shauns@vice.UUCP (Shaun Simpkins) (01/14/84)
I originally tried to post this as a response to
Subject: Re: differences in sounds of CD players
but I couldn't get the hang of our &*&$%!!!postnews.
Yes, exactly. There are serious problems with the slick mag's test reports,
and to the uninitiated they are just plain confusing. Some of this
inappropriateness stems from the limited scope of their test disc, which
apparently has been superseded by one with more variety.
A second reason may be Steven Jay Gould calls `scientific prejudice', a
tendency to use only those facts which support one's predispositions, a
phenomenon often noted on this net.
With all this discussion about time response vs. magnitude (frequency) response,I find it interesting that much-vaunted LP cartridges, when subjected to
the 1KHz square wave test, often exhibit far WORSE transient response than the
worst CD players I've seen, with atrocious overshoot and ringing, albeit at
at a higher frequency than the CD. This then calls into question the
`betterness' of the LP system; the differences that may be perceived, that
we have been conditioned to think of as better, may actually mean the LP
cartridge is worse.
I question, too, what the relevance of 40KHz ringing is to a receiving system
that limits at 20KHz, and therefore WHY magazines even bother to print such
pictures. A fairer test would be to accurately determine the response of the
ear and apply postfiltering to the recovered signal to find out what the
ear ACTUALLY HEARS through an ideal amplifier/speaker chain.
Following this a little further, transient response figures at just one
amplitude are not enough to ascertain the linearity of a transducer. What is
the amplitude of the LP step? 3.54cm/sec? 1cm/sec? System response can
change dramatically as it approaches its dynamic range limits - and these limits
vary with frequency. Look at the frequency response vs. recording level plots
on a tape recorder, for example. How does the transducer behave when it
overloads? Does it round over gracefully or
does it clip (or, perish the thought, bounce around a bit), throwing painful
hash into the output? Does it slew at high amplitude/high frequency?
Perhaps a useful test result would be a graph of the
transducer's linearity vs. amplitude AND frequency, or a Fourier
decomposition of step responses of varying amplitude.
I leave it to you to figure out whether Joe Average could understand this,
either.
It is quite obvious from the test reports that the CD would win hands down
in a linearity contest with a LP cartridge. At high
frequencies, the CD is essentially dependent upon the ideality of its
reconstruction filter and line amplifier, while the cartridge heads into
complex electromechanical aberrationland. Name your distortion.
By the way, have you ever seen test reports on the interchannel phase
difference vs. frequency OR the absolute phase vs. frequency of an LP
cartridge? I can't recall seeing them. Before we criticize CDs for their
`excessive' HF phase delay, shouldn't we find out what we're used to
listening to?
Well, somewhere within this long ramble are some suggestions for better
component tests. It is obvious to me that our present set of LP and CD tests
a) do not adequately test linearity and dynamic range, the essential character-
istics of a linear transducer, b) do not test LPs and CDs in a way that
meaningfully compares each with the other, and c) do not thoroughly consider the
behavior of the receiving system, the listener.
Rather than engage in wild speculation on insufficent data, or engage in
status quo maintainance a la Linn Sondek or Douglas Sax, the industry should
come up with a series of standardized tests for linear transducers and a set
of `tty drivers', if you will, for LPs, CDs, tuners, amplifiers, tape decks,
etc..
Then we can rationally discuss the relative merits of each system and
identify the mechanisms behind each fault. Until then, we're just arguing
religion.
The rambling squash,
--
Shaun Simpkins
uucp: {ucbvax,decvax,chico,pur-ee,cbosg,ihnss}!teklabs!tekcad!vice!shauns
CSnet: shauns@tek
ARPAnet:shauns.tek@rand-relay