Speech Recognition is a feature that allows you to communicate with your bots, and give them orders using your voice (and a mic, of course). However, this feature is not functional by default in the demo version of UT2004. Here's a run-down of what you'll need to do to get it working in the UT2004 demo. Thanks to Sir_Brizz for providing some of this information on our forums. First, install the Microsoft Speech SDK. It can be found over on this page. Download the 68MB 'Speech SDK 5.1' on the 'Download' sidebar. Next, install the SDK, fire up the Windows Control Panel and look under 'Speech'. Make sure your mic is correctly set-up and click 'Train Profile'. You'll be asked to say a number of paragraphs so that the speech recognition software can calibrate itself to your voice. After that, you can train the software as many times as you like (it'll get more and more accurate the more you train it). Next, hop into the UT2004 demo. Make sure 'Speech Recognition' under 'Settings | Game | Misc' is checked. Then make sure you've got an 'Activate Microphone' key set under 'Settings | Input | Configure Keys'. You'll need to hold this down while giving your bots orders. Now you're done! Load up a map and start talking to your bots. Each bot is given a callsign from 'Alpha' through to 'Oscar'. This callsign will appear next to the bot's name above its head when you're close enough, or you can view the callsigns for all your bots on the scoreboard by pressing F1. You should address your bots by their callsign rather than their name. Alternatively, you can use 'everyone' or 'everybody' to address all bots on your team. Each gametype has a different set of bot voice commands. The example list below is from Onslaught. Look in your 'UT2004Demo\Speech' directory for voice profiles for the other game types, although most are identical.

  • 'Defend' - Tells the bot to guard your base.
  • 'Attack' - Sends the bot to attack the enemy base.
  • 'Cover Me' - Tells the bot to watch your back.
  • 'Freelance' - The bot will go off and do what it wants.
  • 'Taunt' or 'Dance' - The bot will do a little jig for you.
  • 'Get Out' - If the bot is in a vehicle, it will exit that vehicle.
  • 'Status' - The bot will tell you what it's doing.
  • 'Gimme' - The bot will throw its current weapon at you.
  • 'Suicide' - The bot will die and respawn.
  • 'Jump' - Boing! The bot will jump on the spot.
  • 'Hold (this) Position' - The bot will stop and defend its current location.

As you can see, most of the commands are fairly easy. To tell a bot to do something, simply say its name and then a command while holding down your chosen 'Activate Microphone', for example...

  • 'Alpha, Defend'
  • 'Echo, Gimme!'
  • 'Everyone, Status'

That's it! If a bot doesn't recognize your command, or gets it wrong, you can always keep training the Speech SDK using the method detailed above.