In Reply to: This Wasn't as Bad as Expected... posted by Todd Krieger on August 25, 2003 at 21:31:58:
"If there is no oversampling, there is no filter function the base media signal is convolved with."Yes there is ...
Theory holds only for a sample train implemented as a train of Dirac
impulses, which is clearly ever so slightly impossible in the known
universe. Practical DACs do no output impulses, but stable levels each with a duration 1/fs. This aberration from theory has a frequency
domain representation which follows a rolloff with the shape of Sin(f)/f. Which happens to be Sinc(f).This may seem to be annoying, as it gives -2dB at 20kHz for a 44.1k system, and associated phase shift (or rather group delay is what we're after, phase shift *can* be quite harmless). On the other hand, this alone is enough to suppress the high-activity region of the first and second image, say 40-48kHz, in a way sufficient for NONOS DACs to get away without any filtering at all.
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- Re: This Wasn't as Bad as Expected... - Werner 08/25/0323:34:54 08/25/03 (3)
- ZOH... - csown 16:50:11 08/26/03 (2)
- Re: ZOH... - Werner 22:32:08 08/26/03 (1)
- Re: ZOH... - csown 07:43:04 08/27/03 (0)