home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!munnari.oz.au!ariel.ucs.unimelb.EDU.AU!ucsvc.ucs.unimelb.edu.au!lugb!news
- Newsgroups: rec.audio
- Subject: Re: CD SOund QUality
- Message-ID: <1992Dec29.024815.29836@lugb.latrobe.edu.au>
- From: MATGBB@LURE.LATROBE.EDU.AU (BYRNES,Graham)
- Date: Tue, 29 Dec 1992 02:48:15 GMT
- Sender: news@lugb.latrobe.edu.au (USENET News System)
- References: <1h17e4INNrkv@usenet.INS.CWRU.Edu> <24459@alice.att.com>
- <1992Dec21.213820.2737@cbnewsh.cb.att.com> <vanz.029i@tragula.equinox.gen.nz>
- <24474@alice.att.com> <shetline-271292021613@128.89.19.80>
- Organization: La Trobe University
- In-Reply-To: shetline@bbn.com's message of 27 Dec 1992 07:35:40 GMT
- X-News-Reader: VMS NEWS 1.24
- Lines: 44
-
- In <shetline-271292021613@128.89.19.80> shetline@bbn.com writes:
- > First, I just want to assure you that I have no problems with CD sound, and
- > I
- > believe that I've got a good idea what the Nyquist theorem is about at a
- > qualitative, if not mathematical, level. But I do have a question for those
- > who
- > understand it better:
- >
- > I wonder what happens to higher frequency signals as they approach half the
- > sample rate? It seems to me that at *exactly* half the sample rate, the
- > amplitude of a signal would be entirely dependent upon its phase
- > relationship
- > with the sampler, going from maximum amplitude when samples were taken at
- > the
- > peeks and troughs, down to nothing if the samples happened at
- > zero-crossing.
-
- Absolutely correct. You cannot accurately reproduce a signal sampled at
- precisely double its frequency. But for anything less, you can.
- >
- > It seems then that slightly lower frequencies would slide in and out of
- > phase
- > with the sampler.
- In a sense, it is this "sliding in and out of phase" that allows the sampling
- to work: there is no point always sampling at the same point on the waveform.
-
- > What, if any, audible effect could this produce? Would
- > low-
- > pass filters eliminate such an effect? Wouldn't the end result be like
- > amplitude modulation, producing a spread of frequencies centered on the
- > original
- > signal frequency, +/- the beat frequency?
- >
- Hmmm, you have almost fallen over it. You get a spread of freq's centred on
- 1/2 the *sampling* frequency. -beat gives the original recorded tone (yes!),
- +beat gives the (first) aliase tone. The job of the low pass filter is to
- get rid of the latter, while keeping the former.
-
- > My apologies if this has already be brought up a sickening number of times.
- >
- > -Kerry
- Well, at least you brought it up in an original way as a question, rather
- than as a pseudo fact (eg "it is obvious that it can't work blah blah blah)
- regards, Graham
-