Re: can anyone help with this one...

["James W. Beauchamp" <jwbeauch@xxxxxxxxxxxxxxxxxxxxxx>]

Brittany,

Probably what is happening is that in the "transposed" speech, the
amplitudes of the components are the same, but their frequencies have
been shifted by a fixed multiplier. A four semitone upward shift would
require a frequency multiplier of 2^(4/12) = 1.26. This will result in
the natural formant resonances being shifted up, changing the quality 
of the voice considerably, such that an adult male voice can sound 
more like an adult female or child voice.

This problem could be fixed using software that keeps the spectral
envelope of the sound close to that of the original. If the program
models speech in terms of an glottal excitation, which changes its
frequency, and a vocal tract modulation, which is invariant with
respect to the original, that may be what you want. The method of
Homomorphic Speech Modelling does something like this. A talk on it
is given at:

http://www.ece.ucsb.edu/Faculty/Rabiner/ece259/digital%20speech%20processing%20course/lectures_new/Lecture%2012_fall_2010.pdf

Pretty mathematical though.

The question is Is there software out there that will do this?
I actually think that the free download program Praat might do it
because it inherently models signals in terms of excitation and
filter. Maybe someone on this list knows of alternative software
products.

Jim

You wrote:
>From: Brittany Guidone <brittanyguidone@xxxxxxxxx>
>Date: Tue, 6 Sep 2011 08:17:59 -0700
>To: AUDITORY@xxxxxxxxxxxxxxx
>Subject: [AUDITORY] can anyone help with this one...
>
>Dear List,I am designing a study in which I planned to "morph" female
> voices into male voicesand vice versa. The method I tried was simply to
> change the pitch (using GarageBand,although I also tried Audacity). At
>first I tried small transpositions, just  3-4 semitones up or down.
> Our task is to identify individual speakers. The problem is that 
>the speakers were easily discernible before the transformation, however,
>after the transposition, none of our pilot subjects were able to 
>differentiate between the various speakers. Can anyone help me 
>understand what is going on? What is known about the acoustic information 
>people use identifying individual speakers? Can anyone suggest a better 
>transformation? Thanks in advance,Brittany 

>From: Brittany Guidone <brittanyguidone@xxxxxxxxx>
>Date: Tue, 6 Sep 2011 15:56:06 +0000
>To: AUDITORY@xxxxxxxxxxxxxxx
>Subject: [AUDITORY]
>
>Dear List,
>
>To explain more about my original question:
>
>I mainly want to know if "morphing" or "changing" the pitch of a male
>voice by p ositive 4 semi tones will change the components of the
>original male voice (befo re it was morphed) in a way that will make the
>two voices have different compone nts or "make up" in comparison to one
>another, besides the fact that they will h ave different pitches.
>
>In other words, when the pitch of a voice is changed in audacity or
>garage band then what other components of the voice are changed (besides
>the pitch shift)?
>
>-Brittany 
>Sent from my Verizon Wireless BlackBerry