Lukrop's Voice Attack Profile
-
I’ve just solved my issue; you can’t just say “request picture” as that alone is just a suffix, ie it would output ‘1’. I have to say the prefix too, so “sentry 1” which would output ‘q’. Prefix then suffix gives ‘q1’ and thus, ‘request picture’ is recognised. I now have the same issue as Wheelchock in that I can’t say “Texaco1, Falcon1, request fuel” as it won’t recognise it. I guess that this is because I’d have to define “Texaco [1…9] Falcon [1…9]” “Texaco [1…9] Viper [1…9]” and so on for EVERY call sign for it to work. Essentially, I think I’m stuck with “Texaco1, request fuel” unless I sit through every permutation of callsign to/from.
From VA, click the spanner, then Recognition tab, then Utilities > Speech Engine Training. Follow the steps to save a word.
Lukrop, just a heads up. I’m guessing your keyboard is not QWERTY, and so I’d recommend you mention that in the new post. You have Tanker menu defined as Z, not Y, so whilst I can say “Texaco 2, request fuel”, it outputs ‘Z1’, not ‘Y1’.
Also, it seriously struggled to pick up “QFE” despite clearly verbalising it and saving a pronunciation to the dictionary; it would incorrectly detect it as ‘qnh’ 95% of the time. I fixed it by changing the word “QFE” in the profile to “Q F E” and it now works perfectly.
Planehazza, I have found a workaround for your callsign,callsign, request issue. I set the request to “request fuel” Then , in the 'pit I say "Texaco 1, Sting 1 ", then pause very briefly, then “request fuel”. I don’t do this with such as wingman comms such as “2,go Fluid”.
Once you get used to doing the pause, it comes quite naturally.
I have a QFE-type problem, only mine is RTB. I’m going to try your tip on that. -
Here’s corrected profile as the original one is not compatible with U2.
-there were some errors… eg , it was very hard for VAC to see difference between “Tower request landing” and “Tower request landing runway” …
- and there were few errors in ATC section (approach/tower)
-corrected
Cheers
Trying to import the profile I get a message claiming that it was created with a newer voice attack version, yet the program detects / shows no updates.
Any idea what could be causing this?
Cheers, Uwe
-
Trying to import the profile I get a message claiming that it was created with a newer voice attack version, yet the program detects / shows no updates.
Any idea what could be causing this?
Cheers, Uwe
This is because the profile was exported from the most recent beta branch version of VoiceAttack (currently 1.8.3.8 )
This is shown when the profile exported as expanded XML at line 44612 and edited with notepad (NPP FTW)<exportvaversion>1.8.3.8</exportvaversion>
AFAIK, you can edit this number @white-fang to reflect the current major release rather than beta, but apparently “no way to know if any keys would be exported that are exclusive to the beta and not present in the current version”, and then also you’d have to distribute the profile uncompressed of course. You can also roll back your VA version to do exporting (or if you have a laptop, toss the file over, and install the standard VA version to use it as an exporter - the install license is per user, not per-machine)
-
No prob , here’s xml expanded .vap .
See line 44837 <exportvaversion>1.8.3.8</exportvaversion> - edit to your liking , I guess? I haven’t use anything ‘exotic’ so it should work with older version(s) just fine… but … how to be certain??
I’m on current beta 1.8.3.8 , as it has some ‘crucial’ fixes vs original 1.8.3. - see changelog on VA site
Or maybe you could update to that beta… , “It seems” that is OK , didn’t notice any problems so far.
Cheers
-
Mate, do you have an Idea… (for VA programming… a bit)
How to make a statement … On button - VA listens , ignores unrecognized , catch first recognizable , stops listening … (simple )
But, there’s no such , I cant’ find, such option that it will “stop listening” AFTER it finds recognizable string (sentence)
Yes there’s “Stop listening” … but I need step above that in statement , “On recognition of (or confidence?) ANY sentence - do next” … ALSO “If UNRECOGNIZED then IGNORE - keep listening”
The latter I’ve found in manual , there is " on case if unrecognized" … but what “if recognized” ?? how to make that statement
…maybe I could do “NOT unrecognized” :D:D:D:D
Please, Thanks
–edit… so technically , I want VA to AUTO-STOP after it found something and executes it , meh could’ve said that easier way
-… hmmm, now I figured… , damn , when you writing about , ideas comes themselves… I can think of 2 dirty hacks there , …
1. add “stop listening” after EVERY command
2. execute “stop listening” as suffix , then try mass - pastebut both are … so so dirty haks, … is there “finer” approach maybe you can think of
-
I’m using a pinky shifted key combo on my WT throttle to enable / disable Voice attack listening for commands, it’s become 2nd nature by now and quite helpful because VAC will play a sound when it starts listening and another when it goes offline.
I think I’m using pinkyshift -> COMM SWITCH FORWARD for this.
With U2 now supporting multiple POV switches I guess you could also use unshifted POV hat on the throttle if you haven’t assigned anything to it yet.
All the best, Uwe
-
I’m using a pinky shifted key combo on my WT throttle to enable / disable Voice attack listening for commands, it’s become 2nd nature by now and quite helpful because VAC will play a sound when it starts listening and another when it goes offline.
I think I’m using pinkyshift -> COMM SWITCH FORWARD for this.
With U2 now supporting multiple POV switches I guess you could also use unshifted POV hat on the throttle if you haven’t assigned anything to it yet.
All the best, Uwe
HA, THANKS
But BOTH are already used in my profile … HAAHAH
/// SHIFT (S3) and COMM FW/BK/LT/RT are “DED - nav keys shortcuts” , so change radio and stp “on the fly” - ha!!! sneaky??!! yarrr !!!
// POV + S3 = UP/DN - Harrier thrust vectoring /// POV +S3 = LT/RT - FOV Inc/Dec… so I need VA to - *On COMMS VHF “Doubletap” - Starts listening - "garbled speech " - VA executes OR ignores and waits for first hit (repeat) - VA Stops listening … So all in one command
Thanks
# Radio Switch/T2-5 FW/BK/R/L SimICPNext 275 -1 -2 0 0x0 116 SimICPPrevious 276 -1 -2 0 0x0 116 SimICPDEDUP 277 -1 -2 0 0x0 116 SimICPDEDDOWN 278 -1 -2 0 0x0 116 AFElevatorTrimUp 0 -1 -3 0 0x0 115 AFIncExhaust 0 -1 -3 1 0x0 -1 AFAileronTrimRight 0 -1 -3 2 0x0 115 FOVDecrease 0 -1 -3 3 0x0 -1 AFElevatorTrimDown 0 -1 -3 4 0x0 115 AFDecExhaust 0 -1 -3 5 0x0 -1 AFAileronTrimLeft 0 -1 -3 6 0x0 115 FOVIncrease 0 -1 -3 7 0x0 -1
-
Methinks you might be best served by an additional button box, dear Sir!
I’m using my old G25 shifter which has a few extra buttons on it which align nicely with, say, Radar snooze / laser arm / master arm toggle, so in fact I still have unused, non-shifted switches left on my HOTAS WT throttle
All the best,
Uwe
-
Great!! … send those extra keys … I’ll find the use for them
Also have some old broken usb stick lying around attic , just need time and idea to build “little black box” around those axis and buttons, then , it can be anything. …think it has 3 axis , that could be radios and hud , but just 4 or 8 buttons …
(initially , the idea was building a simple external mixer for NI Traktor , now not so much…)Cheers
-
I too have an old beta innovations USB board gathering dust in the cellar which I had meant to use to convert my gameport-only Thomas TSW superwheel to USB many years back (that should give you an idea how old this thing is :)). I’ve considered building a pair of rudder pedals out of my old TSW pedals (I eventually threw away the TSW wheel because I had gotten so used to the Momo DFP’s force feedback which I had used during the conversion that there was simply no way of going back to driving GPL using a non-FF wheel), but for now my T500 pedals serve me just fine as a “poor man’s rudder pedal” using BMS’ virtual rudder function that this probably isn’t worth the time.
The 32 buttons, 8 analog and quite a few digital axes on the board leave open many possibilities, though
All the best, Uwe
-
You can get user input in a few ways – if you know what you want them to say, you can pre-define that list of words (separated by semi-colon) and put them in a ‘wait for spoken response’ action as below:
↑ Find these actions under the ‘Other >’ tab and under the ‘Advanced’ list -
- to find more info, press F1 and then CTRL+F for any of the key words shown like ‘{CMD_AFTER}’ or ‘Wildcards’ or ‘Compare’ or ‘Write to log’ (helpful for testing)…
Or if you need to capture something completely unknown, using Wildcards (asterisk) is the way to go – but there are limitations, too. For If/Then/Else statements, you compare one item to another via Text Compare (or INT/BOOL/DEC/etc) like below
↑ I tried this example saying “Record the words Pickle Rick” and the output was ‘The spoken phrase after the command was pickle wreck’ – again, there are some limitations, and evaluating for many similar sounding words is the way to go when trying to detect a keyword you need to use if it is known. if all is unknown, things get dicey… there are far better dictation softwares out there and VA dictation capture uses default WSR, and inherits its limitations.
-
I have been trying the profile with good results too, thanks for sharing it!
However one issue remains: I usually have Voice Attack listening for commands when I press (and keep pressed) pinky shift on the stick and COMM forward on the throttle. That doesn’t seem to be possible with this profile, I have to set VAC to continuous listening so it picks up the voice commands.
Other profiles work fine using my “push to to listen” function above, it’s just this one that’s not working. Any ideas what could be causing this?
Thanks,
Uwe
-
I have been trying the profile with good results too, thanks for sharing it!
However one issue remains: I usually have Voice Attack listening for commands when I press (and keep pressed) pinky shift on the stick and COMM forward on the throttle. That doesn’t seem to be possible with this profile, I have to set VAC to continuous listening so it picks up the voice commands.
Other profiles work fine using my “push to to listen” function above, it’s just this one that’s not working. Any ideas what could be causing this?
Thanks,
Uwe
Has anything changed? Update or started using the Beta VA? Without knowing how you’re doing your push-to-listen mode exactly, I would guess that there is something slightly different between your profiles where it works and the non-functional one. Can you switch to another profile right now and observe the same Joystick buttons combo working? Might double check that something isn’t out of place or even spelled wrong. In the profile where your Pinky+Comm does not work, you could add an action to that command “Write an event to the log” to force the Event Log in VoiceAttack’s main window to show a text notification for yourself when the buttons are pressed/held down. This might help you troubleshoot elsewhere, if it IS getting recognized, but is not doing anything (or not doing what it is supposed to).
Feel free to share a copy of your profile, happy to have a look at it - second set of eyes sometimes helps.
-
Thanks SemplerPDX,
I checked the beta versions and I’m already on the latest one it seems.
I think I’ve solved the problem by editing the profile and making certain that “global recognition buttons” weren’t overridden (they were before, I unchecked the relevant boxes and PTT, or PTR, “push to recognize” now works again).
All the best,
Uwe