octal-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Octal-dev] Pitch Shifter / Extracting Fourier Description


From: Davíð B Franzson
Subject: Re: [Octal-dev] Pitch Shifter / Extracting Fourier Description
Date: Tue, 16 May 2000 16:28:57 +0100

There is some analysis method called LPC (f i remember it correctly) but it
is waaayayayyyy to slow to use in software on a resonable computer ( needs
some gHz to work) Maybe we can use it in octal y3k
----- Original Message -----
From: <address@hidden>
To: <address@hidden>
Sent: 16. maí 2000 17:11
Subject: [Octal-dev] Pitch Shifter / Extracting Fourier Description


>
>   Steve Mosher wrote:
>
>   [ I have a question,  does  anyone  know how  to write a  reasonable
>   [ pitch-shifter?   I know  that  pitch  shifting  audio  data  (i.e.
>   [ -withoutchanging its speed/duration) is gruesome work,  but it can
>   [ be a good correctional component for many effects.
>
>   Pitch shifting without changing the sampling rate would seem  to re-
>   quire an  identification  of the pitch as  frequency  and  amplitude
>   (Fourier description)  and then changing the frequency in the sampl-
>   ing generate phase (reverse Fourier transform).
>
>   I have been experimenting with code to identify the complete Fourier
>   description  of a sample sequence.   Does anyone on the list know of
>   code to perform this process?  Some of my observations are:
>
>   (1) The Fast Fourier Transform  only  detects  frequencies  in whole
>       numbered values  of which useful frequencies in general are not.
>       What is required is a more precise frequency detection sequence.
>       The full Fourier transform can be used to more precisely analyze
>       fractional ranges after the FFT identifies nearby high-amplitude
>       areas.
>
>   (2) The precision of frequency identification correlates to the num-
>       ber of samples used.  I.e., though one may address any frequency
>       below half the sampling  rate,  the degree  of distinction  from
>       nearby  (fractional) frequencies depends  on how long  (how many
>       samples in which) the frequency is present,  and would appear to
>       be an aspect fundamental to all frequency detection systems such
>       as human hearing.
>
>   Neil Nelson address@hidden
>
> _______________________________________________
> Octal-dev mailing list
> address@hidden
> http://mail.gnu.org/mailman/listinfo/octal-dev



reply via email to

[Prev in Thread] Current Thread [Next in Thread]