What about voice commands

[Columbus]

Saw a commercial about Apple's Siri and...

Just wondering, how much would Spring benefit from voice commands?

Currently there are alot of keyboard shortcuts in use and maybe some of them could be doubled by voice commands.

It could even go as far as having some intelligent functions, like giving an order to reclaim all corpses on screen with nearby free constructors, or to call out the names of units you want selected in range of the screen and telling them to go to route a or b or tell certain factories to do something.

I imagine the easiest way to do this would be to have some kind of a lua widget/gadget that communicates with an external app for speech recognition plus translation and then lua performs the requested commands.

There are already lots of free/kind of free solutions/API-s available for speech recognition so the main challenge would be to make the part that deciphers the command that you have said.

Although most speech reco soft is platform specific(like MS-s Speech API), the lua and text deciphering would be universal.

Currently, there doesn't seem to be any known games out there that have this feature.

I guess it would look something like this:
http://www.youtube.com/watch?v=owJS5XwXAEA (not my vid ofc.)

If it's deemed a good idea, I might give a shot at it.

You know... just sayin(let the trollin' begin).


PS. This is not a feature request.

[momfreeek]

Wow, that would be so cool! "Flash team 1 attack now!"

Ok, it probably would be worse than useless but it sounds cool.

[smoth]

Do not want, I am on VoIP when gaming.

[luckywaldo7]

Cool idea, could be fun but not too useful because I don't see how you would be able to meet the precision of keyboard+mouse, even with assistance from a brilliant helper AI. Also like smoth said, VoIP.

If you are interested in development though I'm sure there are loads of projects people could recommend, between lobbies and engine and lua and whatnot.

[Johannes]

Pressing a key after another VS uttering sentences, which is faster and why?

[momfreeek]

q: which is faster? mouse or keyboard?
a: both together!

so.. wouldn't a mixture of all 3 be best? mouse + keyboard + voice?

marking positions and directions are best done with mouse. so what about things like:

"select one penetrator and one jammer"
"save unit 1"
"unit 1 move to [click]"
"vehicle lab build five stumpies"
"all vehicle labs move to [click]"

theoretically that could be done with keyboard commands, but it seems less intuitive to learn "a,v,l,m,[click]"... and there's only so many keys on the keyboard.

[SpliFF]

I think you're forgetting that Apple are bitches and Siri is now and forever locked to their official applications (ie, no API). Before getting excited about it you might want to think about how you would ever connect Siri to spring.

Anyway, this whole Siri thing makes me puke. Dragon Dictate has been around for about 20 years and it controls pretty much anything (ie, virtual keyboard). People are acting like Apple actually invented something here instead of simply binding a voice engine to a chat bot and calling it a fucking revolution like they always do.

Any, back on topic. I looked into this years ago and sadly there is no such thing as a good open-source voice-recognition library. There are some projects like Sphinx but they come across as being incomplete and only really useful for academic work. There use to be an IBM / princeton project called ViaVoice but they killed that off. That pretty much leaves Dragon Dictate as the only viable option but it's non-free.

[jK]

1. http://en.wikipedia.org/wiki/Tom_Clancy's_EndWar#Voice
2.

PS. This is not a feature request.

3. so no comment by me (you know it's better that way)

[MidKnight]

Voice command opens up some cool possibilities for multitasking.
For example: I could queue units and manage my eco with voice commands while my eyes/hands are busy on a raiding mission.

Voice would also let you move units around without flying around everywhere on the map. (Then again, unitgroups lets you do this, too)

[Jools]

What a cool subject, I would say: absolutely!

Voice commands work quite well (depends much on the mic you have), but about 99% can be correctly heard.

The thing is, it's a totally new dimension of commands. You can give commands even though a menu is open, saves lots of clicks. Would be really cool if you could integrate this with lua.

Windows speech recognition (WSR) already works quite well, if you are interested, look at some macros at http://tech.groups.yahoo.com/group/ms-speech/

I already use some macros to pause/unpause/control winamp while in-game, it saves you from the need of exiting full screen.

[Jools]

I think you're forgetting that Apple are bitches and Siri is now and forever locked to their official applications (ie, no API). Before getting excited about it you might want to think about how you would ever connect Siri to spring.

There is an API for WSR together with a macro language. So we just need an interface with lua.

http://blogs.msdn.com/b/tsfaware/archiv ... ition.aspx

[MidKnight]

WSR is platform dependent and varies even from one version of Windows to the other.

A library that could be bundled with Spring is likely a better choice.

[Floris]

Winamp already has build-in shortcuts. I use them a lot. ctrl+alt+...



Voice recognition could be usefull in avoiding menu's

'Hold fire', 'Roam', 'Nano', 'Nuke', 'Fighters', 'Bombers', 'Fight', 'Patrol', 'Cloak'

[Columbus]

Do not want, I am on VoIP when gaming.

You could always use a button to enable voice commands/disable VoIP?

I think you're forgetting that Apple are bitches and Siri is now and forever locked to their official applications (ie, no API). Before getting excited about it you might want to think about how you would ever connect Siri to spring.

Never intended to use Siri. If Apple would make it open, then the planet would start turning backwards

Any, back on topic. I looked into this years ago and sadly there is no such thing as a good open-source voice-recognition library. There are some projects like Sphinx but they come across as being incomplete and only really useful for academic work. There use to be an IBM / princeton project called ViaVoice but they killed that off. That pretty much leaves Dragon Dictate as the only viable option but it's non-free.

Well, many years have passed since then

Currently, the Windows SR does look like the best candidate, although that will leave linux and mac people on the dry.

Also, with Apple making voice recognition "the latest thing" I'll bet there will be many projects popping up to mimic this, maybe even one that we can use.

The nice thing about voice commands is, that when a better reco solution is found, all you have to do is change the API of the command deciphering app, the commands will always work the same.

WSR is platform dependent and varies even from one version of Windows to the other.

A library that could be bundled with Spring is likely a better choice.

The russian roulette equivalent in spring:
"Send Group one to Location Enemy Base" and then it'll Ctrl-D instead.

This will be like motivation for people to upgrade their OS: If they can run spring, then they can sure as hell run Win7
/OT

There is an API for WSR together with a macro language. So we just need an interface with lua.

http://blogs.msdn.com/b/tsfaware/archiv ... ition.aspx

That looks very cool, but it has a downside: then the system will be even more platform dependent, since these macros are for WSR and probably can't be used with anything else.
Still, looks like this way will be easier than I imagined, I'll set up a few tests with it.
Got to read up on Lua and gadgets

Now, Kinect anybody?

[AF]

The issue is not voice recognition, Google have APIs for that.

The issue is now that we have a plaintext natural language input what are we to do? AIs already have a lot of these issues and still haven't satisfied.

For example:

"Attack the enemy"

• With what?
• Where shall I attack the enemy
• From which direction
• Point blank or at a distance
• Which enemy should be hit first
• How do you get there? ( the direct route isn't always the optimum
• In a line? In a surrounding circle? Pincer movement?

"Withdraw units"

• Which units?
• How far should we withdraw?
• Where should they go to?
• Are they withdrawing from their current position or from a targetted position?

etc etc etc, and this is all after you've figured out the meaning of what is being said.


Any voice recognition is going to be limited in what it can do, most likely single or two word commands issued to what is currently selected such as "cloak" or "self destruct". Commands such as attack wouldn't issue the order, they would be the same as pressing the attack button or A, you'd still need to click somewhere.

[Anarchid]

The issue is now that we have a plaintext natural language input what are we to do? AIs already have a lot of these issues and still haven't satisfied.

I'd imagine you'd get to use some kind of a word-set for precision and brevity and all kinds of defaults. Like the military use anyway. Or like any CLI program has.

Then you could pass all the extra parameters just like you pass keys to a shell script. So, just voicing 'attack' would be ignored for lack of parameters, but 'group five long range one time attack base sector A5, commence now' would send a squadron of bombers to raid (fire bombs/rockets from max range and leave once depleted) any known (or unknown) enemy buildings somewhere along the center of the left border of the map. (leaving out which keys would be optional and which would be required for such an order)

After all, it's not like grep has an AI to parse its keys, that shouln't be so hard Just supplant 'natural language' for 'military command language'.

The down side is obviously the user would have to learn that 'language' much as they have to learn keyboard shortcuts, or even harder (keyboard shortcuts don't have much syntax besides modifier key combinations; military-speak like that would have to have quite rigid syntax).

[momfreeek]

selecting units by name seems like the most obvious simple advance:
"select 5 stumpy"
"select vehicle lab"
"build 10 flash"

these aren't things easily done with the keyboard and the mouse is now free to point out targets and directions (which aren't easily identified by voice).

[Jools]

I was thinking of leaving that open to the user. Just intergrate SR with lua and creativity will spring up with something. How about
"nap commander" instead of
ÔÇö Find and select nearest transport
ÔÇö Find enemy commander
ÔÇö to to load enemy commander

(all of the above requires user interaction, whereas the nap command could be run independently)

[Jools]

WSR is platform dependent and varies even from one version of Windows to the other.

Yes, it is for Windows 7. Isn't that was most use anyway. It works both for 64-bit and 32-bit ones.

You can also include it in windows XP, although I think it's not native there.

[Anarchid]

I was thinking of leaving that open to the user. Just intergrate SR with lua and creativity will spring up with something.

That sounds very nice. Now, if we could also hook up different engines to it, then maybe using WSR on windows and, say, Sphinx or google elsewhere would somewhat even out the odds.

On a side note, doesn't Chromium (open-source version of Chrome) has its own speech recognition? Or does that simply use google's remote stuff? I suspect it could be extracted in the former case.