avcs4-bms-manual/The command recognition phrases for wingman are inclusive of the word “to” as shown here:
I’m not sure what could be interfering with these, but you can enable the VoiceAttack event log to show the confidence level of all recognized (and rejected) command phrases. The default confidence level is 50.
If you can take a screenshot of the VoiceAttack event log showing “to rejoin” as unrecognized with proper confidence above your minimum confidence level, we could consider that a problem worth troubleshooting. I cannot reproduce that:
(I culled the “two” from the prefix command for this test, and “to rejoin” worked just fine for me…) -
@SemlerPDX You are a star! Will give this another go on the weekend and will report back!
-
@SemlerPDX Reporting back… added my other screen onto my setup so I could put VA there to monitor what it recognises. Otherwise, absolutely nothing changed… and it works. Rejoin, close up, etc. commands recognised with high confidence. I wonder if it is the mic boom placement that resulted in poor recognition before? But in any case, issue sorted for now, will see if any other gremlins poke their heads out.
Thanks!
-
EDIT: The “Reset” button on the Joystick Options “Test” page solved this problem. I could swear I tried that yesterday, but it didn’t work. Oh well, glad it worked today, and I’m glad VA and AVCS4 are back to fighting strength!
Oh, and I got the idea to try “Reset” again from this thread on the VA forum, to which SemlerPDX recently contributed:
https://forum.voiceattack.com/smf/index.php?topic=4219.0
The original problem:
I’m updating to BMS 4.37, and getting VoiceAttack to work is the last item on my to-do list. Unfortunately, I’m having trouble getting VoiceAttack to “see” any of the buttons on my Winwing F-16 throttle.
I’ve updated VoiceAttack, and I’m using AVCS4 Core 1.11 and AVCS Voice Control Radios 1.41.
I’ve been able to use a voice command to turn on PTT mode in AVCS4. I am hoping to use PTT in the “default” mode, where it only requires a PTT switch be actuated in order to do in-game commands over the virtual radio, like talking to wingmen, talking to ATC, etc…
I can’t set a PTT button, though. When I go into VoiceAttack’s settings (starting with the wrench icon in the lower-right corner of the VoiceAttack window), and then to Joystick Settings, I can Enable my Winwing base + F-16 throttle. But when I click “Test”, and VoiceAttack asks me to press buttons on that device, it does not see any of my button presses.
I was able to use this base+throttle in 4.36 with AVCS4 1.40, in PTT “default” mode, but I can’t get VoiceAttack to “see” this base+throttle any more.
I might be boneheading something, or maybe VoiceAttack has an issue with the Winwing device after the recent Winwing firmware update I did.
Just curious if anyone else is having the same problem…
-
@SemlerPDX When I use VA, my voice attack transmissions are repeated by the sim - both aurally and in text on the screen. I have noticed in some voice attack BMS videos, the sim does not repeat the command. How do you stop the repeat transmissions by the sim when using VA ? thanks.
-
It’s in your FalconBMS Setup.
Go to Setup >> Sound >> deactivate “Player Voice”.cheers
Magic -
@zap737 for the comms box being shown, use this line to disable them
“//
set g_bDisableCommsMenu 1 // If set, the comms menu (tower/awacs/wingman etc.) will no longer be drawn.”
It’s in the Config file -
@zap737 In addition to what was stated above, you can also move the Radio Answers from the screen to off-screen to hide it (afaik) by changing the size/position data in the
3DUiSettings.ini
file under\Falcon BMS 4.37\User\Config\
-
@Magic49th Thankyou @Magic49th, @SemlerPDX , @Quasi_Stellar
-
what am I doing wrong? I cant seem to make the commands work by using callsigns at the beginning of my phrase
look first command at the bottom gets recognized but the others above it with a callsign just dont work, I also tried other callsigns than hammer
-
@Pekins Absolutely anything said before the actual command of ‘taxi for departure’ will be discarded, and the actual command will be recognized. These are “wildcard” commands. The problem you are experiencing is a recognition issue - “exit for departure” is not a command. You can try enunciating a bit more.
If you have not trained your Windows Speech Profile (at least three times completely through the entire Speech Training session), that would be the place to start.
Check out this VoiceAttack page, the section a little lower down “Train your speech engine”
https://voiceattack.com/howto.aspx -
I have “Enable Audio Feedback” turned on, but I don’t hear the feedback sound when I issue a command. VA / AVCS4 is hearing and executing the commands fine, I just don’t hear the little sound. I heard it in BMS 4.36, but not since I switched to 4.37 and updated VA and the AVCS4 profile.
Just curious if there is a step I’ve forgotten, in order to hear the little sound. Otherwise, everything is working fine.
-
@Mylonite470 said in AVCS4 Immersive Voice Control Radios for VoiceAttack:
I have “Enable Audio Feedback” turned on, but I don’t hear the feedback sound when I issue a command. VA / AVCS4 is hearing and executing the commands fine, I just don’t hear the little sound. I heard it in BMS 4.36, but not since I switched to 4.37 and updated VA and the AVCS4 profile.
Just curious if there is a step I’ve forgotten, in order to hear the little sound. Otherwise, everything is working fine.
Please open your VoiceAttack Options menu (the wrench/spanner icon on the main window), and navigate to the Audio tab along the top.
Verify your audio settings including output devices, and volumes (defaults are around 50 afaik)
You should have “Integrated Components” selected as shown in the image below, unless you have some non-standard setup and have changed this purposefully to “Legacy Audio”
AVCS would visibly display errors in the Event Log if the audio files were not found and not properly played, so unless you see errors we could assume that all the paths are correct - barring that you could open the VoiceAttack program folder, and check the “Sounds” folder there at
%ProgramFiles(x86)%\VoiceAttack\Sounds\AVCS
(…unless you have VA installed through Steam or in another path)You should see the three audio files in that folder for LEFT, RIGHT, and CENTER. Under the
%ProgramFiles(x86)%\VoiceAttack\Apps\AVCS
path, you should see all the operating files and folders for my AVCS profiles including AVCS4 BMS.If any of these files/folders are not present, there should be errors shown in VA Event Log when launching my profiles, but again you can manually verify they are where they should be as stated.
Let me know if something was off in your Audio options - if you can hear text-to-speech but not sounds, this would be the place to start looking.
-
@SemlerPDX you say its a recognition/enunciation issue which is a fair conclusion when viewing that screenshot but as you can see when given the “request taxi for departure” command it works every time, it hasn’t failed me once. The problem tho comes when I add the callsign just before it. Of course I have tried multiple callsigns i.e Spade, Hawkeye etc… the result is always the same.
With that being said, I have been using this VA profile for BMS for the longest time, which doesn’t use any type of plugin, just pure VA, and it doesn’t have any issues recognizing the callsigns… only reason Im looking for an alternative is because you have to manually load the profile corresponding to the flight callsign every time. I am really intrigued and also kind of frustrated to find out what’s going on.
If anyone would be willing to troubleshoot this with me on discord you can add AlexGotcha#3455
-
@Pekins ,
I have found that VA / AVCS (for me at least) requires just a short pause between the “throwaway” part and the command part - less than half a second. For instance if I say “Gunsan Tower Mustang 4-1 ready for departure” the command may not be recognized. However if I say “Gunsan Tower Mustang 4-1…ready for departure” the command is recognized and acted upon. Give it a try. This is especially noticeable if you are a “fast-talker” by nature.Give that a try and see if you have better success with AVCS/VA recognizing your commands. You can say almost anything that doesn’t include a command, but just give it a brief pause before the actual command.
Hope this helps.
Regards,
Tomcattwo
(VoiceClone)7