home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!cs.utexas.edu!sdd.hp.com!spool.mu.edu!agate!ucbvax!ucdavis!matthews
- From: matthews@eecs.ucdavis.edu (Thomas W. Matthews)
- Newsgroups: rec.audio
- Subject: Re: CD Sound (again?)
- Message-ID: <20597@ucdavis.ucdavis.edu>
- Date: 28 Dec 92 19:31:57 GMT
- References: <1992Dec28.052109.13426@leland.Stanford.EDU> <1hm6geINNhpi@gap.caltech.edu> <24496@alice.att.com>
- Sender: usenet@ucdavis.ucdavis.edu
- Organization: Division of Electrical Engineering and Computer Science, UC Davis
- Lines: 28
-
- In my work with sampled-data detectors, I find it is often necessary to
- clarify something about the Nyquist sampling theorem. Maybe some aspect
- of this principle would be appropriate for the FAQ on sampling.
-
- In my application (not audio), I care only about the value of the input
- waveform at the sample times. I will not be attempting to reconstruct the
- waveform BETWEEN the sample times. Hence, the input signal need not be
- bandlimited to half the sampling frequency.
-
- However, it is the goal of audio devices to reconstruct the input waveform
- from the samples, including the part of the waveform between the samples.
- The Nyquist sampling theorem addresses this problem. In the digital vs
- analog debate, I have heard it said that digital leaves out part of the sound
- (the part between the sample times). It may not be widely understood that
- the goal of digital recording is to reconstruct the entire signal. Maybe the
- FAQ should state that explicitly.
-
- Once the goal of digital recording has been established, one can apply
- the Nyquist sampling theorem to determine a minimum sampling rate depending
- on signal bandwidth, given ideal samplers and ideal filters. Practical
- filters certainly introduce undesirable effects; these can be mitigated by
- increasing the sampling frequency (to decrease required filter complexity) or
- by doing more of ther filtering in the digital domain (certain types of
- filter responses are possible there). The "sinc interpolation" refered to
- in an earlier article can be called filtering - convolution with a sinc
- function results in a "brick wall" filter.
-
- Tom Matthews
-