James Chandler Jr
jchandjr at bellsouth.net
Sun Feb 27 18:30:58 EST 2005
Had been reading on sinc halfband filters, which are -6 db at the halfband Fc
with symmetrical shape on both sides of Fc (when viewed on a linear
freq-vs-amplitude plot). Just assumed that a polyphase IIR halfband filter would
look the same, but obviously IIR filters don't often have such tidy shapes.
----- Original Message -----
From: "Erik de Castro Lopo" <erikd-music-dsp at mega-nerd.com>
> On Fri, 25 Feb 2005 11:46:15 -0500
> "James Chandler Jr" <jchandjr at bellsouth.net> wrote:
>> Erik de Castro Lopo had earlier mentioned that a half-band filter, regardless
>> its sharpness and band-stop attenuation, is only -6 dB at the half band freq.
>> Far from perfect for up/downsampling, if the audio really will contain
>> substantial high frequencies.
> This is right for the upsample by 2 case (although I think it may
> only be -3dB, not -6dB).
> In the more general resampling case (up and down sampling) the anti
> aliasing filter must have its stop band extend all the way down to
> the lowest possible alias frequency.
> In the case of Dave's half band filter, the -3db point is at the
> lowest possible alias filter and the filter's transition band
> extends into the alias frequencies.
More information about the music-dsp