Subject: Re: Proposal: Audio: Sampling rate conversion
To: TAMURA Kent <kent@netbsd.org>
From: Wojciech Puchar <wojtek@chylonia.3miasto.net>
List: tech-kern
Date: 03/04/2002 17:52:43
>     on 02/03/04, Lennart Augustsson <lennart@augustsson.net> writes:
> > Can you describe the algoritm you use to perform the sample rate
> > conversion?
>
> Simple resampling like "dst[i] = src[i*srcrate/dstrate]" is used
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

this generates very bad outputs - frequencies between dstrate/2 and
srcrate/2 is "mirrored" as dstrate-(srcrate/2) to dstrate/2.

unfortunately there is no better way to do it without overloading CPU