keyclone + voice recognition a marriage made in heaven.

[shaeman]

This forum is amazing. I learn something new everyday and a quick post with a video yesterday has (possibly) made my ideal setup come true.

Being a masochist I wanted to multi box 5 different classes (ultimate aim), but rather having one character played well, and the vast majority of the other classes abilities unnused was not something I wanted.

However macroing up and assigning keys, and then trying to remember key presses for these infrequently used but vital abilities would be a nightmare.

I watched a video of someone two boxing with voice recognition yesterday and saw some flaws with it and wondered how it would scale for 5 characters. I initially discounted as a replacement for keyclone.

This morning it suddenly occured to me that keyclone in conjunction with voice recognition may be the answer.

I'm now imagining a setup like the following.

The most frequently used direct damage and healing spells are going to be hotkeyed. The class specific abilities that may not be used all the time but can play a vital part will have a voice command and a unique (but very odd ) key combination assigned.

So for example - I'm playing, i notice my warlock is climbing the aggro table I say soul shatter and it drops (or I say Aggro Dump). My hunter is now climbing the table so I can yell feigndeath (more likely playdead).

I see an add coming along but notice it's a beast - I say hibernate.

I'm thinking that most of the commands I would setup wouldn't be emergency type stuff but generally stuff that may save a lot of moving to another screen and trying to trigger an ability that isnt part of my usual key mashing.

I can see this helping for setting up distracts on my hunter to the tank or to the pet. I can fire off a misdirect if a mob is heading for the healer.

I can banish or enslave on my warlock, reseduce.... all whilst being able to focus on the main dps.

I just think this is going to make mixing multiple classes and using their strenghts and varied abilities easier.

Unfortunately I'm at work so can't get to try this out, but I think tonight I'll be buying something like vr commander and trying it out.

I'll keep people posted on progress.

[gbremset]

Oh.. That sounds awesome.. If you get it to work, give us a guide..

However... Wouldn't this be borderline Automation, hence breaking the TOU/EULA? Discuss please. :)

[Glaxa]

Nice idea.

I don't think it is classed as automation because you are still giving the command to cast a spell, just like pressing a button. WoW allows people to use voice commands instead of pressing a key when playing solo so don't see why this breaks any rules.

[shaeman]

I think this is within the terms of service, it doesn't give any competitive advantage as such, certainly not the way I would use it.

The triggering voice command in essence simulates a key stroke.(I wouldn't do anything other than simulate a keypress - no delays loops etc). The key press may trigger an in wow macro.

I did see a video of someone using software like this (but you've got me worried now ). Is there a list of prohibited software available from blizzard?

[Drizzit]

I read about this too... I believe there is a wiki on this and a couple of posts. The down side is depending on what software you use there might be a problem. Think about it the computer record the word, lets say "heal", you recorded it right when you computer boots up and you are relaxed. You bind that world to Healing wave, since you will only use it when you need it quick. Your in a battle and you get a bad pull, your heart is pumping faster and you get tunnel vision and a little nervous (omg i'm going to die), you yell "HEAL" noting happens and you keep yelling nothing happens and you die. The reason why it didn't work is not the program it is you "heal" <> "HEAL". Like i said it depends on the program, if it is a good program it wont matter, but a bad program you will not get the heal. There is one program in the wiki but most virus scans find a virus in it. What might be good is some of the buffs that you can do out of combat. But it is something worth looking into.

[shaeman]

I'd pretty much keep my voice commands to abilities that can be triggered prior to combat (hunter misdirections etc), and to reactive abilities that may not absolutely break the fight if they dont work.

Healing and main damage dealing spells are gonig to be hotkeyed - the video I saw had a few moments where heals weren't recognised and that to me would be unforgivable.

The software I found (at $10 bucks it seemed worth a try and I wont be upset if it doesn't work) was http://www.vrcommander.com/

Please don't purchase it yet - I haven't tried it so can't recommend it.

The reason I went for this rather than something like dragon dictate is the lack of training (as it states - dragon dictate is always looking for context from many many possibilities, this one is looking for a limited set). Also it seems to have a setting for continuous recognition, so could be slightly more responsive as a result.

I have ideas for this - so voicecommands such as hunterfocus, warlockfocus could set the focus for the hunter to the mains current target , then set the warlocks focus to the next target. I could then pretty much say misdirecttank or misdrectpet and say banish to have the warlock banish it's focus.

In theory combining it with keyclone would give me the best of all worlds .... In reality I may turn out to be disappointed.

Either way I will try this on my paladin, boomkin , shadow priest midlevel characters tonight. (they haven't been multiboxed together yet, and have no macro setup so seem like likely candidates to have).

[Drizzit]

I would suggest that you use 2 words for all commands. Like "Shield me", "Bubble Me", "Follow Me"... Reason is it will be harder when you talk to someone that you toons will do that (granted that shield and bubble will probably not be used that much in everyday life, but whole knows). One of you kids could comes in with bubbles ("bubble" is the word to use bubble) and you to tell them to "stop blowing bubbles", well you pally just bubbled and lost all agro and now you have a wipe... dam kids lol. Another person said to put please at the end of everything "Bubble please" because we all know no one says please anymore lol

[shaeman]

Good point.

I'll probably have commands of the format <action> <class>

e.g

bubble druid, shield hunter.

Distract pet

distract tank

banish focus

fear focus etc

[Drizzit]

Distract pet

distract tank

banish focus

fear focus

I think that these would be great ones. Many times when i see a mob about to break fear and have to find the fear button and then go bad to my mob... Great example is the lion from kara (Wizard of Oz)

[shaeman]

Ah yes - Roar.

I put in a focus fear macro on my lock specifically for that encounter and to take out one of the elementals in the shade of aran fight. It's come in handy in many an instance too.

[Fleecy]

Check out my post in another thread. I've been using Voice Recognition to control my characters for months now.

Voice control ('http://www.dual-boxing.com/forums/index.php?page=Thread&postID=85073#post85073')

When I first started using voice control I took it a little too far and automated a few things. After chatting to a GM about it I realised my mistake and removed the autonation component. But I've had a number of chats with GMs and controling the game by voice is a perfectly legal way of playing. As long as you are not exceeding the functionality of the in-game macro system. Eg.. no waits or timed presses, also not casting more than 1 spell per voice command. I *had* a voice command "Buff Group" which would send CTRL-1 (Priest Fort), wait 1 seconds and send CTRL-2 (Priest Spirit). That is autonation. To make it legal I changed it to a cast sequence spell. I would still give the voice command "Buff Group" but I would say it twice.

[shaeman]

Thanks fleecy. Good to know that if you keep within the boundries it's fine.

[Xuthus]

I've done this it is awesome ,but there are a few flaws .Such as if you get excited you speech changes .You need to basically talk animatronically ,as with keyclone there is a delay up to a full second. I would still be using it, but built a new comp and haven't got it all reset up.A few tweaks and you've got your dream come true. GL

[Bradster]

It's a neat idea but voice will never compare to the speed of pressing a key instead, in PVE it could work out nicely. The other down side is when you press a key it understands and perform that action. Voice on the other hand "heal", "Heeealll" "HEEEAALLLLL" you die. Then after you rez, you wife comes in pissed that you woke her up, she tells you to go die and in a fire, your locks then cast hellfire and kill themselves, worst yet your sleeping on the couch tonight.

[Binaryzero]

I set up a system to do this a while back (when Vista First came out) and it worked very well overall. I used to to to control my priest for my priest and hunter team. I used http://www.codeplex.com/KeySimVoiceCommander to do it. getting it running is a bit rough but it worked well when done. It would NOT be hard to use it work keyclone and control a full team.

[blast3r]

I'm probably going to play around with this idea this weekend as well. This would make it a lot easier to run multiple classes!

Xinzu Fear
Xinza Enslave
Xinze Earthbind
Step Back
Turn left
Turn Right

Omg I can't wait to play around with this stuff.

Oh, and:

Gank Horde :) just for fun.

[Depherios]

Itep Back
Turn left
Turn RightWatch out with these... not sure what blizz would call automation. I suppose you could use "Turn Left" "Stop Turning" -- but if they stop automatically that could perhaps be construed as automation?

[shaeman]

Just an update - I tried this out today. It absolutely rocks. It's beyond a marriage made in heaven.

I did a fraps of it - but it recorded my wow sound, but didn't get my voice commentary which made my video useless. I'll try and resolve that.

Basically this works exactly as I hoped it would.

I was running with 3 locks to try it out. I had them all keycloned up. Then I activated the voice recognition software.

Configuration was simple - create a template. Then within this template create a new command (by typing the name you would say) then set it so it does a keypress. This key corresponds to the hotkeys. You then save and load the template and voila voice input.

In this way if i said demonskin all three would pop demonskin. It was seamless. No voice recognition training. it took the command 95% of the time if not more.

I really cant wait to try this out on a group of different classes.

The software which i bought is $10. Is called VR commander available at vrcommander.com (I don't want to sound like an ad), but this software gave me the same warm and fuzzy feeling I got from keyclone.

Like I say - I'll see if i can fraps something, then figure out how I encode it and try and upload it.

[Catamer]

I would think that it would be ok as long as the voice commands mapped to single key commands.
I believe the G15 macro's are only ok as long as they are mapped to single key commands.
You can't make external macro's either by this voice software or by the G15 keyboard doing a long sequence of events or timed events.

sounds interesting, I want to see the video of it in action!

[Fleecy]

I've had many many chats with GMs over the functionality of voice controlling games and avoiding Blizzard's defination of automation. I have the following move type commands setup, and I have confirmed with a GM they are fine.

Giving the voice "Turn 3 O'Clock" and my characters will turn to the 3 O'Clock position relative to their current position. To do this the voice command triggers the Right arrow key presses and holds it down for about 1 seconds. Likewise for "Turn 6 O'Clock" but holds the key down for about 2.5 seconds.

What you can not do is send more than ONE key press per voice command. You have to look at it like this, If you can perform the same function by pressing a single mapped key (or modified key combination - ctrl alt shift) on your keyboard, such as "A" then you can do it via a voice command.

Xuthus is right, your voice does change when you get excited and it can sometimes inpact the recognition performance. But I've been using Dragon for quite a while and its been "trained" very well over that time. Regardless of the pitch in my voice or the scream as I yell for a heal its usually accurate.

But there are times when it doesn't do what its told and is a great cause of frustration! Because you think if you had just pressed the button instead of using the voice command then you would have been able to kill that boss and not wiped. But thats a decision you have to make for yourself. I've become so reliant on voice control that I don't remember half of my key bindings for my abilities. Basically without voice control I would have to learn to button bash again.

[shaeman]

As a further commendation for the vr commander software, there is no training for your voice required.

I put the commands in - and tried it. It worked first time. i then handed the headphones and mic to my son (who's voice hasn't broken yet) and got him to speak the commands and it recognised them when he spoke them. It really is phenomenal. I imagine if it can handle the pitch shift between an adult male and a child then it should do fine when I'm squeeling for a heal . (although I'm not going to use it in that circumstance).

I can't wait to give it a go on multiple different classes. Tomorrow I will start testing it with my druid and warrior, but at least now I know my dream of controlling 5 different classes reasonably well (and utilising more of their abilities) is possible.

[Groovy]

What ever happened to dragon naturally speaking?

[shaeman]

Dragon naturally speaking is still available at nuance.com (or something like that).

I had an earlier version (my first foray into speech recognition) and it worked pretty well. The more time you spent training it the better it got.

It's more pricey than the voice recognition I'm using specifically for games, but dragon has a whole load more feaures.

Supposedly windows vista has got pretty good speech recognition built in - and I think someone on this forum was talking about interfacing with that for us multiboxers.

[Korruptor]

All this Voice stuff made me think of the movie Dune:

Paul Atreides(Muad'Dib): "Some thoughts have a certain sound, that being the equivalent to a form. Through sound and motion, you will be able to paralyze nerves, shatter bones, set fires, suffocate an enemy or burst his organs. We will kill until no Harkonnen breathes Arakeen air."

[shaeman]

And mua'dib was the killing word . Words have power

:)

[Anozireth]

I missed this post yesterday or I would have jumped in earlier.

I'm working on what will be a free program to do this designed specifically for WoW. There's more info in This Thread ('http://www.dual-boxing.com/forums/index.php?page=Thread&threadID=9427'), but basically it will allow you to map keys to voice commands or to use your voice for in-game chat. I'd eventually like to add some multiboxing specific features, but likely not in the first release. Test version will require Vista (works on 32 or 64 bit), XP maybe later. I'm nearly to the point that I'm ready to give it to a few people to test, maybe by the end of the weekend if I have the time.

[The IT Monkey]

Giving the voice "Turn 3 O'Clock" and my characters will turn to the 3 O'Clock position relative to their current position. To do this the voice command triggers the Right arrow key presses and holds it down for about 1 seconds. Likewise for "Turn 6 O'Clock" but holds the key down for about 2.5 seconds.
I know you have stated that you have talked to many GM's about it... but from my understanding and everyone I have ever talked to TIME DELAYED or TIMED KEY PRESSES are a no no. I would love to be wrong about this. LOVE LOVE LOVE to be wrong about this. Last time we had this conversation with someone using voice command I believe it was the same exact issue. After further clarification with a Blue on the forums (they are the be all end all of policy in my opinion as you can POINT to it and say "but they said" where you can't do that with a GM) in where the poster pointed out that he was using a TIMED function the Blue said that was against the rules.

Not trying to rain on your parade... would like to use such a device myself.... but would HATE to see you get in trouble for something you thought was okay. Would also hate to see people follow your example without it specificly be okayed by Blizzard.

[Anozireth]

I'm curious as to some of the delay/automation issues with my program as well. Right now I've got it so I can tell it I want to say something in /g for example, and then it brings up a box to dictate into, then I say "send" and it sends it all at once. It's technically sending dozens of keypresses at once, but it's not doing anything I didn't tell it to do. I have a hard time believing this would be an issue for chat stuff.

I've also got it so I can say something like "Mark" and it sends "/cast Mark of the Wild". That's clearly sending more than one key press for one voice command, but it is only performing one action and there's no delay aside from the time it takes the voice recognition engine to process the command. This same action could be achieved by sending a single keypress that's mapped to an in-game macro that does the same thing. Currently program would probably allow for macros longer than 255 characters (haven't tested that), but I think there's already in-game addons that do that anyway (right?).

Lastly, I added the ability to put modifiers on single key presses, so I could map a voice command like "Barkskin" to Ctrl+1 for example. Again, it's technically sending multiple key presses(one each for Ctrl and 1), but only one action is happening per voice command.

I'm curious to see what the ruling would be on things like this. The key thing is that I'm not allowing any delays; each command gets sent all at once. There is only one game action taking place per voice command.

[The IT Monkey]

I'm curious as to some of the delay/automation issues with my program as well. Right now I've got it so I can tell it I want to say something in /g for example, and then it brings up a box to dictate into, then I say "send" and it sends it all at once. It's technically sending dozens of keypresses at once, but it's not doing anything I didn't tell it to do. I have a hard time believing this would be an issue for chat stuff.

I've also got it so I can say something like "Mark" and it sends "/cast Mark of the Wild". That's clearly sending more than one key press for one voice command, but it is only performing one action and there's no delay aside from the time it takes the voice recognition engine to process the command. This same action could be achieved by sending a single keypress that's mapped to an in-game macro that does the same thing. Currently program would probably allow for macros longer than 255 characters (haven't tested that), but I think there's already in-game addons that do that anyway (right?).

Lastly, I added the ability to put modifiers on single key presses, so I could map a voice command like "Barkskin" to Ctrl+1 for example. Again, it's technically sending multiple key presses(one each for Ctrl and 1), but only one action is happening per voice command.

I'm curious to see what the ruling would be on things like this. The key thing is that I'm not allowing any delays; each command gets sent all at once. There is only one game action taking place per voice command.All of the above (with the exception of the chat stuff) is supported by in game mechanics and is clearly allowed by Blizzard. I would equate "per voice command" with "per key press". I think you are clearly covered hear. If you were to program it with something like "back up" and it sent a key press to your crew that backed them up for 10 seconds by sending that one key press for 10 seconds and then releasing... that would be automation (at least as far as I understand Blizzard's stance on the issue) and therefor be against the rules.

As far as the chat things go I wouldn't think they have a problem with it.

I would, if I were you and had your programming skills, ensure that your program was not customizable to the point of being to add delays, But that is just my 2 c

[Anozireth]

I'm almost ready to release my voice recognition program for a small alpha test. See this thread ('http://www.dual-boxing.com/forums/index.php?page=Thread&postID=88230') for more details.

[Anozireth]

A bump to note the alpha release of my voice recognition program. See Update! WoW VoiceBox Alpha release! ('http://www.dual-boxing.com/forums/index.php?page=Thread&postID=89205#post89205') for details.

[Simulacra]

I tried the vrcommander trial - not bad, a little unresponsive here and there but not bad at all. I really like saying ATTACK and they do lol. One warning, when flying don't cough as it tends to unmount and it's a loooong way down.

[Toned]

Mute your mic ?

[Simulacra]

Mute your mic ?Yep I learned the hard way, I have to have the mike right in front of me and it's got a big OFF button thank god. The prob with vrcommander is that it just stops responding after awhile.....no idea why, so what I have to do is stop it and start again, anyone else having the same prob?

On another note, after years of playing wow my wife finally took an interest last night as I was controlling the chars with voice lol. I got her saying MOUNT, STANDUP, SITDOWN, JUMP lol - she had a bit of fun there, it brought the control freak out in her.

[Fleecy]

All of the above (with the exception of the chat stuff) is supported by in game mechanics and is clearly allowed by Blizzard. I would equate "per voice command" with "per key press". I think you are clearly covered hear. If you were to program it with something like "back up" and it sent a key press to your crew that backed them up for 10 seconds by sending that one key press for 10 seconds and then releasing... that would be automation (at least as far as I understand Blizzard's stance on the issue) and therefor be against the rules.

As far as the chat things go I wouldn't think they have a problem with it.

I would, if I were you and had your programming skills, ensure that your program was not customizable to the point of being to add delays, But that is just my 2 cI've had many chats with GMs over the definition of automation and the use of voice recognition to control the game. Basically 1 command = 1 key press - voice or physically pressing a key. As for the voice command "Back Up"; it presses and holds the down arrow key for X seconds. While it is a "timed" or "delayed" ability and it does not voilate the "automation" ruleing. I've had that chat with a number of GM's.

[Malekyth]

Xuthus is right, your voice does change when you get excited and it can sometimes inpact the recognition performance. But I've been using Dragon for quite a while and its been "trained" very well over that time. Regardless of the pitch in my voice or the scream as I yell for a heal its usually accurate.

I find, to my own amusement, that the commands I say to Voice Commander tend to be repeated so often that they settle into their own distinct timbers, or tones, or whatever I'm trying to say. I mean, I say "dot" at least twice per fight (castsequence, coa then corruption), so I tend to say it exactly the same way every time. "Drain life" too. Somehow I got myself locked into saying "Burn 'Em" (immolate, searing pain castsequence) like a supervillain directing his minions to turn the flamethrowers on the poor defenseless village. But maybe I'm hearing what I want to hear, I'll need to shoot some video some time and see whether I'm really as consistent when dealing with voice recognition as I think.

[Quarto]

Installed Dragon today. Got it working basically, however I cannot get it to send the same command to each instance. Using Keyclone, am I missing something?

If I mouse over an instance, it will send the command to that window. Basically I got the hard stuff done....just gotta make it work for all windows!

[Anozireth]

Installed Dragon today. Got it working basically, however I cannot get it to send the same command to each instance. Using Keyclone, am I missing something?

If I mouse over an instance, it will send the command to that window. Basically I got the hard stuff done....just gotta make it work for all windows!If you're running Vista, you may want to give my program a try. I'm trying to design it specifically for multiboxing, so the ability to broadcast to multiple instances is built in. Click the link in the signature for more details.

[Sarduci]

Installed Dragon today. Got it working basically, however I cannot get it to send the same command to each instance. Using Keyclone, am I missing something?

If I mouse over an instance, it will send the command to that window. Basically I got the hard stuff done....just gotta make it work for all windows!If you are using keyclone then as long as it is not on you DNP list it should be working fine. Unless keyclone requires some specific kind of input that only Rob would know about off hand.

[Deadly1]

I just set this up last night on my 3 box, 5 account setup, it was awesome! Took a little while to get it configured, and if you are modifying the program on the fly you have to remember to reload the template after you make changes, but its fairly simple to figure out, and the tutorial on VRcommander is 'fairly' thorough....

I am right now boosting my other 4 accounts, and placing the Macros for follow, grouping, "accept all", and stop following on voice command has let me quit hitting the other keyboards for anything.... It was nice to walk near my alts and say "follow" have them all snap to attention and start following. "Party", 1 send out the invites, "Accept", they all accept group invite and I get to take off running. Entering instance is a snap, just have to say follow, instantly back on follow again... was very nice, or any other encounter, fear... ect. The program does have to be spoken to very clearly to "hear" whats being told, but for me it opened my eyes to what i could do. I have friends that said Dragon Naturally speaking is better on learning your own voice, and different pitches, so i might swap the program, but the concept is incredible.

I loved it from a boosting perspective it was VERY handy..