Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/5/84; site mordor.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!ucbvax!ucdavis!lll-crg!mordor!sjc
From: sjc@mordor.UUCP (Steve Correll)
Newsgroups: net.audio
Subject: Re: nut.audio: The \"ear\" vs. \"the instrument\"
Message-ID: <3785@mordor.UUCP>
Date: Fri, 4-Oct-85 14:05:53 EDT
Article-I.D.: mordor.3785
Posted: Fri Oct  4 14:05:53 1985
Date-Received: Sat, 5-Oct-85 14:55:45 EDT
References: <625@decwrl.UUCP>
Organization: S-1 Project, LLNL
Lines: 33

> Those CD cables probably just have a rolled off high-end.  Perhaps one
> or both cables has frequency dependant phase delay to compensate for a
> single DAC machine, but I doubt it.

Using a single DAC in a CD player does not introduce
frequency-dependent phase distortion ("group delay distortion").  When
one channel advances to the next sample 1/(2*44100) second later than
the other, the delay is constant and independent of the frequency which
the samples are representing. Imagine (I fear people are getting tired
of reading this) that you are using a two-DAC CD player but you are
sitting 1/3 inch closer to one speaker than to the other. The sound
from the farther speaker will, due to the finite propogation speed of
sound in air, be delayed exactly the same amount as if the player were
using only one DAC.  This delay is independent of frequency. If a 20kHZ
sinusoid and a 20Hz sinusoid cross the zero axis at the same instant,
and you delay each of them by 1/(2*44100) second, they will still cross
the zero axis together.

An unscrupulous audio manufacturer can make this appear detrimental by
expressing the delay in degrees rather than seconds, since one degree
of a 20kHz sinusoid is shorter than one degree of a 20Hz sinusoid, but
that's like comparing feet and meters. If a CD player did somehow
manage to delay both tones by exactly the same number of degrees, then
it would be introducing frequency-dependent delay with a vengeance!

The phase distortions which *do* occur in the CD process result from
filters which delay some frequencies more (when measured in seconds)
than others, causing our 20kHz and 20Hz sinusoids no longer to cross
the zero axis together.

-- 
                                                           --Steve Correll
sjc@s1-c.ARPA, ...!decvax!decwrl!mordor!sjc, or ...!ucbvax!dual!mordor!sjc