[music-dsp] OT: CD Perfection

Joshua Scholar joshscholar at yahoo.com
Fri Oct 15 01:17:56 EDT 2004


I'm not an expert but here's my 1/2 expert opinion :)

44k and 48k sampling have nyquest frequencies too close to 20k for analog
filtering to work very well, so you have to do as you said at a higher
sampling rate, and then refilter in the digital domain and downsample.  Keep
in mind that high quality analog filtering requires expensive precision
componants and even then probably isn't very good compared with software
filtering - and software filtering is always perfect and only requires
sufficient processing power.  It's much cheaper and much more reliable to
depend on software.

Early equipment did as you said at lower sampling rate, and I'm sure that
some really cheap sound cards work that way - and the results are bad, often
very bad.

Oversampling in A to D is something else.

By the way I suspect that the preference for digital filters in A to D may
have some historic reason.

When CD players first came out, highfi magazines used phase-delay between
the channels and within a channel as one of the primary measurements of the
quality of phonograph cartridges.  CD manufacturers wanted all of their
tests to come out perfect in the reviews, so they prefered linear phase
filtering which can only be done in the digital domain to avoid the latter.

High end CD players had dual D to A converters just to avoid having a phase
delay measurement between the channels show up in the reviews (even though
the actual result of a 1/88000 of a second delay is a coherent channel delay
which moves the stereo image over by a fraction of an inch but doesn't
affect the sound otherwise)...

A cheaper solution than dual converters was to oversample a lot, and
interleave the channels on the oversample giving a smaller measured delay
between the channels.

----- Original Message ----- 
From: Philip Mcleod
To: a list for musical digital signal processing
Sent: Thursday, October 14, 2004 5:32 PM
Subject: Re: [music-dsp] OT: CD Perfection


I thought A/D converters first used an analog filter to band limit the
signal to under the nyquist limit of the sampling rate. Then just sampled
at the required rate.
Maybe oversampling is used in D/A conversion though? Doesn't Analoge CD
used 8 times oversampling?
Is there any experts who can tell us the real story of DAC's and ADC's in
modern sound cards. (or if you are an expert Tony, then I just don't know
what I am talking about)

Phil

On Thu, 14 Oct 2004, Bob Cain wrote:

>
>
> Tony Scharf wrote:
>
> > the key is right in that paragraph.  ANY downsampling I have ever heard
has
> > resulted in a greater degree of signal degredation than simply recording
at
> > the lower rate.
>
> All audio A/D converters oversample, filter and downsample
> within the device.
>
>
> Bob
> --
>
> "Things should be described as simply as possible, but no
> simpler."
>
>                                               A. Einstein
> --
> dupswapdrop -- the music-dsp mailing list and website:
> subscription info, FAQ, source code archive, list archive, book reviews,
dsp links
> http://shoko.calarts.edu/musicdsp
> http://ceait.calarts.edu/mailman/listinfo/music-dsp
>

--
dupswapdrop -- the music-dsp mailing list and website:
subscription info, FAQ, source code archive, list archive, book reviews, dsp
links
http://shoko.calarts.edu/musicdsp
http://ceait.calarts.edu/mailman/listinfo/music-dsp




More information about the music-dsp mailing list