Voice commands

[Krozgen]

I've messed around with some basic speech recognition applications before, and with knowledge of the windows API, it's always fun to simulate messages sent to applications, such as commanding winamp to hit the "next" button.

My question is... winamp has a set of messages which can be sent, using the windows API post/send message commands, that control its windows (such as play, next, volume up, etc.) Would it be possible to implement such messages in DefCon? It really wouldn't be too difficult, I think, and would allow for external programs to interface with Defcon more efficinently.

I guess what I'm trying to say here is... I want to hold up my mic, give the kill word, and have the whole world be blown to itty bitty pieces ... without going through the trouble of simulating a mouse click.

Please let me know

[elDiablo]

Its actually a lot more difficult then you think. Winamp just has to recognise one word (eg, "close" or "play"), while in DefCon you would have to split up sentences ("Fire nuke at Moscow", etc) to work out exactly what the play wants to do. And then you would have to work out where everything is on the map.

And how would you move a sub? "Submarine. Move to 50km North of X"? Thats probably tripling the complexity of DefCon just with giving units path finding, I would expect.

And obviously, each player would have to train up DefCon or the Windows speech recognition software first. Which takes HOURS. I did only the basic training in Windows to see if I could just "speak out" my essays, and it took ages, and still made loads of mistakes...

Thats not even getting to the fact that Iv have a history of releasing games for PC, Mac and Linux, which would mean getting some form of speech recognition software that runs in Linux and some that runs on Macs, and then making sure they all work! Which is again, tripling the amount of work.

Some RTS (I forget which) tried using speech recognition, and it just annoyed most people Which is a shame...

Still, nice idea, but its probably more work then its worth

[xander]

I think he was talking more about there being some kind of interface between Defcon and any other application, not necessarily speech recognition. However, the point of complexity still stands.

xander

[Babylon5]

That'll be Hey You, Pikachu! for the N64

[Krozgen]

Xander hit the nail on the head

It's not building the speech into the game that I'm talking about, moreso ... well, as Xander put it, "some kind of interface between Defcon and any other application". When one thinks about it, it's actually not too hard... at least in my opinion.

Oh, and btw... about recognizing the sentences, it's less hard than one would think. One way to control certain things might be sectors, or dynamic naming for external controls. Sectors might be... "Submarines" "Row 6" "Column 4" etc., though far more compact and faster. The dynamic naming would be an optional display name - so Moscow is city 13, Chicago 14, New York 15, etc. By having a command that displays these numbers, one might say... "Display", "thirteen" (indicating object 13), "Fire", "twenty" or somesuch.

Just an idea for more fun to a game that already seems excellent

[Luigi300]

The idea's not that bad. "Submarine twelve, move to thirty-two comma twenty-one."

[shinygerbil]

Unfortunately, as has been pointed out, the Windows API does not run on Linux or Mac. It may be possible to use DCOP commands on Linux - I believe they have a similar functionality. But then we open ourselves up to the possibility of extreme keyboard shortcuts/scripts. Potentially.

Steve

[Luigi300]

Macs have built-in voice recognition, Ambrosia'd probably use that when porting it.

[ewanm]

Voice recognition on linux would be more difficult due to the element of choice, although it is implemented to an extent by external programs.
And why not write an api to recieve commands from other programs, it dosen't need to give the position of subs, launch sites, ... etc. Just reicieve calls like "fire","move",..., and send a signal when it is ready to take the next call.

[xander]

Voice recognition on linux would be more difficult due to the element of choice, although it is implemented to an extent by external programs.
And why not write an api to recieve commands from other programs, it dosen't need to give the position of subs, launch sites, ... etc. Just reicieve calls like "fire","move",..., and send a signal when it is ready to take the next call.

I think that allowing other programs to interact with Defcon would be a bit unfair. Everyone should be playing the same game -- without some obnoxious kid hacking up Defcon to play itself, or exploit holes in the scripting system. If you want to play a game, play it. Don't have your computer do it for you.

xander

[shinygerbil]

you'd end up with:

Code: Select all

##MyScript.txt
##============
dcop defcon Nukes FireTeam1
dcop defcon Nukes FireTeam2
dcop defcon Nukes FireTeam3
dcop defcon Nukes FireTeam4
dcop defcon Defences BuildDefences

and repeat..

[DTNC Vicious]

THis would be awesome!

[xander]

xander

[DTNC Vicious]

yeah i revived it

[Duck'N'Cover]

Forgive me if I'm wrong, but didn't Clancy's endwar for the xbox 360 have voice activated commands? Maybe I read that somewhere in a review and misunderstood.