home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: rec.audio
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!ames!sun-barr!decwrl!pa.dec.com!engage.pko.dec.com!nntpd.lkg.dec.com!ryn.mro4.dec.com!3d.enet.dec.com!roth
- From: roth@3d.enet.dec.com (Jim Roth)
- Subject: Re: Anti-aliasing on the recording end?
- Message-ID: <1993Jan25.033014.7033@ryn.mro4.dec.com>
- Sender: news@ryn.mro4.dec.com (USENET News System)
- Organization: Digital Equipment Corporation
- Date: 24 JAN 93 22:24:54
- Lines: 32
-
-
- In article <shetline-230193113049@128.89.19.90>, shetline@bbn.com (Kerry Shetline) writes...
- >In article <1993Jan23.073000.13305@news.columbia.edu>,
-
- >The advantage of recording with a higher sampling rate is that you don't
- >need steep analog input filters. But this does mean that there can be
- >frequencies greater than 22.05 emmbedded within the data stream. So you
- >can't just throw away the extra samples. The data has to be processed to
- >remove the higher frequencies.
-
- This is *NOT* the reason for oversampling A to D converters. It's in fact
- a rather minor side benefit.
-
- Also note well that while it was fashionable to blame minimum phase "brick
- wall" filters for all that was wrong with digital audio in the early days,
- there are just as steep filters in oversampling ADC's, except that they're
- implemented digitally, and that the phase distortion of minimum phase
- antialias filters is inaudible. You need about an order of magnitude
- more group delay dispersion than typical analog minimum phase filters
- have to start having a problem. The problems with old style filters
- had nothing do do with group delay, or the fact that they're minimum
- phase.
-
- (This fact about filtering can be easily demonstrated with some DSP
- hardware: record some audio and process it with a minimum phase digital
- filter (say, a recursive filter) twice; then process it again twice
- except in time reverse on of the passes. The second signal will have
- the same amplitude response but clearly linear phase. Now see if you
- can hear a difference. You won't unless the phase distortion is *really*
- horrible, like from making many passes thru the filters...)
-
- - Jim
-