AVCS4 Immersive Voice Control Radios for VoiceAttack
-
It does import without errors, everything breaks once I try switching to this profile. Once I do, VA hangs with high CPU/RAM usage and will not work upon restart until I actually clear the dat file in appdata (deleting the profile in process). I don’t even have any log files, it seems it never got chance to produce them. And it actually worked for some time (after I completely reinstalled VA, I would add) but then it ultimately broke again and I couldn’t get it to work since.
-
It does import without errors, everything breaks once I try switching to this profile. Once I do, VA hangs with high CPU/RAM usage and will not work upon restart until I actually clear the dat file in appdata (deleting the profile in process). I don’t even have any log files, it seems it never got chance to produce them. And it actually worked for some time (after I completely reinstalled VA, I would add) but then it ultimately broke again and I couldn’t get it to work since.
Understood.
I’m surprised there isn’t a fault log at all, just for the random things that happen over time with VoiceAttack, not involving my own profile, but other issues - if VA ever crashes under other circumstances, the file will exist. I can only assume in all the time you’ve used VoiceAttack that it’s never crashed – also, could be that a Steam Version has the file in a different location, and if that is the case, the reasons for all of this would be in that log.
Again, this is a very unique issue, there are at least a hundred people who have used this profile including myself, through various versions from the standalone to the Steam Edition, and this has not popped up yet. It must be something related to your OS environment, other software that may be running including anti-virus or malware protection, or the VoiceAttack installation itself. I’d still like to help you get past this issue, but if you cannot locate a proper VoiceAttack fault log file, I’m afraid only some of the moderators at VoiceAttack would be able to help you.
…
If you’re still interested in working through this, I’d suggest posting up this issue on the VoiceAttack forums in the Issues section, or pop on over to the VA Discord to the People-helping-People channel where many people are happy to help others with VoiceAttack. I’m sure they are familiar enough with my profile to know it’s not been created improperly or is unable to function within the VoiceAttack program as designed, so should direct you towards technical help regarding your VA installation, etc.Again, just to be clear - the SOP would be to download my profile from the download section on my website, to extract the files from the zip, and then to import the .vax package labeled “IMPORT THIS”. I assume you’ve done that, but just wanted to be sure you’re not trying to import the download .zip - apologies, sometimes troubleshooting questions can sound downright insulting, and that’s not my intent. Best wishes and good luck! I’ll be following any support issue you post on the VA website or discord to make sure I learn what went wrong for myself as well.
-
It did not crash indeed. It just freezes and maintains high CPU/RAM usage forever until I kill it. And I’m not retarded, I wouldn’t be able to get it working at all if I was trying to import the zip. It did work for some time until this stuff started happening again. I’ll try removing my speech recognition profile and regenerating it from scratch.
-
It did not crash indeed. It just freezes and maintains high CPU/RAM usage forever until I kill it. And I’m not retarded, I wouldn’t be able to get it working at all if I was trying to import the zip. It did work for some time until this stuff started happening again. I’ll try removing my speech recognition profile and regenerating it from scratch.
Roger that. Again, no offense intended, just asking questions. Also, removing the speech recognition profile may have nothing to do with this issue with no information pointing to that as an issue, but crazy things can happen in VA when the WSR profile gets borked. You may also want to backup that profile before hand so you could restore it if it wasn’t the issue - here’s a tool for the job: https://www.dropbox.com/s/9b6ffxxkm00bu2o/Windows_Speech_Profile_Backup_Tool.rar?dl=0
*re. crashes, I did not mean with this profile issue, I meant ever in the history of the time you’ve owned and used VoiceAttack, again, assuming this isn’t the first time/first profile you’ve been using.
-
I never had issues like those before. I used VAICOM for DCS (which is a plugin, not just a profile) and the UOAF profile for Falcon. That was before 1.8 version of VA came along though.
Resetting the Speech profile didn’t work, I’m still having the issue. I’m gonna try digging deeper.
EDIT: Upon digging deeper I discovered I had multiple config directories in %localappdata%\VoiceAttack.com. Deleting them and letting VA regenerate config seems to have fixed the issue for now.
EDIT2: Bingo! I found out what was triggering the issue. It gets triggered by checking “Allow Command Segment Info for Composite Commands”. That’s why deleting config fixed it. It seems that VA can’t handle a profile this complex with this option on, at least on my system. I’ll let it sit for some time and see if it unhangs, if it doesn’t I’ll just refrain from using this option.
-
I never had issues like those before. I used VAICOM for DCS (which is a plugin, not just a profile) and the UOAF profile for Falcon. That was before 1.8 version of VA came along though.
Resetting the Speech profile didn’t work, I’m still having the issue. I’m gonna try digging deeper.
EDIT: Upon digging deeper I discovered I had multiple config directories in %localappdata%\VoiceAttack.com. Deleting them and letting VA regenerate config seems to have fixed the issue for now.
EDIT2: Bingo! I found out what was triggering the issue. It gets triggered by checking “Allow Command Segment Info for Composite Commands”. That’s why deleting config fixed it. It seems that VA can’t handle a profile this complex with this option on, at least on my system. I’ll let it sit for some time and see if it unhangs, if it doesn’t I’ll just refrain from using this option.
Oh boy, that would just about do it. Likely less an issue with your system, and more a limit with the 32-bit VoiceAttack application, combined with the prefix/suffix use in AVCS4.
I suppose you had a profile that was using that or needed that? I will make sure to inform people about that on the download page, readme, etc. and I put a note on this reply to hopefully help anyone searching through this post in future.IIRC that’s an option for rare use case scenarios that Gary added a ways back to help simplify prefix and suffix command words evaluations, and I use much less resource intensive methods in my AVCS profiles. I have 3 prefixes each with over 1,500 potential combinations, and that’s before all the non-wildcard radio commands themselves (if it’s to a flight member, it’s a prefix/suffix combo - if it’s to the AI like tower/awacs/tanker, it’s a wildcard command). It’s actually best to not get into prefix/suffix use when it can be avoided and when CMDSEGMENT evaluation is heavily needed; each method has it’s purposes and using a large number of prefixes for a rather smaller number of suffix commands that all use that command starting phrase is about the only time it’s the best method to use over all the ways you can do things in VA, especially for dynamic multi-part composite commands.
A few commands with a few thousand options, or a several hundred commands with a few options…
I’m big on using the fastest methods for items, it’s why the eval for playing Audio Feedback exists in each and every command instead of just a single ‘function’ command that gets called from any command. These types of combos are also great for switches/toggles/on and off commands that all share those sentence structures in a dynamic multi-part composite format, and using methods other than CMDSEGMENT to pass the type of prefix used to the suffix with the body of the command actions.
-
It’s not really necessary for me, I’ll just keep it off. I’ve kept VA running to see what happens and it never recovered and eventually crashed when I suspended and woke up the system, this time leaving a fault log behind:
29/02/2020 16:58:10*************************************************************************************** System.InvalidOperationException: Collection was modified; enumeration operation may not execute. at System.ThrowHelper.ThrowInvalidOperationException(ExceptionResource resource) at System.Collections.Generic.List`1.Enumerator.MoveNextRare() at System.Collections.Generic.List`1.Enumerator.MoveNext() at VoiceAttack.frmMain.(String , Boolean , List`1& ) at VoiceAttack.frmMain..() at System.Threading.ThreadHelper.ThreadStart_Context(Object state) at System.Threading.ExecutionContext.RunInternal(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean preserveSyncCtx) at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean preserveSyncCtx) at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state) at System.Threading.ThreadHelper.ThreadStart() ***VoiceAttack v1.8.3 - Windows 10.0.18363.0
So far the profile works really well, though I have doubts about the PTT mode not completely disabling/enabling listening, since Speech gathers training data all the time when it listens. Doesn’t seem to be a problem so far, we’ll see how it goes.
If I were to suggest something: add alternative syntax for “Report Overhead Break”, something like " <number>in a break". While the idea of having the profile work just by reading the radio menu options out loud is very convenient, some of those options are clearly descriptory rather than showing what the message would be like, and this is an example of that.
EDIT: Plugin support also works fine now, so there shouldn’t be any problems with VAICOM after all, at least unless it craps itself without CMDSEGMENT.</number>
-
It’s not really necessary for me, I’ll just keep it off. I’ve kept VA running to see what happens and it never recovered and eventually crashed when I suspended and woke up the system, this time leaving a fault log behind:
29/02/2020 16:58:10*************************************************************************************** System.InvalidOperationException: Collection was modified; enumeration operation may not execute. at System.ThrowHelper.ThrowInvalidOperationException(ExceptionResource resource) at System.Collections.Generic.List`1.Enumerator.MoveNextRare() at System.Collections.Generic.List`1.Enumerator.MoveNext() at VoiceAttack.frmMain.(String , Boolean , List`1& ) at VoiceAttack.frmMain..() at System.Threading.ThreadHelper.ThreadStart_Context(Object state) at System.Threading.ExecutionContext.RunInternal(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean preserveSyncCtx) at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state, Boolean preserveSyncCtx) at System.Threading.ExecutionContext.Run(ExecutionContext executionContext, ContextCallback callback, Object state) at System.Threading.ThreadHelper.ThreadStart() ***VoiceAttack v1.8.3 - Windows 10.0.18363.0
So far the profile works really well, though I have doubts about the PTT mode not completely disabling/enabling listening, since Speech gathers training data all the time when it listens. Doesn’t seem to be a problem so far, we’ll see how it goes.
If I were to suggest something: add alternative syntax for “Report Overhead Break”, something like " <number>in a break". While the idea of having the profile work just by reading the radio menu options out loud is very convenient, some of those options are clearly descriptory rather than showing what the message would be like, and this is an example of that.
EDIT: Plugin support also works fine now, so there shouldn’t be any problems with VAICOM after all, at least unless it craps itself without CMDSEGMENT.</number>
Good to know there is a viable fault for this issue.
Regarding PTT mode – again, use that link I provided to manage your Speech Profile and backup a working version you can fall back upon, and periodically created a timestamped new profile if you use adaptive recognition, so you have backups to fall back upon (as opposed to falling back to the very first backup). I’ve been working heavily with VoiceAttack for almost 5 years. For a few years, I let adaptive recognition keep running, and eventually, it got pretty borked out. Common phrases just had too many recognition errors, and I had to fall back.
I turned on adaptive recognition again, but backed up periodically, and then fell back again when issues arose, to a recent profile. Then I turned off adaptive recognition. I keep recognition confidence set at 50% now and everything fires perfectly. In my personal experience, adaptive recognition should not be used indefinitely for a number of reasons, and including grunts or sounds that get recognized as words often enough to become regular recognitions that can cause even further issues especially with a few single word commands that may be in use.Try to think about the long run for using VA - get your profile trained through Windows Speech Training a full 3 times, and either freeze it there by turning off adaptive recognition, or at the very least, backup the WSR profile so you don’t have to do WSR training again unless your voice changes or you get a new microphone that is different enough to cause issues with recognition. And if you leave adaptive recognition on, backup your WSR profile regularly and be prepared for the date when you’ll want to turn this option off for good.
https://www.dropbox.com/s/9b6ffxxkm00bu2o/Windows_Speech_Profile_Backup_Tool.rar?dl=0
The optional Push To Talk method is designed to be an option for those who do not already use a PTT mode for VoiceAttack but want/need one for the AVCS4 BMS profile only – you can most certainly craft your own overarching PTT mode that actually disables the Listening Mode in VoiceAttack – but to try and limit everything that VA hears (throughout time) in an effort to ensure speech training doesn’t get poorly trained is much more labor intensive than just letting it listen, and falling back to a WSR Profile backup if things ever become problematic.
Regarding alternate syntax commands: Tools and examples have been provided inside the profile for alternate syntax command creation, and I’ll add your suggestion to the already existing alternates, too. Basically, just duplicate any alternate syntax command examples already present in that category, or follow the formatting of one of the examples to create a new command in the correct category… All these do is call the main command by name, so inside it would be an “execute command by name” and for this example would be “Report Overhead Break”
Expand the category named: AI COMMS 3 - ATC 3 - WILDCARDS (alternate syntax)
Copy any of the commands there (duplicate) and change the when I say (and description) to your desired usage so, maybe ‘[1…4] [in;on;an;anna] [a;] break’ if you want to get a little free on the catching of this phrase
Change the Execute action to: Execute command, ‘Report Overhead Break’ (by name)There’s no way I could plan for every alternate syntax or preferred method, so I made sure it’s easy to create these items as needed. I should really poll the users to see if they have any useful alternates that should become a permanent part of AVCS4 for everyone. For now, it should be easy to modify or add to the profile in any way. The next version (coming soon, I promise) will include a voice command builder system that should allow such command additions to be created with a much more simple GUI for merely inputting the new desired “when I say” phrase - but for now, the only option is the manual examples I provided for these new alternate syntax commands.
Additionally, the method for adding additional callsigns in the current version is far too technical, but the next version will provide a GUI for adding and saving additional callsigns to the list.
The wildcard use lets us pretty much say anything before the hardset commands, and adding alternates lets us expand that even further.
-
Hey question, what does “Audio Feedback” do? By default, its turned off.
-
Hey question, what does “Audio Feedback” do? By default, its turned off.
Audio Feedback Mode provides an audio cue when a command is successfully recognized. The sound is like a radio cue-up static sound, and can be set to Stereo Left/Right/Center as desired.
When VA is working perfectly and we have max recognition, it’s almost not needed…. but for those frustrating times when you say a command and seemingly nothing happens, and you feel like repeating yourself to be sure the VA Command fires, this option gives that “command recognized” feedback without breaking immersion… a very light radio-like sound that you can distinguish from the game or other sounds.
It lets you know your voice control is operating without needing to see the VA window and it’s event log of recognized phrases/commands, or lack-thereof.
-
Not familiar with how this profile works, but is it safe to rebind default keys ( ie. rebinding q for awacs page, to z)? Or are the commands bound to certain key inputs ?
-
Not familiar with how this profile works, but is it safe to rebind default keys ( ie. rebinding q for awacs page, to z)? Or are the commands bound to certain key inputs ?
All the keys used by this profile are defined in the KEYPRESS_VARIABLES command – it’s greyed out. If your keyboard is a non QWERTY, like AZERTY or QWERTZ. these options are available on voice command “Change Default Keyboard Layout”
Otherwise, if you set non-standard keys in BMS, you’ll want to change the corresponding keys for that command through the KEYPRESS_VARIABLES init command. (These options are part of a voice command config editor in v1.5 – coming soon, I promise!!)Basically, everything the profile uses to function is defined in those init commands that are mostly greyed out - they’re called when the profile is loaded, from speech spoken by the profile to the keys it uses, the time a key is held down and then released and the time between keypresses for macros. Any questions, feel free to ask! Cheers!
-
Amazing, thank you for this profile.
Question: every once in a while I get “no keys released”. What is the cause and is there anything I can do about it?
Still, keep up the great work!
Tnx
Bikeman -
Amazing, thank you for this profile.
Question: every once in a while I get “no keys released”. What is the cause and is there anything I can do about it?
Still, keep up the great work!
Tnx
BikemanI assume that is an action message due to some command that cleans up any keys that are held down. If the profile functions, and there are no other errors or issues, just ignore it. I do try my best to make sure the event log is not populated by needless notifications such as that, and my profile’s methods are specifically designed to work around actions that would throw an error message during a ‘false’ evaluation or such.
Just ignore it, and if you have any issues, feel free to post up and I’ll help. :drink:
-
Thanks for this profile, it’s excellent - I can’t get push to talk to work though.
I’ve it on (enabled).
I’ve assigned joystick buttons correctly in the profile and commands are executed in voice attack when I press the buttons.
Is it supposed to enable listening in voice attack? I have it off by default, if it’s on, push to talk doesn’t really work, as every command is listened to and processed nonetheless. -
Thanks for this profile, it’s excellent - I can’t get push to talk to work though.
I’ve it on (enabled).
I’ve assigned joystick buttons correctly in the profile and commands are executed in voice attack when I press the buttons.
Is it supposed to enable listening in voice attack? I have it off by default, if it’s on, push to talk doesn’t really work, as every command is listened to and processed nonetheless.The PTT in AVCS4 is a sub-mode and does not have anything to do with the overall Enable/Disable Listening feature in VoiceAttack. I totally get that the concept is a bit odd on the surface, don’t feel at all bad for not ‘getting’ it on first look. I think I’ll include a better infographic on this mode for future releases.
Think of it as button(s) you hold down so that AVCS4 Voice Commands can function - it does NOT affect any other voice commands. This design allows you to have your own profiles with other voice commands that are NOT restricted by my AVCS4 PTT mode. My PTT is only designed to stop AVCS4 BMS Radio Commands from firing IF the button(s) are not held down.
Inside each AVCS4 voice command is a check that boils down to this: “IF PTT MODE ON, and IF PTT BUTTON(s) NOT PUSHED DOWN == EXIT COMMAND”
I hope that makes sense… If not, feel free to reply and I’ll take another swing at it.
*Pushing hard to get the new version of AVCS4 into public beta test before June, but life keeps getting into the way - Spring comes early in Oregon, had to get the veggie garden going and the old rototiller refuses to RAMP start … spent my week tilling a 10x10’ bed by hand. Fun fun. Tons of stuff is now on voice command, live profile updates and voice control bug reporting, and better user personalizations from adding new callsigns (or removing some), even adding to or editing the included voice commands are now on voice commands themselves, so you can edit a profile without ever even opening the profile to “edit” in the traditional sense. Voice Calculator is cool - use it all the time during development… in my standard AVCS4 “Any Way you Say It” fashion, so for example, “Tell me the square root of Pi and be quick about it!”
…back to work. Cheers!
-
Dear all,
I have enjoyed this warthog profile (Morphine’s Profile for BMS 4.33 and Beyond by Gastone) since years. Congrats again for the great work.
I have been now using VoiceAttack and the great AVCS4 profile but I am struggling with using the push-to-talk linked to VHF/UHF. When pressing these, there is no reactions from VoiceAttack. (It does work when I use a specific PTT button for this (iff in))
Does anyone use this profile with AVCS4 and does it work with the UHF/VHF function ?
-
Does anyone use this profile with AVCS4 and does it work with the UHF/VHF function ?
I use this profile also. Just go into the settings (little wrench bottom right) select the hotkeys page and there you can set the PTT to the UHF button on your throttle. I did that and it seems to work fine
-
I use this profile also. Just go into the settings (little wrench bottom right) select the hotkeys page and there you can set the PTT to the UHF button on your throttle. I did that and it seems to work fine
Thanks for your feedback. However, in multiplayer, this wouldn’t be a good solution as if you do so, when you would need to control your wingman, it would transmit on the UHF channel and everybody would hear it. In that case, you should have transmitted on to VHF.
Until now, I did use the IFF in on the throttle to speak to Voiceattack without emitting in Multiplayer. But this is not ideal as I would rather the multiplayer people hear my voice when i speak on UHF, but not when I speak on VHF (while giving orders to voiceattack)
-
Thanks for your feedback. However, in multiplayer, this wouldn’t be a good solution as if you do so, when you would need to control your wingman, it would transmit on the UHF channel and everybody would hear it. In that case, you should have transmitted on to VHF.
Until now, I did use the IFF in on the throttle to speak to Voiceattack without emitting in Multiplayer. But this is not ideal as I would rather the multiplayer people hear my voice when i speak on UHF, but not when I speak on VHF (while giving orders to voiceattack)
You do not need to set BOTH buttons for PTT, if you don’t want to. You can use either PTT button. Please view the reply just above your first post here on Page 7 of this thread (click here), it explains the PTT mode in plain examples.
If you don’t want to transmit on UHF, then why are you not just using the VHF for PTT? Remember also, this is an optional system – and it is turned OFF by default. If this mode is ON, then you have to push one of the PTT buttons you set for your controller in VoiceAttack and in both PTT Push/Release commands (VHF and/or UHF) in order for a command to fire (they will be recognized, but exit any AVCS4 Command if PTT ON and PTT BUTTON NOT PRESSED)
This Radio System is designed to allow Immersive Control of AI Wingmen through the in-game Radio Menus. If a Human is your #2 Wingman, but an AI Computer Controlled Pilot is #3 Element Leader, then AVCS4 allows you to set your #2 Wingman as “Human” –using the command “Set a Human Pilot” and assigning #2 Wingman as Human-- during the flight, if you say, “Two Say Fuel”, the in-game Radio Menu command will NOT fire (because you set that pilot to Human). But if you say, “Three Say Fuel”, it WILL fire because it assumes all pilots (in this example, #3 Element Lead) are NOT human.
Other Human Pilots do NOT need to hear you speak to your Computer Pilots such as #3 in the example above, but AVCS4 lets this occur for immersion and you can use whatever PTT button you feel is appropriate (VHF or UHF), or not at all (PTT MODE OFF).
Use VHF or UHF as you wish, it doesn’t matter to AVCS4, or BMS.