[SciPy-User] bandlimiting sample sequence

Jerry lanceboyle at qwest.net
Wed Jan 12 00:54:32 EST 2011


You should be careful here. As I understand it, your samples have been  
generated by a program ("soft-generated"). ___THEY ARE ALREADY  
BANDLIMITED.___ Any and all aliasing has already been done and you  
can't do anything about it. For example, say you want to generate  
samples of a square wave so you make a sequence like for example -1 -1  
-1 1 1 1 -1 -1 -1 1 1 1 and so on. Those are samples of a non- 
bandlimited analog square wave and the aliasing of what would be out- 
of-band harmonics is already done and there is nothing that you can do  
about it. Make such a wave and play it as an audio file and you will  
hear what I mean. Low pass filtering will not remove the aliasing.  
What you need to do is to construct a bunch of truncated sinc  
functions of the proper weights--that will give you approximate  
samples of a bandlimited square wave--the approximation depends on how  
long your sinc function is and how you truncate it--a nice window is  
good. Look up windowed sinc functions.


On Jan 11, 2011, at 3:39 AM, Andrew Gaydenko wrote:

> Hi!
>
> Say, I have given sample rate Fs and a sequence of (soft-generated)  
> samples.
> The aim is to limit band below Fs/2. I mean any limiting params are  
> also given
> (for example, attenuation at 0.9 * Fs/2 is -1db or less, attenuation  
> at Fs/2
> must be -60db or more, gain oscillation in some frequencies range  
> (say, from
> 0Hz to 0.8 * Fs/2) must be in +/-0.01db range.
>
> Has anybody examples (or references to publically available  
> examples) I can
> take to understand how to band limit a sequence decribed way?
>
>
> Andrew
> _______________________________________________
> SciPy-User mailing list
> SciPy-User at scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-user
>




More information about the SciPy-User mailing list