In Reply to: Re: May not be undesirable posted by Dave Kingsland on December 7, 2006 at 09:44:49:
"In his experiments, the reconstruction filter is a perfect 22.05 KHz brick wall, not an FIR filter."I wonder how this could be done.... I guess he could have done a simulation of an "infinite order analog LPF"..... But such a filter would obviously not be feasible in the real world.
If this were to be a simulation of a "perfect" linear-phase digital filter, the oversample rate and output word length would both need to be infinite. (The "first alias" would have to pushed out to infinity. The gradations of the oversampled data would also need to be infinitesimal, necessitating infinite word length.) And an impulse response whose length at least twice as long as the audio track, preferably even longer....
Note this is all in the context of a *perfect* filter.....
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Re: May not be undesirable - Todd Krieger 00:09:23 12/08/06 (1)
- Re: May not be undesirable - Dave Kingsland 09:06:44 12/08/06 (0)