Announcement

Collapse
No announcement yet.

Speech Recognition in UT3

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Speech Recognition in UT3

    I've heard a lot about it.
    Originally posted by http://www.beyondunreal.com/content/articles/168_1.php
    UT3 will bring major advances in the artificial intelligence that controls UT bots, as well as the way in which players communicate with them. UT3 players will be able to give bots specific orders using their own voices. Whereas in UT or UT2004 you'd use the voice menu to tell your bots to "attack", you can now give orders like "there's a sniper on the tower, go get him". Bots will also respond to particular questions. For example, if you asked "is there anyone in the cave?", they might reply "no, the cave is clear".
    Do you think this will be implemented?
    Sounds nice but complicated.

    And i heard that UT3 engine includes FaceFX
    FaceFX can be used to animate characters by hand or generate lip-synchronization and speech gesture data automatically from audio dialogue. FaceFX uses industry leading VoiceIn® speech recognition technology from Fonix to perfectly synchronize the audio to the mouth movements.
    Basically, this means that if Epic aren't too lazy, we could see characters moving their lips according to that we're saying via microphone in voice chat. (This feature is not approved. Just my guess.)

    #2
    I can't say whether the voice recognition was implemented into the demo or not as i didn't get much of a hands-on, but the recognition was one of the first (if not THE VERY) overhauls to the Unreal universe/engine. I remember reading a post elsewhere that was something along the lines of a gentleman testing out the voice commands told a bot to "go left", to which the bot replied "going left" as it started heading left away from the player. Startled that the bot had figured out what to do the very first time, the guy exclaimed (with mic still on) "He got it right!" After which, the bot started heading right

    As for the FaceFX, that one's iffy. We may see it implemented in the movies, as at least in that effect, a lot of games and cgi movies are trying to get speech patterns and mouth movements to sync up just right and be more realistic. Whether we see it in-game or not is anyone's best guess, but as of now, i'd say no because i would think that the memory requirements to keep track of six characters' lip movements, let alone 32+ (especially for servers) would be enough to break even the best of machines currently built (let alone on the market).

    Comment


      #3
      all indications so far are showing, that the "advanced voice recognition" functionality, descriped in the very early interviews about ut3, has been dropped and is not implemented.

      Comment


        #4
        Originally posted by GuntiNDDS View Post
        all indications so far are showing, that the "advanced voice recognition" functionality, descriped in the very early interviews about ut3, has been dropped and is not implemented.
        For real? And they don't want make "simple" edition of VR? Oh fried crisps!!! They'll pay...

        Comment


          #5
          Originally posted by GuntiNDDS View Post
          all indications so far are showing, that the "advanced voice recognition" functionality, descriped in the very early interviews about ut3, has been dropped and is not implemented.
          So sad. I really thought that this will be so cool when it will be implemented.

          Comment


            #6
            Vent ftw

            Comment


              #7
              Would this take extra CPU or grpahics power to caculate facial movements and such?

              Comment


                #8
                Originally posted by PricklyPoo View Post
                Would this take extra CPU or grpahics power to caculate facial movements and such?
                I think they need to be calculated only if you're 4 or less meters ahead of other player facing him while he talk. So i think no.

                Of course if you meet dozens of smiths with their "Mr. Anderson. We've missed you...", your PC will burn to ashes.

                Comment


                  #9
                  the feature with the lips moving sounds very advanced to me, not sure if its for UT3 or for UT5 lol

                  Comment

                  Working...
                  X