[net.audio] differences in sounds of CD players

pmr@drufl.UUCP (Rastocny) (01/10/84)

I am a bit puzzled by some of the comments I've read on the net.
Some (not all) of these puzzles are summarized and listed below.

> "...distortion specs on CD players are far better than the
> ear has the ability to detect..."

> "...people could not hear distortion that was artificially
> introduced until it reached a level of about 0.5%..."

> "...the ear cannot hear phase differences above 10KHz..."

> "...digital is emphatically superior to analog..."

If all of these statements are true, and since the specs between
digital players are too insignificant for the ear to detect
(0.05% THD, etc.), then why does a Sony CD player sound different
from a Luxman CD player (or a Hitachi, or a Technics, etc.)?

		Yours for higher fidelity,
		Phil Rastocny
		AT&T-ISL
		..!drufl!pmr

rfg@hound.UUCP (R.GRANTGES) (01/11/84)

Are you, Mr. Rastocny, saying they do sound different from your
own two-eared experience or is this just another 
rhetorical question designed to provoke the easily provoked.
The critical comment I have read to date preponderantly
indicates they all do sound alike. When answering you should
also indicate if you hear the difference between monster cable
and any old 14 gauge wire so we can calibrate on the sensitivity
of your hearing. Thanks, Dick Grantges hound!rfg

5121cdd@houxm.UUCP (C.DORY) (01/11/84)

Boys, Boys...are we forgetting all the anolog circuitry in the CD player?
There are about as many unique realizations of these analog circuits
(i.e., anti-aliasing filter, reconstruction filter, line-level circuitry, etc.)
as there are currently available CD players.  Not to mention, some
manufacturers utilize over-sampling techniques, both digital and analog
filtering, etc., etc...  If, of course, you believe the rave reviews in
Audio, Stereo Review, and High Fidelity, these varying executions of CD
players make no difference whatsoever -- remember, these mags live off the
advertising revenue from audio-equipment manufacturers.  The CD has received
the biggest media hype the audio industry has seen. (Has anyone seen a bad
review in one of these mags lately?)

There are other IMPORTANT aspects to audio equipment performance than
simply THD distortion specs and magnitude response (erroneously called
"frequency response") alone.  What about time-domain response?  We are
attempting (through the recording chain) to recreate musical performances
so that we can enjoy them time and time again.  Music is NOT time independent --a bassoon sounds like a bassoon when its frequency and time components
are reproduced accurately.  (These are not usubstantiated claims --
much of the recent work of Deane Jensen - Jensen Transformers, John Bau -
Spica Loudspeakers, Mile Nestorovic - Nestorovic Loudspeakers, John Meyer -
Meyer Sound Labs, Patrick Duran - Horus Music, John Curl, (the list goes
on...) is focused on TIME DOMAIN accuracy of their components.

It is very important, also, to note what type of music is listened to in
making subjective judgments in audio equipment.  No offense, but there
is little "true stereo" information in current Rock/Pop and most of
the Jazz recordings -- Supertramp - "Crime of the Century" and 
Pink Floyd - "Dark Side of the Moon" and "The Wall" are notable exceptions,
however.  Most of the Pop/Rock and jazz recordings are made with
the recording console as an integral "instrument" in the compositions
and, therefore, these recordings are multi-mono.  True stereo ONLY comes
from simple miking techniques (two and at most three mics with the
best results from coincident techniques).

What I'm getting at, is that unless that your source is of unimpeachable
quality (I know of few), your system is accurate (well?), you have an
excellent listening room (noisy fridge, flimsy walls?) there is no way
you can make many of the subjective judgments that I see made here on the
net and in the mags.



                                      Craig Dory -- AT&T Bell Laboratories
                                                         Holmdel, NJ

pmr@drufl.UUCP (Rastocny) (01/11/84)

Interresting, and I agree.  But several engineering types have claimed
that time-domain related errors are undetectable by the human ear
at higher frequencies (names withheld).  (CD systems are uniform at
lower freq's and get into trouble usually over 5KHz.)  Others say that
they all should sound the same since distortion (et. al.) specs are
insignificantly different from one CD unit to another.  Some,
understandably, think that minute differences in loudness sway a
person's opinion of superiority regardless of the reference system's
quality.  You all can't be right, can you?  Who is?

Before things get too enFLAMEd, I suppose I should explain why I started
this.  I feel that the current set of specs used by everyone in the top
three audio mag's is inadequate in describing sonic differences that
are currently described in subjective magazines.  Anything that we can
hear, we should be able to measure.  Can anyone offer opinions as to
what specs would more correctly describes these differences? (e.g., why
isn't there a plot of the power supply impedance vs. frequency at full
load?  Or supply ripple at the output devices under full load?  Or
a time-domain plot?)

		Yours for higher fidelity,
		Phil Rastocny
		AT&T-ISL
		..!drufl!pmr

gregr@tekig1.UUCP (Greg Rogers) (01/12/84)

Phil, you seem to want to provoke a debate on the sound of CD's based only
on unsupported assertions of their sound quality.  Is it correct to assume
that you have no knowledge of, or respect for scientific method?  I have yet
to find a single person who has been able to demonstrate any significant
audible difference between CD players under a scientifically controlled
double-blind test.  If anyone is aware of any such test that supports your
assertion will they please provide a reference to the test.  This will 
really be of benefit to those of us who spend a great deal of time
discovering the physical, measurable differences between audio components
that do contribute to audible differences. 

It must be stressed EMPHATICALLY that any perception of differences between
audio components that is NOT the result of controlled double-blind testing
is simply worthless information given the performance levels of todays 
electronic components.  (It is likely that the audibly gross differences
between speakers and cartridges makes such rigorous testing unnecessary
to merely prove differences exist but continued improvements in these
components may make or already have made even this a dangerous exception.)
Note carefully the word CONTROLLED above.  This implies that all external
variables that can influence the sound of the units under test are removed.
I believe that we are all aware that when comparing two nearly identically
components if the sound levels are not precisely matched (0.2 dB has been
documented) that the louder observation will usually sound better and at 
the least different.  Hence I repeat any claims of sound differences between
CD players NOT made under controlled double-blind testing is WORTHLESS.

But maybe the real question should be, "Given the amazing measured performance
levels of even the very first CD players, could they still sound different?"
The answer to that is very simple - "Of course they could".  It seems unlikely
(even contradictory to many controlled tests) that the current set of parametersbeing measured would be responsible.  But the possibility exists that we are
not measuring some previously unconsidered parameter that does have an effect
at an audible level.  But it is clearly up to those that claim to hear a 
difference to prove through proper testing that a difference exists before
we need to search for some heretofore unknown parameter.  Until that day comes
the evidence suggests that even the first of the CD players are capable of
reproducing the recorded CD's with no significant degradations or differences.

(Ah yes, that word "significant" again, there is no question that measurable
differences exist between players, but our previous knowledge maintains
that they are not audible.)

For anyone not yet bored I should give an example of discovering a new 
parameter that affected sound quality but proved tough to discover.  It was
long believed by many (seemed obvious to most critical listeners) that
moving magnet input stages of phono preamps sounded different between
components.  Some "objective" reviewers who based their opinions only on
measurements claimed that differences didn't exist.  This was due to the fact
that they made measurements by driving the input of the preamp from signal
generators which didn't accurately simulate the complex interaction between
the output impedance of the cartridges and the input loading of the preamp.  
Hence their measurement technique was faulty.  (Their real sin here was not
trusting their ears to make them question the measurements).  Controlled 
double-blind testing provided ample evidence that differences did in fact exist.Having proven the existence of differences it was relatively straightforward to
isolate the cause and design better measurement techniques and in fact to
design better circuits to eliminate the interaction problem.  The difference
in this example to today is that testing did show audible differences between
components.  Critics of CD's should not take too much encouragement from this
example because the reverse has also proven true.  Despite the best efforts
of a rather famous amplifier designer he was unable after several attempts
to prove his amplifier to sound audibly different than another under such
rigorous testing.  If anyone would like further information on testing
methodology or any example I noted consult the back issues of the 
Audio Engineering Society Journal.

				Still working on REAL improvements,
				Greg Rogers 
 
 
 

jj@rabbit.UUCP (01/12/84)

Well, Phil, you say, and I paraphrase:
	1) Some people say that time domain behavior doesn't
matter above 5kHz or so
	2) Some people say that CD distortion is so low
that there's no sonic difference
	3) Some say (understandibly) that minute differences in
playback levels will influence listeners more

Then you say,
	"Well, how can you all be right?"

It's a nice debating technique to say that, but it's also TRUE.  
They are ALL right about what they've said.  Why?
1<-->2  Phase distortion ::= Time domain distortion
and the people who are (informedly) saying 1 and 2 are saying,
in part, the same thing.  Furthermore, 1 is simply a subset
of 2.
The people who claim 3) are pointing out (VERY correctly) that
minute <.1dB> differences in playback levels will mask ALL of
1) and 2) when the spec's are at the CD level.   

I don't understand just WHAT you even intended to say with your
article, given the incorrect (but effective in debate) assumtions of the 
first two paragraphs. <Unless the ONLY point was making debate 
that is intended only to influence uninformed people.>


Another really offensive debating technique you use is to say
...don't get into trouble until above 5kHz...

What do you mean getting into trouble? Apparently a distortion
that's 40-50dB better than any vinyl ever made, and a phase response
that's SMOOTHLY changing, rather than raggedly jumping all over the
place like a cutting lathe and mastering deck, is getting into trouble,
while the gargantuan CHANGES in group delay (let's use the relevant
terms, here) that one gets from a tape deck (not to mention the
self-erasure and scrape/flutter noise) aren't trouble at all.
I'm confused, Phil,  what are you saying, perhaps that analog
recordings are even worse, and not worth anything at all?

Didn't think so.

Do me a favor, Phil, co-post your articles to net.flame.  This is
net.audio, I'm trying to be polite, and I'm trying to keep my sheer
annoyance down to a loud scream.  If you post an article to net.flame
<where it belongs> I'll reply in the tone that I feel is really
appropriate. (If I have time and energy, I  sometimes wonder if
it all matters in net.audio.)

ENOUGH!

======================
Everything you know is wrong!
Distortion is Fidelity
Weakness is Strength
Assertion is Truth

rabbit!jj
<welcome to 1984, Phil>

rentsch@unc.UUCP (Tim Rentsch) (01/14/84)

Just to relate a story:

Some years ago people were thinking about deciding on a standard for
FM broadcasts (in terms of frequency response).  Most people (engineering
and technical types) wanted to follow the AM standard, i.e., limit to
5 KHz.  Some people at Bell Labs did some experimentation and came out
with the "radical" limit of 8 KHz (beyond which no difference could be
heard).  Well, the radical fringe had its say, and as you all know FM
standardized on limiting to 15 KHz.

Nowadays, we know that 20 KHz is right.  (Or do we?)

"From transducer to transducer, sound reproduction that's better than
ears."  This is the goal.  When it is achieved, identical will mean
identical -- to all people, in all circumstances.

rmd@hpcnoa.UUCP (01/19/84)

#R:drufl:-77300:hpcnoa:30200015:000:579
hpcnoa!rmd    Jan 14 16:03:00 1984

CD  players  certainly  sound more  alike than any other  class of audio
playback equipment.  (i.e.  turntables and cassette decks)

I personally  think that it is stretching  things to say that CD players
sound significantly  different from each other.  However, there are some
things like frequency  response which are slightly different and I could
believe that these differences could be audible if you spend enough time
looking for them.

Let me ask a similar  question:  why does  'high end'  analog  equipment
sound different?  It can't all be right.

Rick Dow
hpfcla!rmd

gwes@inmet.UUCP (01/20/84)

#R:drufl:-77300:inmet:2600029:000:1716
inmet!gwes    Jan 16 14:00:00 1984

Re: distortion perceptions, time delays, and reviews

1) Up to 3% (or so) of CORRELATED 2nd harmonic distortion is often not perceived.
It is often first noticed as "warmth" in the sound.  HIGHER ORDER or
UNCORRELATED distortions (time domain) are much more easily perceived,
at the low levels of .05% or below.  In general, the higher order distortions
are much more unpleasant in smaller quantities.

2) The one area of time domain measurements I have done indicates extreme
sensitivity to time delays for ambience and placement information.  Movements
of speaker elements of approximately 1 inch (= 80 microseconds) can be
distinguished by comparisons.  Proper time correlation is recognized as an
"open window" effect, with instrument placement becoming definite.

Time delays in filters seem to generate the same type of effects.  I have not
been able to run controlled experiments to properly examine this, however,
the psychoacoustics folks often use Bessel filters for their experiments
because of the uniform delay characteristics.  They don't (often) care so
much about the amplitude ripple.

3) If you really want to get useful information from the advertising noise
in Audio, etc., read the reviews carefully.  Absence of praise and exact wording
is very important.  The classic example is the review that praises the
beautiful styling of the unit, and leaves performance to last.  The reviewer
is probably obligated to praise SOMETHING.  If he doesn't like the sound,
you will notice qualifiers or lack of definition in the review.  

The graphs and tables DO give you the objective data.  Read them.
The subjective data can be obtained by subtracting the hype offset.

Geoff Steckel (harpo!inmet!gwes)

pmr@drufl.UUCP (01/20/84)

Rick Dow's comments are well taken.  Why do high-end components all
sound different?  Could be a couple of things but I'm not picking
on anything in particular:
	1. The quality of capacitors used in the signal and feedback
	   paths.  Low ESR caps "leak" less and pass a more accurate
	   and correspondingly less distorted version of the original
	   signal.  Low end components use budget capacitors that
	   have significantly higher ESR.
	2. Quality of circuit design.  Low-end components are concerned
	   about things like parts count and individual component
	   prices to make a product sellable in a highly competitive
	   market.  Once you decide that price is no object, you can
	   design more sophisticated and correspondingly more sonically
	   accurate circuits, but the number of people who are able to
	   afford this equipment is very limited.
	3. Attention to detail.  Although it may be hotly debated as
	   to whether design considerations of wide bandwidth and
	   well-regulated power supplies is in fact audible, most
	   high-end stuff will use these considerations in their
	   designs.  Things like board layout, positions of input
	   and output stages with respect to the rear panel jacks,
	   and regulators for each gain stage are commonly incorporated
	   in the high end.  Low end stuff may not have a regulator
	   even on the front-end circuitry.

How does this translate into sonic reproduction?  Very well in most
cases of high-end gear.  But not all high-end equipment sounds good.
There seems to be an art involved as well as a science.  After all,
Fourier never claimed that his mathematical models in any way depicted
reality.  They were just that, models -- nothing more, nothing less --
but still good approximations.

I just don't know why high-end sounds better but I can point to some
circuit differences between low and high end that must be at least
part of the reason.

		Yours for higher fidelity,
		Phil Rastocny
		AT&T-ISL
		..!drufl!pmr