Menu

From Speech Recognition to Lie Recognition?

2015-03-18
2015-03-20
  • Travis Banger

    Travis Banger - 2015-03-18

    There must be folks around here who know a lot about Voice Identification? Actually, the preferred term seems to be Speaker Identification.

    Can anybody please help with this project?:

    http://goo.gl/mREmd2
    [Click into the folder "O'Reilly" and then into "spectrograms"]
    [For those unfamiliar with Google Drive, you only click ONCE]

    Additionally see the folders named "Learning Material".

    The first question that I have is: how to normalize the formants, trying to make them identical?

    If this is not the proper venue (what says you, Nickolay?), you may e-mail me to ramon at patriot dot net

    Spasibo.

     

    Last edit: Travis Banger 2015-03-18
    • Nickolay V. Shmyrev

      Hi Travis

      The subject is fun but I don't see any technical possibility to investigate here.

      The first question that I have is: how to normalize the formants, trying to make them identical?

      There is probably no such way to "normalize". If you need to compare two tracks you might extract some features from them, but in general it is not easy to do.

      If this is not the proper venue

      From technical point of view this might be interesting. Relevant link is

      http://languagelog.ldc.upenn.edu/nll/?p=3554

       
      • Travis Banger

        Travis Banger - 2015-03-20

        Hi Nick:

        There was a serious confusion, and I am afraid it was caused by my ambiguous subject line. I will repost.

         

Log in to post a comment.