1. Forum
  2. Usenet
  3. SCI.IMAGE.PROCESSING

  • VOICE RECOGNITION BASED ON SPECTROGRAM

    From manaswi.navin@gmail.com@21:1/5 to All on Sat Jan 7 13:14:32 2017
    Can we find out total number of speakers and their duration by looking at/analysing spectrogram.!
    [image description] (https://drive.google.com/drive/folders/0B4rwzcsr5hevdEJlam9scTRodTg)

    By just looking at the image, I can see some pattern, but I am looking for right solution in terms of opencv code(python)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Martin Leese@21:1/5 to manaswi.navin@gmail.com on Sun Jan 8 12:18:28 2017
    manaswi.navin@gmail.com wrote:
    Can we find out total number of speakers and their duration by looking at/analysing spectrogram.!
    [image description] (https://drive.google.com/drive/folders/0B4rwzcsr5hevdEJlam9scTRodTg)

    In general, no. Different speakers can
    use similar frequency ranges, so frequency
    doesn't work for this. The ear/brain uses
    spatial processing (search for "cocktail
    party effect").

    By just looking at the image, I can see some pattern, but I am looking for right solution in terms of opencv code(python)

    I can't, because I do not have permission
    to view the file.

    --
    Regards,
    Martin Leese
    E-mail: please@see.Web.for.e-mail.INVALID
    Web: http://members.tripod.com/martin_leese/

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)