Announcement

Collapse
No announcement yet.

I have the most powerful speech control interface.

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

  • I have the most powerful speech control interface.

    Hi,

    AFAIK I have the most powerful speech control interface core technology. (I worded this subject to get your attention, not to be arrogant.) I'm looking for software developers to join me unless someone has developed core technology better than mine. (If you have better technology maybe I'll join you.) I'll also share my technology in exchange for technology which helps my project.

    Here is a list of some of the technical features of my software. I don't have these features listed on my website because I don't want competitors to steal my ideas or some large corporation to sue me for patent infringement of their trivial ideas - e.g. Amazon 1-click patent.

    technical specs of my software (partial list)
    --------------------------------
    - My software handles very large collections
    - Startup time is < 3 sec. with a music collection of 14,000 songs, 1500 albums, and 800 artist
    - the .exe (executable) file is only 144k.
    - my hardware - athlon 2000, 1gb sdram, udma 7200rpm hd
    - My software should easily be able to handle a collection at least ten times this size.
    - Music data is loaded from Itunes's xml file (Itunes takes much longer to startup than my software)
    - Speech recognition of song, album, artist, and genre names is higher than all other SR jukeboxes because I have developed an intelligent algorithm to alter the names so that the computer will recognize them. Here is one simple example: My software renames 'Crosby, Stills & Nash' to 'Crosby, Stills and Nash'
    - I can call up individual songs, albums, and artists
    - I'm using winamp to play the music (The Itunes api sucks. I'm still trying to decide which player to use.)
    - My software is written in C++ and embedded Perl and uses SAPI 5.1, xml, and windows xp
    (The only reason I am using windows is because they have by far the best SR engine, AFAIK.)
    - TTS (text-to-speech) replies for nearly all commands
    * I'll be adding AT&T natural voices by the end of this month.
    --------------------------------

    I've also done experimentation with hands-free microphone technology. I can get >90% accuracy with an omni mic. in a quiet room standing ten feet away. In a noisy, reflective room (like a coffee shop) I can get >90% from about two feet away.

    I'm developing my technology for whole house audio systems, but I am also interested in plugging it into car pc systems. I'll put my technology anywhere I can make money.

    Let me know if you are interested in learning more about my project or if there is another project comparable to mine. I saw this thread, and checked each of these websites to find out if their software is comparable to mine - "List of front-ends/other useful apps"http://www.mp3car.com/vbulletin/showthread.php?t=34218. I didn't see any projects with much similarity to mine.

    Thanks,
    Jeff
    http://IntelligentJukebox.com
    [email protected]
    "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo, 1802-1885

  • #2
    Intelligent Jukebox Command Reference

    All these commands are working except "open genre", which will be working within a couple days. (Please do not distribute this, or I'll have to send someone to liquidate you. )

    Intelligent Jukebox Command Reference
    ------------------------------------

    Attention commands:

    [hello] computer
    stop listening (Disables all commands except the above.)
    Quit this program (closes the program)

    ------------------------------------
    Music selection commands:

    Open artist <artist>
    Open [album] <album> (must be called after "open artist")
    List albums (lists albums for currently opened artist)
    Open [album] number <n> (opens an album of the artist by number)
    E.g. "Open album number fourteen."
    Open [song] <song> opens a song for the last album, artist, or genre selected
    Open genre <genre> opens a genre (you can open a song or artist from this genre afterwards)
    enqueue [the music] (queues up the currently selected album or song into the playlist)

    ------------------------------------
    Winamp voice commands:

    previous [track]
    next [track]
    play [the music] (starts or unpauses the music)
    * if "Open artist" or "Open album" was the last command it will
    clear the playlist and load the album
    * if "Open song" was the last command it will
    clear the playlist and load the song
    pause [the music] (Say "play the music" to unpause.)
    stop [the music]
    clear [the playlist]
    toggle shuffle

    -----------------------------------------------------------
    Other commands:

    Tell me about my music collection
    reply: "Your music collection has <n> artists, <n> songs, and <n> songs."

    -----------------------------------------------------------
    Syntax Notes:

    - Words with braces around them (e.g. [hello]) are optional, i.e. you
    don't have to speak them. To get the computer's attention you can say
    "Hello Computer" or "Computer"
    - You should speak the option words if the speech recognition is not
    accurate enough for you
    - Words surrounded by "less than" and "greater than" signs (e.g.
    <artist>) need to be filled in with another word or phrase. E.g. "Open
    artist <artist>" becomes "Open artist The Beatles."
    "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo, 1802-1885

    Comment


    • #3
      Nice, sounds like something I'd definitely like to use in the home to increase my media center PC's capabilities. I've been playing with ol33l's software as well as one called Realize Voice, and done the whole training and tweaking bit to get it as accurate as possible. I'm curious on how well the voice recognition input works while music is playing etc. I always see this as an ongoing battle.

      The one thing I'm considering is some kind of wireless mic that operates on bluetooth or something. Something small and discrete while I could utilize in both the house and car and have it automatically connect to either depending on it's location. This puts the mic in a nice, close proximity to the user and should allow it to be more louder and clearer then my desk mic's or car mic. Have you thought of, or come across any good products to use for this type of input?? Most bluetooth headsets for cell phone don't provide the high fidelity I'm looking for, so I'm waiting for technology to catch up...
      -GiZak-
      Carputer Project: 98% Complete!! *tweakin'*
      Xenarc 7TSV | Opus 150 | Biostar M7NCG400 | Athlon Mobile2200 | Seagate 160Gb | Kingston 2x256mb | BT Holux GR-231 GPS | LiteOn CD-R/DVD| Keyhole Rev. Cam | Zippy Mini-glow Keyboard

      Comment


      • #4
        WoW man This definately sounds gr8. Hope u r done with it soon...
        FOR OUR SAKE !!!!!!!!!!!!!!!!!!!!!!!!
        LOL
        good work
        Commell LV-677 DC 6 x 6" Mother Board
        2.6 GHz; 1 Gb Ram
        120 Gb Laptop HD
        7" InDash Xenarc MTD - X7000 Screen
        Gyration Mouse ; GPS Mouse ; ELM Scan ; TPMS
        Parrot MK 6000 & iPhone
        ----------
        Re done the carputer...Into the Glove Box !!

        Comment


        • #5
          How did you get iTunes to automatically export the xml
          PowerVoice v1 | NaviVoice Source
          GammaControl v2.4
          SKINbedder v3

          1995 Lexus SC300

          Comment


          • #6
            Originally posted by 0l33l
            How did you get iTunes to automatically export the xml
            Exporting the data was easy.

            copyItunesXml.bat:
            --------------------------------
            copy "%USERPROFILE%\My Documents\My Music\iTunes\iTunes Music Library.xml" "<IJ exe dir>\xml\iTunes Music Library.xml"
            --------------------------------
            Parsing and transforming the data took a lot of work. I probably spent more than 60 hours on this and used knowledge from my senior level computer science courses.
            "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo, 1802-1885

            Comment


            • #7
              not sure if i missed this in your post, but do you have to do any kind of voice training?
              01101100 01101001 01110001 01110101 01101001 01100100 01011111 01110011
              01101101 01101111 01101011 01100101

              beer replenishment fund
              http://www.mp3car.com/vbulletin/forumdisplay.php?f=93
              mp3car live search

              i have joost invites, just hit me up for one.

              Comment


              • #8
                Originally posted by liquid_smoke
                not sure if i missed this in your post, but do you have to do any kind of voice training?
                Yes, since this uses SAPI.
                PowerVoice v1 | NaviVoice Source
                GammaControl v2.4
                SKINbedder v3

                1995 Lexus SC300

                Comment


                • #9
                  Originally posted by kykeon
                  Exporting the data was easy.

                  copyItunesXml.bat:
                  --------------------------------
                  copy "%USERPROFILE%\My Documents\My Music\iTunes\iTunes Music Library.xml" "<IJ exe dir>\xml\iTunes Music Library.xml"
                  --------------------------------
                  Parsing and transforming the data took a lot of work. I probably spent more than 60 hours on this and used knowledge from my senior level computer science courses.
                  I mean how do you have it automatically created
                  I've got my music in the same folder, and I can't find an XML file anywere there or in the iTunes folder

                  [edit]: Never mind... I found it
                  PowerVoice v1 | NaviVoice Source
                  GammaControl v2.4
                  SKINbedder v3

                  1995 Lexus SC300

                  Comment


                  • #10
                    beta release

                    I'm going to try to get a beta release done by the end of this month. I hope to get a few people here to test it.

                    I'll let you all know as soon as the beta is ready.
                    thanks
                    "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo, 1802-1885

                    Comment


                    • #11
                      Surely if you are using the sapi then your recognition is as good or bad as anyone elses?

                      Comment


                      • #12
                        oooo competition for Navivoice... Boris get working harder!

                        Comment


                        • #13
                          Originally posted by ElKeeed
                          Surely if you are using the sapi then your recognition is as good or bad as anyone elses?
                          Try getting any jukebox software out there to recognize AC/DC or any other name which is not obvious to the computer how to pronounce.
                          "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo, 1802-1885

                          Comment


                          • #14
                            Originally posted by kykeon
                            Try getting any jukebox software out there to recognize AC/DC or any other name which is not obvious to the computer how to pronounce.
                            Not to be nitpicky, but changing the names to "A C D C" is not really revolutionary.

                            But the idea is sound, surprise no one has done it before.
                            [H]4 Life
                            My next generation Front End is right on schedule.
                            It will be done sometime in the next generation.
                            I'm a lesbian too.
                            I am for hire!

                            Comment


                            • #15
                              Originally posted by Scouse Monkey
                              oooo competition for Navivoice... Boris get working harder!
                              Yeh, yeh yeh.
                              PowerVoice v1 | NaviVoice Source
                              GammaControl v2.4
                              SKINbedder v3

                              1995 Lexus SC300

                              Comment

                              Working...
                              X