VoiceAttack - Dictation mode for Comms

Gary has been working on a new Beta of Voice Attack. One of the neat new features he has added is a dictation to speech ability.

I have been playing with it and came up with a 9 Command Profile that will allow you to Dictate messages to the comms in ED.

HOW IT WORKS:
First you need to use the Voice Attack Beta that is available here. It will replace the current version you have. All the features and setups that you have will keep working, but it will add the extra features.
This will not work with the regular VA.


The profile contains 3 Static commands for convenience. You do not have to use them.

If you do want to use them, You will have to edit three commands: Standard Greeting, Standard Reply, and Standard Target.
The first line in the command sets the text you want to be sent if you say this command. You can edit this as you want.

"Standard Greeting" is sent only in Local Chat. (ie: Hello CMDR.)
"Standard Reply" is sent only in reply to someone who messages you directly. (ie: Hello CMDR. How can I help you?)
"Standard Target" is sent only to your Current Target. (ie: Stand down. Do not run or you will be destroyed. Wait for instructions.)

I made these specifically to be static as a quick response. You do not need them for the rest of the dictation to work.

VOICE DICTATION COMMANDS:
These are the voice commands and how they work.

"Open a Channel"
This starts the dictation recording. It will take anything you sayand recognize it as best as the system can. It will keep recognizing till you give it one of the other commands. Pauses will be considered Sentence stops. It will automatically Capitalize the first letter of the sentence and put a period at the end.

Examples:
You Say: hello commander {you pause} how are you
System recognizes: Hello commander. How are you.
You say: hello {pause} whats up {pause} i dunno
System recognizes: Hello. What is up. I can all. (its not perfect :) )

Once you have spoken your message, you have to give it one of 4 commands or it will keep recording.

"Close Channel"
Cancels dictation and clears the dictation buffer.

"Send message"
Sends your message in Local Chat. ( uses /l )

"Send Reply"
Sends your message as a reply to the last person that directly messaged you. ( uses /r )

"Send to Wing"
Sends your message to your Wing. ( uses /w )

Remember: You can cancel your message at any time by pausing and then saying Close Channel.

CAVEATS:
It is not perfect. As you know, Voice recognition in Voice Attack is looking for a specific word or two, so it works quite well and will ever only be as good as you train the speech system. It will be less accurate if you use longer sentences. Use shorter multiple sentences for better accuracy.

NOTE: While it is waiting for your dictation, YOU CAN NOT USE ANY OTHER VOICE ATTACK COMMANDS. These will not be considered VA Commands, but will be added to the Dictation recording. It will keep recording until it runs out of memory (no clue how long that is), you decide there is a preoblem, logout and reboot your PC, swearing at me, and sending me a nasty message, or you Cancel or complete the command with one of the 4 other commands.

ie: You say OPEN CHANNEL and it starts recording. You Say HELLO COMMANDER {you pause} POWER TO SHEILDS {you pause} SEND MESSAGE
In Local Chat is enters "Hello commander. Power to sheilds.

It does not force focus on the Comms pannel so it should not interfere with what you are doing.

I am not repsponsible for what you send, or the results from using this. You are on your own.

HOW TO GET THIS INTO YOUR NORMAL VA PROFILE:

If you don't want to use the Static Commands:
Open the profile, and right click on each command and select Copy To and then select your normal profile. If you dont want the standard messages, don't copy them.

If you want to use the static commands:

Open your regular profile and edit your profile. In the bottom left is a button to Import Commands. Point it to the Dictation Mode Commands-Profile.vap file and it will add them to your existing profile.

All the commands are set to the category of "Comms Dictation" so they are easy to find if you sort by Category.

If you don't like the way they work, you can remove them from your profile. (Remember to remove all of them to avoid issues.)

Enjoy and if you have any issues or questions, feel free, but as I said, Use at your Own Risk - No Warranty expressed or implied!


 

Attachments

  • Dictation Mode Commands-Profile.zip
    6.8 KB · Views: 2,267
Last edited:
Thanks for putting this together. It sounds a lot like what I threw together for my profile when Gary put in the dictation commands. Nice to have a neat package for someone to incorporate into their own if they want to try this out.

The biggest Achilles heel in the whole thing is not VA's part, but in Windows word recognition, which can be quite hit or miss. The reason that VA usually works pretty well is that it doesn't have to figure out the exact match of a word spoken, it just has to get a close match to trigger the commands. Dictation on the other hand has to be accurate, and Windows isn't as good as others.

The best that I saw on the computer so far used Google through an API, much like Siri and Cortana do with phones, and it was much better, because it had a bigger and better database and processing to work with. But then Google killed that API. :(

Kind of hoping that when Win 10 comes out public, it will be better in this area.
 
Kind of hoping that when Win 10 comes out public, it will be better in this area.
I don't see why this wouldn't be true, cortana's recognition engine is 100% local and as long as it is open then voiceattack should be able to select it as speech engine.

Edit: local as in, the speech engine itself doesn't need online ability to work, however certain functions within it would need it, stuff not stored locally.
 
Last edited:
NOTE: While it is waiting for your dictation, YOU CAN NOT USE ANY OTHER VOICE ATTACK COMMANDS. These will not be considered VA Commands, but will be added to the Dictation recording. It will keep recording until it runs out of memory (no clue how long that is), you decide there is a preoblem, logout and reboot your PC, swearing at me, and sending me a nasty message, or you Cancel or complete the command with one of the 4 other commands.

I'm able to issue other VoiceAttack commands while dictation is running in v1.5.7.70... it's on by design. Should be working for you, too. Thanks for sharing this bit of awesomeness, btw :)

Gary
 
Last edited:
I'm able to issue other VoiceAttack commands while dictation is running in v1.5.7.70... it's on by design. Should be working for you, too. Thanks for sharing this bit of awesomeness, btw :)

Gary

Then maybe it has changed. It didn't work for me when I tried it before, but seems to be ok now.

No Problem. It is simple code even for a Luddite like me to come up with. Hopefully this will be implemented into even more complex profiles. I can see it being used for a Captains Log type situation, where you can dictate notes to a logging system that will save the information for you and then you can search it. WAY beyond my abilities ;)
 
I want to use this, but I wish the tech was a little more responsive. I've seen some say that working with your windows speech recognition program can help make it work better do you know anything about that? Specifically, I saw this guy, who clearly has a very heavy accent, but his VA works really good, in his comments he said that's how he got it to work better, by fooling the windows speech recognition that his accent was clear English.

[video=youtube;s4QkltClArw]https://www.youtube.com/watch?v=s4QkltClArw[/video]
 
Last edited:
Nice work. This looks great! I'm getting an error when trying to import it though..

Error Loading profile dada.
[There is an error in XML document (18,12.]

Any ideas?
 
Nice work. This looks great! I'm getting an error when trying to import it though..

Error Loading profile dada.
[There is an error in XML document (18,12.]

Any ideas?

Did you install the Voice Attack Beta that I linked to? It will not work with the regular Voice Attack.

- - - Updated - - -

I want to use this, but I wish the tech was a little more responsive. I've seen some say that working with your windows speech recognition program can help make it work better do you know anything about that? Specifically, I saw this guy, who clearly has a very heavy accent, but his VA works really good, in his comments he said that's how he got it to work better, by fooling the windows speech recognition that his accent was clear English.

You need to do 4-5 sessions with the Speech recognition Training tool in Windows for best results. The more you do it, the better it gets.
 
Did you install the Voice Attack Beta that I linked to? It will not work with the regular Voice Attack.

- - - Updated - - -



You need to do 4-5 sessions with the Speech recognition Training tool in Windows for best results. The more you do it, the better it gets.

Ha! I'm an idiot! Sorry :)
 
Forgive me for raising this necro thread from the dead, but can the commands be imported into the new version of VA or is it still limited to the VA Beta?

Cheers.
 
Great stuff, Loriath! Just tried it and it works like a charm. Added the line to the Extended VA pack and added a few other lines, great to be able to finally send greetings quickly.
Just a side note: Using your text macro also triggers the "Landing gear" command, since the "L" seems also to be interpreted as a Command, not just as a chat Send To Local command. Maybe I missed a setting there.
 
Last edited:
Great stuff, Loriath! Just tried it and it works like a charm. Added the line to the Extended VA pack and added a few other lines, great to be able to finally send greetings quickly.
Just a side note: Using your text macro also triggers the "Landing gear" command, since the "L" seems also to be interpreted as a Command, not just as a chat Send To Local command. Maybe I missed a setting there.

It should be pressing Enter to activate coms, then / and then L then space to activate local Chat. I have not checked it in a while but should still be the same. Check yours and let me know.
 
I independently tried to do this using the VA dictation commands but despite getting the comm's box open no words ever appear!

I then found this thread -same problem. VA recognises what I say but does not "print" them out.

I tested this in M$ Word and still no words appearing... any ideas?

VA Version 1.5.8 FULL
 
Last edited:
I independently tried to do this using the VA dictation commands but despite getting the comm's box open no words ever appear!

I then found this thread -same problem. VA recognises what I say but does not "print" them out.

I tested this in M$ Word and still no words appearing... any ideas?

VA Version 1.5.8 FULL

Are you completing the command with one of the send commands?
 
Are you completing the command with one of the send commands?

I finally figured it out and saw the send commands seperate - which is what confused me. I'm working on wrapping this all into one voice comm's command.

BTW this is awesome stuuf - VA works almost flawlessly in converting speech to text!


And when it doesn't the results always seem to be amusing ;)
 
Back
Top Bottom