AVCS4 Immersive Voice Control Radios for VoiceAttack
-
So, I played around a bit more with AVCS4, have since found a solution for global noise filter (which improved recognition in VA massively - the 10-20dB static noise I had was ruining it) but I noticed another issue, likely a limitation of VA itself (at least versions 1.8.5+) - the wildcards simply do not work anymore. Like almost at all. If I’m lucky I sometimes get a 20-30% confidence match but 99% of the time even if I set command weight to 99 and confidence to something like 10, it will still fail to recognize the command if there’s anything in front of it, throwing “unrecognized”. At 100% weight it literally fails to even evaluate them, showing no message in log. If I do not say anything before the command on the other hand, it almost always recognizes it with confidence above 90% at all times - which again, is it even supposed to work this way? Shouldn’t it drop the command if there’s a wildcard marker in front of it and nothing is said? This makes me think that wildcards might have been completely deprecated at some point recently (it’s listed as “unsupported” in VA manual after all). Maybe alternative methods could be used for this, like having a database of valid callsigns and recognizing the command if one of those is followed by player callsign and a command? The airbase names would be the biggest issue here, but maybe it’s doable if there’s a relatively fool-proof way for end users to make their own?
-
So, I played around a bit more with AVCS4, have since found a solution for global noise filter (which improved recognition in VA massively - the 10-20dB static noise I had was ruining it) but I noticed another issue, likely a limitation of VA itself (at least versions 1.8.5+) - the wildcards simply do not work anymore. Like almost at all. If I’m lucky I sometimes get a 20-30% confidence match but 99% of the time even if I set command weight to 99 and confidence to something like 10, it will still fail to recognize the command if there’s anything in front of it, throwing “unrecognized”. At 100% weight it literally fails to even evaluate them, showing no message in log. If I do not say anything before the command on the other hand, it almost always recognizes it with confidence above 90% at all times - which again, is it even supposed to work this way? Shouldn’t it drop the command if there’s a wildcard marker in front of it and nothing is said? This makes me think that wildcards might have been completely deprecated at some point recently (it’s listed as “unsupported” in VA manual after all). Maybe alternative methods could be used for this, like having a database of valid callsigns and recognizing the command if one of those is followed by player callsign and a command? The airbase names would be the biggest issue here, but maybe it’s doable if there’s a relatively fool-proof way for end users to make their own?
Wildcards have not be depreciated, which would imply that they’d been replaced… They are 100% working, and an excellent solution for situations such as my AVCS4 BMS profile handles. I believe you may have created these issues by messing with settings that don’t need to be messed with. Standard confidence levels and weight at around 50 and 50 should be more than sufficient. What matters most for recognition is a good microphone, going through the Windows Speech Recognition Training a total of three times completely through, and enunciating your speech when giving commands to VoiceAttack.
A wildcard command will fire even if it is just the wildcard key that was spoken, meaning you don’t HAVE to say something before it, but you CAN say anything before it. I would suggest you monitor what it thought you said, and checking there to see what issues are preventing the proper recognition of the command. This time of year, air conditioners can cause background noise, as well as other things - best to start at the beginning and troubleshoot why YOU are not getting your commands recognized while the rest of us are using it daily without issue.
Check out the images in this thread regarding troubleshooting, to enable the command each time any speech is unrecognized:
https://imgur.com/a/xuclI0a -
Is there any way to listen to what VA actually hears? Maybe it skips the noise filter?
EDIT: ok, I give up I tried rebuilding speech profile and it still won’t recognize wildcards. I guess it’s my audio. Can’t do anything about it though, I would need to purchase a new sound card and can’t afford it right now (the onboard soundcard is the source of the noise, doesn’t matter what mic I plug in it’s still there). I’ll stick to keyboard. Thanks for help anyway.
-
I understand your frustration. VoiceAttack won’t work properly for any profile unless it can hear us speak clearly - most modern motherboards have no issues, but sounds like you’re having some bad luck with your hardware there. Best wishes and good luck to you - hope a new sound card is the key, definitely the easiest and most affordable option, rather than gutting the motherboard (and possibly CPU/RAM and CPU cooler, too).
Is there any way to listen to what VA actually hears? Maybe it skips the noise filter?
Under the command “F_AVCS_VAS_INVALID_COMMAND”, add a line that contains a “Write a Value to the Event Log” action, using the token below:
Write [blue] 'VoiceAttack just heard: {LASTSPOKEN}' to log Make sure you have this command set to fire anytime a phrase is unrecongized (open the profile to edit it, the click 'Options' at the top left): [img]https://i.imgur.com/zKjrYNH.png[/img] as far as direct audio capture goes, recording your mic in any way should reveal what VoiceAttack and the Windows Speech Recognition hears....[/blue]
-
I was trying out the Raygun command on my AI Wingman. I was on UHF 6 and also tried the correct VHF frequency. Why was there not a Buddy Spike response?
Also VoiceAttack recognized the command “two Ragun” derived command, but what does the 2 unrecognized responses following the derived command mean?
LorikEolmin informed me in the General Discussion forum that Raygun should work with the correct VHF frequency and it does work with keyboard commands.
-
@jc1:
I was trying out the Raygun command on my AI Wingman. I was on UHF 6 and also tried the correct VHF frequency. Why was there not a Buddy Spike response?
Also VoiceAttack recognized the command “two Ragun” derived command, but what does the 2 unrecognized responses following the derived command mean?
LorikEolmin informed me in the General Discussion forum that Raygun should work with the correct VHF frequency and it does work with keyboard commands.
That is definitely an odd bug there … if you look in the KEYPRESS_VARIABLES command, scroll down, you can see the GVAR clearly listed as:
Set text [KEY_WINGMAN_Raygun] to '3'
…I have no idea why you are getting such errors… Restarting VoiceAttack and/or re-initializing the AVCS4 profile should fix that, definitely. No idea how the GVAR is getting lost… I have not been using this previous version for a long time, as I’m days from dropping a beta test for the next version which is a complete overhaul with no loss of functionality (and possibly far better stability in regards to GVAR errors like this, at the very least including live debugging of voice commands to create a bug report). If it were the new version, I’d simply ask you to run the “Create a Bug Report” command so I could see a live feed of all the actions that command is trying to take, and the keys it is trying to press.
As it is, your report is good enough to see that hard-set key variables are not being ‘remembered’ somehow, Raygun is no different from any of the many other commands - it’s a variable that resolves to a key. If that happens again, run the command, ‘Reinitialize Systems’, then when it’s done, try the command again (no need to close BMS, or anything). Let me know what happens. That screenshot really helps, too - thanks!
EDIT: AAHHHHH!!! I see it NOW!!! Ragun and not Raygun – holy hot hell, can’t believe that typo wasn’t caught until now!!! It is even in what would be the new beta of AVCS4 BMS (but in THIS case, I can change the value on the AVCS server powered through my website, and push the update to all users with a simple update prompt when loading the profile).
Solution: Open the actual voice command for Raygun and change the word “Ragun” to “Raygun” and all will work fine.
Beta for AVS4 BMS (1.3.9) will be released this week, following a quick round of private public testing with a few willing volunteers.
-
That is definitely an odd bug there … if you look in the KEYPRESS_VARIABLES command, scroll down, you can see the GVAR clearly listed as:
Set text [KEY_WINGMAN_Raygun] to '3'
…I have no idea why you are getting such errors… Restarting VoiceAttack and/or re-initializing the AVCS4 profile should fix that, definitely. No idea how the GVAR is getting lost… I have not been using this previous version for a long time, as I’m days from dropping a beta test for the next version which is a complete overhaul with no loss of functionality (and possibly far better stability in regards to GVAR errors like this, at the very least including live debugging of voice commands to create a bug report). If it were the new version, I’d simply ask you to run the “Create a Bug Report” command so I could see a live feed of all the actions that command is trying to take, and the keys it is trying to press.
As it is, your report is good enough to see that hard-set key variables are not being ‘remembered’ somehow, Raygun is no different from any of the many other commands - it’s a variable that resolves to a key. If that happens again, run the command, ‘Reinitialize Systems’, then when it’s done, try the command again (no need to close BMS, or anything). Let me know what happens. That screenshot really helps, too - thanks!
EDIT: AAHHHHH!!! I see it NOW!!! Ragun and not Raygun – holy hot hell, can’t believe that typo wasn’t caught until now!!! It is even in what would be the new beta of AVCS4 BMS (but in THIS case, I can change the value on the AVCS server powered through my website, and push the update to all users with a simple update prompt when loading the profile).
Solution: Open the actual voice command for Raygun and change the word “Ragun” to “Raygun” and all will work fine.
Beta for AVS4 BMS (1.3.9) will be released this week, following a quick round of private public testing with a few willing volunteers.
Your Solution above worked
-
v1.3.90b Full Overhaul & AVCS CORE Integration Update (beta)
I’m very proud to finally offer the next version of AVCS4 BMS because it includes an entire set of profile control and support from my new AVCS CORE Framework Profile for VoiceAttack. While AVCS4 BMS has not functionally changed much from what users have come to expect, AVCS CORE adds entirely new features that remove any previous need to manually edit any commands in the traditional sense to modify any configuration settings or otherwise.
The only significant new command added directly to AVCS4 BMS is “Edit the Callsigns List” (and phrase variations of it). This opens a simple Get Choice menu to offer options to add/change/edit/remove any callsigns recognized by the profile.
After that, most everything comes from the required AVCS CORE Profile Framework that will load before AVCS4 BMS, adding commands and features, as well as allowing auto-profile switching by setting AVCS4 BMS as a “Preferred Profile” in the CORE Options Menu (say, “Open Core Options”, poke around, have fun!)
Full details here: https://veterans-gaming.com/index.php?/semlerpdx-avcs/home.html/Please be sure to download the AVCS CORE 0.90b Profile from the download section as well, it is now a requirement for any AVCS4 game profile I release, such as the upcoming Elite Dangerous profile, and more…
BETA Release v1.3.90 Changelog Aug2020 New Commands: -AVCS CORE Commands Included into AVCS4 BMS Profile to lessen profile switching to CORE when it was needed -"Create a new Voice Authorization Phrase" and other included commands from Voice Authorization System -VAS can allow QCC commands to be locked behind a required authorization phrase as set by users, in up to 10 weighted levels -Integerated Bug Reporting and debugging options from AVCS CORE for easy troubleshooting -Versioning allows for Mutiple prior versions (reinitialization required) for easy profile version rollback if desired -Included UPDATE command to allow users to import previous saved user settings and saved Quick Commands from previous version of profile -Full profile control through command, "Open CORE Options Menu" -Better user support through command, "Open Help and Support Menu" (will expand in time) -User settings save file control through command, "Open Save File Menu" (includes save file backup and restore system) -...and much more (see posts at avcs homepage) Fixes/Improvements: -Set First Use to present user with Get Choice box for Default Keyboard Layout (QWERTY/AZERTY/etc..) -Setting 'CurrentBackupIndex' and 'SaveFileBackupCount' to 'Not Set' inside F_SFS_LOAD_ALL (was getting corrupted in save file) -Separated 'new_profiles.txt' file from AVCS4 profiles and AVCS CORE profiles to handle two imports before first time run/initialization of either -Added Duplicate Profile check from AVCS CORE, informs user of non-functional AVCS profiles (if ending in 'dash number' such as '-1') -Edited Preferred Profile command to allow for any user profile to be set to be auto-switched to after AVCS CORE finishes initializing ___ BETA Release v1.3.80 Changelog Mar2020 New Commands: -"Add/Edit Callsigns List" commands to add or remove callsigns recognized -"Add a new Voice Command" and others included from Quick Command Creator system - QCC can be used to create additional phrasings, just set action to execute another command (by name) -"Edit Keypress Timings" to open options menu to set custom timings Fixes/Improvements: -AVCS CORE Profile and Updates Management integeration (see AVCS CORE changelog) -Full Profile Update Control added with user choice, only downloads if allowed by user (only checks 8 bytes of a version number before that) -Profile Updates set to allow choice for majority of fixes/changes to occur without new VoiceAttack Profile download/import *called "LIVE UPDATES" -Physical Profile Updates set to allow previous versions to remain installed without conflict, and also import/uninstall previous version files *called "MAJOR UPDATES" -Fixed typo in command "Raygun" -- corrected from "Ragun" (first fix via new live updates system) -Created 'Add New Callsigns' system using Get-Choice boxes -Set most profile variables to config files in VA Apps Folder (edit with "Open Advanced Core Options" command) -Set most commands to Text Variables, read from config files
-
-
Major Profile Update – should be much faster, better, stronger than before…
basically I culled the AVCS CORE App for VoiceAttack, too slow, plugin would be too much, went with simple inline functions instead and everything is pretty much the same except for faster initialization.AVCS4 Falcon BMS Radios v1.3.91(beta):
https://veterans-gaming.com/index.php?/files/file/35-avcs4-voice-control-radios-for-falcon-bms/AVCS CORE Profile Framework for VoiceAttack v0.91(beta):
https://veterans-gaming.com/index.php?/files/file/37-avcs-core-profile-framework-for-voiceattack/ *(required) -
NOTE: This beta test is for functionality and flow of interactions with my various systems. It is a working beta for the updates, multiple profiles initialization, and first time user initialization which grabs the latest config files for said profile(s) from VG’s secure website AVCS pages (simple HTM pages hosting flat text files of the various default settings, TTS phrases, and Voice Command phrases), and places them in the VA Apps Folder under AVCS CORE and the appropriate profile name folder. This is how my profile can get “live updates”, or minor patch fixes, as I have put so much of the operating settings into these config files. If someone reports an incorrect TTS phrase, I can change it on the website and then “push” the update as a ‘Get User Choice’ box when you all load AVCS CORE next (“view changelog”, “update”, “postpone update”).
The initialization time is currently rather slow, depending on your CPU! If you already enjoy using AVCS4 BMS, DO NOT delete the old version so you can continue to use a working, fast AVCS4 profile when you get sick of helping me test this new, slow loading version! It will not interfere with the beta, and operates out of a separate VA Apps folder (“AVCS4” as opposed to the new “AVCS” folder for the betas)
I’m clocking 45 seconds to a minute for CORE to load itself and AVCS4 BMS 1.3.91(beta) on my old laptop i5 4690k - slower in less powerful CPU’s! Understand that this will be resolved before the end of the Beta tests, and I expect to get it on par with the old stand-alone AVCS4 BMS initialization time, closer to 20 seconds or less. If I didn’t release it now to test functions and flow of interactions, I’d miss out on valuable feedback or bugs I may have missed while I continue work on optimization and speed of initialization.
Thank you for bearing with these long load times during the beta tests! Feedback is encouraged, don’t hesitate to tell me if something is messed up or just plain stupid. -Sem
-
Hi,port my english.
Cant use 1.3.91 prof. I imported core prof and avcs 1.39 prof.Im trying to run it,but it resets on core prof.9:43:14.068 Inline function execution exception: Exception has been thrown by the target of an invocation. Access to the path 'C:\Windows\system32\91D60AC8C0E3D' is denied. 9:43:17.128 AVCS ERROR! AVCS CORE Online Database is offline, and cannot install this profile - or this AVCS Profile Version has been replaced by an update. 9:43:17.128 AVCS ERROR! AVCS CORE Online Database is offline, and cannot install this profile - or this AVCS Profile Version has been replaced by an update. ```Is this temporary problem with your servers, or this problem with my PC or firewall? The 1.1 prof is fully worked.
-
Hi,port my english.
Cant use 1.3.91 prof. I imported core prof and avcs 1.39 prof.Im trying to run it,but it resets on core prof.9:43:14.068 Inline function execution exception: Exception has been thrown by the target of an invocation. Access to the path 'C:\Windows\system32\91D60AC8C0E3D' is denied. 9:43:17.128 AVCS ERROR! AVCS CORE Online Database is offline, and cannot install this profile - or this AVCS Profile Version has been replaced by an update. 9:43:17.128 AVCS ERROR! AVCS CORE Online Database is offline, and cannot install this profile - or this AVCS Profile Version has been replaced by an update. ```Is this temporary problem with your servers, or this problem with my PC or firewall? The 1.1 prof is fully worked.
More of a VoiceAttack issue – VoiceAttack must be Run As Admin for a majority of it’s Windows interactive functions that I use in AVCS, in this case, file access. It can’t read a redundant profile version ID to make sure everything is in place, so other safeguards I put in place throw that error so it cannot proceed in a crippled state (potentially causing further issues). The error message itself is not specific enough, and I’ll refine these sorts of things before the end of this Beta test period.
Open VoiceAttack Options and click the last tab on the right, and check the box circled:
RSVP if that doesn’t work, happy to help new users!
-
Thank you very much.Everythink worked fine.
-
I apologize for how basic this question is, but I’m confused by the need to set up KEYPRESS_UHF_XMT_RELEASE and KEYPRESS_VHF_XMT_RELEASE.
I use a TM Cougar throttle, and I have set the KEYPRESS_UHF_XMT_PRESS and KEYPRESS_VHF_XMT_PRESS functions to the UHF and VHF “actuation directions” (for lack of a better term) of the Comms switch on my Cougar throttle. That part I understand.
For the …_RELEASE functions in AVCS4, should I set those to the exact same Comms switch “actuation directions”? In other words, for UHF, the _PRESS button and the _RELEASE button should be identical? And likewise for VHF?
If so, then I guess I’m a bit confused as to why. I would have assumed that by setting the _PRESS buttons, AVCS4 would be listening to my spoken command during the whole time I’m holding the Comms switch in one of the comms “directions”. (Sorry for this use of “directions”, but it’s the best term I can think of, for a 4-way hat-type switch).
Or is it intended that those switch actuations are supposed to toggle the sim’s comms functions on and off? If so, that sounds different than how I would expect the real HOTAS Comms switch to work IRL.
Sorry for the confusion!
-
I apologize for how basic this question is, but I’m confused by the need to set up KEYPRESS_UHF_XMT_RELEASE and KEYPRESS_VHF_XMT_RELEASE.
I use a TM Cougar throttle, and I have set the KEYPRESS_UHF_XMT_PRESS and KEYPRESS_VHF_XMT_PRESS functions to the UHF and VHF “actuation directions” (for lack of a better term) of the Comms switch on my Cougar throttle. That part I understand.
For the …_RELEASE functions in AVCS4, should I set those to the exact same Comms switch “actuation directions”? In other words, for UHF, the _PRESS button and the _RELEASE button should be identical? And likewise for VHF?
If so, then I guess I’m a bit confused as to why. I would have assumed that by setting the _PRESS buttons, AVCS4 would be listening to my spoken command during the whole time I’m holding the Comms switch in one of the comms “directions”. (Sorry for this use of “directions”, but it’s the best term I can think of, for a 4-way hat-type switch).
Or is it intended that those switch actuations are supposed to toggle the sim’s comms functions on and off? If so, that sounds different than how I would expect the real HOTAS Comms switch to work IRL.
Sorry for the confusion!
Yes, both _PRESS and _RELEASE would need to be the same joystick button respectively, in your case a directional on a 4-way HAT. The main feature of my PTT mode is that it does NOT effect Global Listening, but only my AVCS4 BMS Radio commands. This is so that non-BMS commands can continue to function, perhaps a “Mute Teamspeak” command, maybe a “Pause Track IR” or “Reset” command, or anything else that a person may add for themselves.
PTT functions by setting a boolean to “TRUE” if the PTT button is held down. When it detects a release of a PTT button, it changes that boolean back to “FALSE” after a brief delay. Each command checks first if PTT mode is ON, and then if the boolean is “TRUE”, and if both of these conditions are met, the recognized command will execute – otherwise the recognized command will exit before it presses any keys. This is why both a _PRESS and _RELEASE must be the same button, and must follow instructions inside each command (greyed out text, comments). EDIT: for clarity, the boolean “TRUE/FALSE” is not specific to either UHF or VHF, but is the same regardless of which PTT key was pressed/released - this allows users to ONLY set a UHF -or- VHF if desired, or both.
TMI; details:
When I use PTT, and I am flying with a friends, and maybe our flight also has 2 AI computer controlled pilots, I don’t want AVCS4 BMS recognizing any Radio Commands unless I have the UHF or VHF buttons depressed, but I do want my TrackIR “Pause” and “Reset” commands to fire anytime I say the phrase, whether I have the AVCS4 PTT button(s) pressed or not. This has the added bonus of letting users create their own Global Listening ON/OFF commands separate from AVCS4 functions. I realize not everyone is a “power user” of VoiceAttack, and some features may be unused - but for those who need the extra functionality, everything is in place to be as compatible as possible with a “local” listening ON/OFF mode specific only to the AVCS4 profile.In these new .91 Beta Tests, this can be helpful when using the QCC “Quick Command Creator” which presents options before saving the newly created command, including enabling PTT mode for that command, or not. Deciding which user created commands can use PTT mode, and require the PTT key(s) to be pressed in order for them to execute after recognition is another reason the PTT mode is not a Global Listening toggle, but rather command specific. Personally, I already used the QCC system to create a few ‘alternate phrasings’ for existing commands, where the existing Radio Command seems wooden or unnatural, and where the new QCC “Advanced Voice Command(s)” is set to execute the OTHER existing command with the wooden/unnatural phrasing. The final options pop-up when creating these presents the option to require PTT buttons to be pressed (if PTT mode is ON), or even play Audio Feedback (if AFM is ON) when recognized, or even a forced confirmation before executing, (“Are you sure?”).
This PTT concept has been slightly confusing for more than one person, so please don’t feel bad. The failure is on me for not finding a way to make this concept more understandable. I’m apparently a better amateur programmer than a teacher, and it’s something I need to work on. I have plans to create PDF guides with graphics, better explanations, and step-by-step instructions for various things, but I’m only one man and my time is divided between personal projects like this and managing the teams and servers that keep the VG Gaming Community up and running day after day.
Hope that helps! If not, or if you have any questions, I’m happy to assist!
-
Thank you very much, SemlerPDX, I really appreciate it!
Now I see how the PTT system works, and I can see why it’s necessary to have a particular PTT button (UHF, say) specified for both PRESS and RELEASE. That makes sense. Thank you very much for explaining it clearly like that.
And thanks heaps for working on this, I can see why it’s not at all easy to carve out time to write guides, manuals, etc… Just creating AVCS4 must be a pretty herculean task! Like most things in BMS, I dream of becoming a “power user” someday, but I should go through things step by step.
I haven’t really been able to get AVCS4 to work yet, but I only just downloaded the components yesterday, and have just begun working with it. Among other things, I see a lot of “unrecognized catch-all” messages after I try to issue commands like “Approach, request overhead approach” or “Approach, request unrestricted approach”. I did the Microsoft speech-recognition training again, but that didn’t seem to help much. Maybe I need to do it a couple of more times.
Oh, one thing I just thought of, wondering if it matters:
I use a Cougar throttle with one of those TUSBA R2 serial-to-USB adapters. Looking at the calibration program for the TUSBA R2, I can see that when my throttle is below AB, it indicates a continuous “DX button 15 is being pushed” signal. And when I go into AB, that changes to a continuous “button 16” signal. (It seems like it should give a continuous “button 14” signal when I’m in idle, but it doesn’t for some reason. Might be due to my having the Idle Cutoff function enabled, which I use a mouse button for.)
I have wondered if the fact that the TUSBA R2 is issuing these “continuously pushed button 15” or “continuously pushed button 16” signals would be a problem for VA and/or AVCS4. Last night when I was trying to map the UHF and VHF actuations to the XMT functions in AVCS4, I noticed that my UHF (= button 3) was sometimes seen as “button 3 + button 15”. That seemed strange until I realized the “button 15” signal was coming from the TUSBA R2. At other times, the “button 15” signal didn’t seem to be picked up by VA, and I was able to get a “clean button assignment” of just Button 3 or Button 2, as needed.
Does this sound like something that could cause issues? I wonder if there’s a way to have VA ignore certain buttons as irrelevant. Perhaps a way to tell VA “Yeah, just don’t ever pay any attention to buttons 14, 15, or 16. Those dudes are nuts. Just let 'em rant over there in the corner and pay them no mind”.
At all events, thanks heaps for your hard work on this, and for taking the time to explain it! :drink:
-
I haven’t really been able to get AVCS4 to work yet, but I only just downloaded the components yesterday, and have just begun working with it. Among other things, I see a lot of “unrecognized catch-all” messages after I try to issue commands like “Approach, request overhead approach” or “Approach, request unrestricted approach”. I did the Microsoft speech-recognition training again, but that didn’t seem to help much. Maybe I need to do it a couple of more times.
It is highly recommended to run through the Windows Speech Recognition Training session completely, three times (at least). Best tip would be to speak naturally, and to not over-enunciate else the “way” you get recognized on average would NEED to be over-enunciated, as well. Make sure the room is quiet, and when you are done, you can save your WSR Profile to a backup location (maybe USB flash) so that unless you get a new microphone or your voice changes to a degree that recognition begins to fail, you can always fall back to a former WSR profile following Windows reinstallation, or even new version of Windows, etc. so that you never have to run through the WSR Training Sessions ever again.
Please also press F1 when VoiceAttack is the active window to open the VA Help Manual, and check out the sections “VoiceAttack Quick Start Guide (v1.8.6+)” and “Troubleshooting Guide”
Here’s a backup tool for WSR Profiles:
https://www.dropbox.com/s/9b6ffxxkm00bu2o/Windows_Speech_Profile_Backup_Tool.rar?dl=1I use a Cougar throttle with one of those TUSBA R2 serial-to-USB adapters. Looking at the calibration program for the TUSBA R2, I can see that when my throttle is below AB, it indicates a continuous “DX button 15 is being pushed” signal. And when I go into AB, that changes to a continuous “button 16” signal. (It seems like it should give a continuous “button 14” signal when I’m in idle, but it doesn’t for some reason. Might be due to my having the Idle Cutoff function enabled, which I use a mouse button for.)
I have wondered if the fact that the TUSBA R2 is issuing these “continuously pushed button 15” or “continuously pushed button 16” signals would be a problem for VA and/or AVCS4. Last night when I was trying to map the UHF and VHF actuations to the XMT functions in AVCS4, I noticed that my UHF (= button 3) was sometimes seen as “button 3 + button 15”. That seemed strange until I realized the “button 15” signal was coming from the TUSBA R2. At other times, the “button 15” signal didn’t seem to be picked up by VA, and I was able to get a “clean button assignment” of just Button 3 or Button 2, as needed.
Does this sound like something that could cause issues? I wonder if there’s a way to have VA ignore certain buttons as irrelevant. Perhaps a way to tell VA “Yeah, just don’t ever pay any attention to buttons 14, 15, or 16. Those dudes are nuts. Just let 'em rant over there in the corner and pay them no mind”.
AFAIK, VA already does this: if you don’t program a joystick button to “do” anything, the fact that it is pressed is completely ignored by VA. But I am slightly confused here why one button would intermittently be recognized as a completely different button - I don’t use TSUBA or Cougar, though I wish I could. Not in the cards right now.
Please also remember that PTT mode is totally optional, though I realize it seems rather desirable for most users of AVCS4 BMS, which is why it exists in the first place.
Once you have WSR trained up well, and recognition is no longer a barrier to using VA as easily as it is intended to use, you might find that PTT could be redundant - unless you find VA firing commands due to normal conversation. Because AVCS4 BMS uses a large number of “wildcard” recognition commands, where any phrase that “ends with” those commands fires, it is potentially easy to call a command when not intended, though rarely. That is where PTT may be the best solution, or for other examples from my previous reply. But in the end, remember it is an optional mode with very specific uses, and if those cases don’t apply to you, you can feel free to use AVCS4 BMS without PTT Mode ON with great success and equivalent levels of immersion and functionality.
-
Oh, one thing I just thought of, wondering if it matters:
I use a Cougar throttle with one of those TUSBA R2 serial-to-USB adapters. Looking at the calibration program for the TUSBA R2, I can see that when my throttle is below AB, it indicates a continuous “DX button 15 is being pushed” signal. And when I go into AB, that changes to a continuous “button 16” signal. (It seems like it should give a continuous “button 14” signal when I’m in idle, but it doesn’t for some reason. Might be due to my having the Idle Cutoff function enabled, which I use a mouse button for.)
I have wondered if the fact that the TUSBA R2 is issuing these “continuously pushed button 15” or “continuously pushed button 16” signals would be a problem for VA and/or AVCS4. Last night when I was trying to map the UHF and VHF actuations to the XMT functions in AVCS4, I noticed that my UHF (= button 3) was sometimes seen as “button 3 + button 15”. That seemed strange until I realized the “button 15” signal was coming from the TUSBA R2. At other times, the “button 15” signal didn’t seem to be picked up by VA, and I was able to get a “clean button assignment” of just Button 3 or Button 2, as needed.
I finally figured this one out – you have a two-stage trigger I assume, and it functions by “always” holding one button down and by pulling the trigger, you release this button, and then can keep pulling to press the second trigger button. For this reason, many programs might interpret that first button as a button being held down.
A Beta version of VoiceAttack addressed an issue with joysticks not too long ago, and if anyone is unable to get beyond this issue to properly assign the button(s) they want to any VoiceAttack command (like my PTT function commands), you might download and install the latest Beta Version of VoiceAttack to avoid problems with setting up a joystick. It may not resolve the constant pressing of one of the trigger buttons, but pulling in on the trigger slightly to release the first trigger button, and then pressing the actual desired button should allow setting any button to a command:
Gary at VoiceAttack wrote …
VoiceAttack beta v1.8.6.9 is now available (Standalone and Steam)Hi, all - This version of VoiceAttack addresses an issue with the joystick screens appearing to be locked up if more than two buttons are down at once on your sticks prior to the screen loading (this seems to be the case with some of the newer sticks designed and used with flight sims).
Gary at VoiceAttack wrote …
VoiceAttack beta v1.8.6.16 is now available (Standalone and Steam)Heya, folks. This version of VoiceAttack contains another set of bug fixes - still working toward a full release.
Hope you’re all having a good nightThe beta can be downloaded here: http://www.voiceattack.com/beta
The zipped version of the beta is available here: http://www.voiceattack.com/betazipped
The Steam beta version can be downloaded here: http://www.voiceattack.com/getsteambeta
The rollback for this beta to the previous beta is here: http://www.voiceattack.com/filesend.aspx?id=VoiceAttackInstaller__18614.exe
The rollback for this beta to the previous Steam beta is here: http://www.voiceattack.com/filesend.aspx?id=VoiceAttackSteam_18614.zipAs always, the release notes can be found here: http://www.voiceattack.com/changesbeta
(please do report to the VA Discord or VA website forums if you have issues with the Voiceattack Beta and/or this issue)
*EDIT: The VoiceAttack Update v1.8.7 on October 15th 2020 has now included all of the fixes/additions of the previous Beta version listed above, and is now the most recent/updated version of VA as of the time of this edit.
-
Hello, thank you for the great work you did so far! I have difficulty finding the right phrase to say/recognized for ATC Commands/Approach.
I tried sayingApproach Request Vectors for Visual Approach
Request Vectors for Visual Approach
Request landing
Visual Approach
yet none of them resulted in the game issuing a voice command. I was on the correct frequency and was able to issue the command with keyboard.
However, the game did pick up when I said:Say again
and issued the in game command correctly.
What is the correct command one needs to speak for
ATC Commands/Approach Page/Request Vectors for Visual Approach
and```
ATC Commands/Approach Page/Request Vectors for Instrument ApproachI am using AVCS4 Falcon BMS Radios v1.3.91b and AVS Core 0.91b with latest BMS updates.