[MUD-Dev] [TECH] Voice in MO* - Phoneme Decomposition and Reconstruction

Hans-Henrik Staerfeldt hhs at cbs.dtu.dk
Wed May 29 10:44:02 New Zealand Standard Time 2002

On Fri, 24 May 2002, Eli Stevens wrote:

>         ***
>       **   *
>     **      ****          **
>    *            **       *  *
>   *----------------*----*----*-------------*  Repeats...
>                     ** *      *           *
>                       *        **        *
>                                  *      *
>                                   **  **
>                                     **

> What I would like to know is if there is a good way (i.e. not
> brute force) to figure out when a wave has finished and is now
> repeating.  In a perfect world, you could compare the 1st to the
> 2nd, 3rd, 4th etc. until a match was found, then increment to the
> 2nd and the match + 1, compare, and they too would match up.

> I suspect that someone has already solved this problem (it seems
> simple enough to be fairly common).  However, I have had a hard time
> coming up with decent search terms.  :/ Anyone?  Anyone?  :)

I believe that you want to calculate the auto-correlation function
of a window of the sample. This will tell you the wavelengths that
are most strongly expressed in the sample window. As far as i know,
you can use FFT to do this for speedup. Then select the strongest
peak, and this will be the stongest period observed.


To speed up processing, you could move the window by the period
found so you don't have to do it for each sample. Remember to choose
your window length to accomodate the full waveform of the longest
period you want to detect.

please anyone correct me anyone if i made a blunder here. :-)

Hans Henrik Stærfeldt   |    bombman at diku.dk    | work:  hhs at cbs.dtu.dk      |
Address:                |___  +45 40383492    __|__       +45 45252471     __|
DTU, Kemitorvet,        | Scientific programmer at Center for Biological     |
bygn 208, CBS.          |  Sequence Analysis, Technical University of Denmark|

MUD-Dev mailing list
MUD-Dev at kanga.nu

More information about the MUD-Dev mailing list