Audio Critic the 22 r schematic

Issue No. 22

Display until arrival of

Issue No. 23.

Is this the format of the future in power amplifiers?

(See the analog electronics reviews.)

Retail price: U.S. $7.50, Can. $8.50

In this issue:

A number of exceptionally competent (and in some
cases controversial) loudspeaker designs are
evaluated, with special emphasis on subwoofers.

In an editorial that will raise some eyebrows, and
even some blood pressures, your Editor analyzes
the hypocrisy of the high priests of the High End.

We continue our survey of amplifiers and preamps.
The question of when, if ever, absolute polarity is

audible is clarified in a letter from a top authority.
Plus other test reports, all our regular

columns, and the return of our popular
CD capsule reviews (oodles of them).

pdf 1

Contents

Paradoxes and Ironies of the Audio World:

The Doctor Zaius Syndrome

By Peter Aczel, Editor and Publisher

Loudspeakers Are Getting Better and Better

By Peter Aczel, Editor and Publisher

12 ACI "Spirit"
13 Bag End ELF Systems S10E-C and S18E-C (continued from Issue No. 21)
14 What the Bag End ELF System Does and Doesn't (sidebar by Dr. David Rich)

15 Velodyne DF-661 (continued from Issue No. 21)
21 Velodyne Servo F-1500R
23 Win SM-8
43 Snell Acoustics Type A (last-minute mini preview)

Good Things Are Still Happening in
Analog Electronics

By Peter Aczel, Editor and Publisher
& David A. Rich, Ph.D., Contributing Technical Editor

25 Line-Level Preamplifier: Aragon 18k (Reviewed by Peter Aczel)
27 Line-Level Preamplifier: Bryston BP20 (Reviewed by Peter Aczel)
27 Mono Power Amplifier: Marantz MA500 (Reviewed by Peter Aczel)
28 Stereo Power Amplifier: PSE Studio IV (Reviewed by David Rich)
29 Line-Level Preamplifier: Rotel RHA-10 (Reviewed by Peter Aczel)
30 Stereo Power Amplifier: Rotel RHB-10 (Reviewed by Peter Aczel)
30 Passive Control Unit: Rotel RHC-10 (Reviewed by Peter Aczel)
31 Stereo Power Amplifier: Sunfire (Preview by Peter Aczel and David Rich)

Catching Up on the Digital Scene

By Peter Aczel, Editor and Publisher
& David A. Rich, Ph.D., Contributing Technical Editor

33 Compact Disc Player: Denon DCD-2700 (Reviewed by David Rich)
35 Outboard D/A Converter with Transport: Deltec Precision Audio PDM 2 and T1

(Reviewed by Peter Aczel)

36 Compact Disc Player: Enlightened Audio Designs CD-1000

(Reviewed by Peter Aczel)

41 Compact Disc Player: Marantz CD-63 and CD-63SE (Reviewed by David Rich)

Large-Screen TV for Home Theater:
Is a 40" Direct-View Tube Big Enough?

By Peter Aczel, Editor and Publisher

45 40" Direct-View Color TV: Mitsubishi CS-40601

Obstructionism By Tom Nousaine

Hip Boots Wading through the Mire of Misinformation in the Audio Press

Four commentaries by the Editor

Recorded Music Editor's Grab Bag of CDs, New or Fairly Recent

ISSUE NO. 22 • WINTER 1994-95 1

10

12

25

33

45

47
50

52

pdf 2

From the Editor/Publisher:

This issue is dated Winter 1994-95. The
last issue, No. 21, was dated Spring 1994.
That was late spring; this issue goes to

press in early winter, so the gap is smaller

than the apparent nine months but bad
enough. What are we doing about it? Lots
of things—but we have learned, painfully,
not to make promises before the implemen­tation is a reality. Three things are certain:
(1) something has to give; (2) we are here
to stay, regardless; (3) we are not chang­ing our editorial stance. That still leaves a
number of viable scenarios to choose from.

Issue No. 22

Winter 1994-95

Editor and Publisher
Contributing Technical Editor
Contributing Editor at Large
Technical Consultant
Columnist
Cartoonist and Illustrator
Business Manager

Peter Aczel

David Rich

David Ranada

Steven Norsworthy

Tom Nousaine

Tom Aczel

Bodil Aczel

The Audio Critic® (ISSN 0146-4701) is published quarterly for $24

per year by Critic Publications, Inc., 1380 Masi Road, Quakertown, PA

18951-5221. Second-class postage paid at Quakertown, PA. Postmaster:
Send address changes to The Audio Critic, P.O. Box 978, Quakertown,
PA 18951-0978.

The Audio Critic is an advisory service and technical review for
consumers of sophisticated audio equipment. Any conclusion, rating,
recommendation, criticism, or caveat published by The Audio Critic
represents the personal findings and judgments of the Editor and the
Staff, based only on the equipment available to their scrutiny and on
their knowledge of the subject, and is therefore not offered to the reader
as an infallible truth nor as an irreversible opinion applying to all extant
and forthcoming samples of a particular product. Address all editorial
correspondence to The Editor, The Audio Critic, P.O. Box 978,
Quakertown, PA 18951-0978.

Contents of this issue copyright © 1994 by Critic Publications, Inc. All rights
reserved under international and Pan-American copyright conventions. Repro­duction in whole or in part is prohibited without the prior written permission of
the Publisher. Paraphrasing of product reviews for advertising or commercial
purposes is also prohibited without prior written permission. The Audio Critic
will use all available means to prevent or prosecute any such unauthorized use
of its material or its name.

Subscription Information and Rates

First of all, you don't absolutely need one of our printed subscrip­tion blanks. If you wish, simply write your name and address as legibly
as possible on any piece of paper. Preferably print or type. Enclose with
payment. That's all. Or, if you prefer, use VISA or MasterCard, either
by mail or by telephone.

Secondly, we have only two subscription rates. If you live in the
U.S., Canada, or Mexico, you pay $24 for four consecutive issues
(scheduled to be mailed at approximately quarterly intervals). If you
live in any other country, you pay $38 for a four-issue subscription by
airmail. All payments from abroad, including Canada, must be in U.S.
funds, collectable in the U.S. without a service charge.

You may start your subscription with any issue, although we feel
that new subscribers should have a few back issues to gain a better un­derstanding of what The Audio Critic is all about. We still have Issues
No. 11, 13, 14, and 16 through 21 in stock. Issues earlier than No. 11
are now out of print, as are No. 12 and No. 15. Please specify which
issues you want (at $24 per four).

One more thing. We don't sell single issues by mail. You'll find
those at somewhat higher cost at selected newsdealers, bookstores, and
audio stores.

Address all subscriptions to The Audio Critic, P.O. Box 978,
Quakertown, PA 18951-0978. VISA/MasterCard: [phone].

2 THE AUDIO CRITIC

pdf 3

Box 978

Letters to the Editor

An archetypal letter we seem to get again and again lists in loving detail all the components in the
writer's system, down to interconnects and tiptoes. In nearly every case it's quite unclear what the

letter writer wants. Our official blessing? Recommended changes? Recognition as a blood brother?
Please, all you audio addicts, if you insist on talking about your equipment and want to get our
attention, make sure you explain how it all ties in with our editorial concerns. Letters printed here
may or may not be excerpted at the discretion of the Editor. Ellipsis (...) indicates omission. Address
all editorial correspondence to the Editor, The Audio Critic, P.O. Box 978, Quakertown, PA 18951.

The Audio Critic:

...I have read Issue No. 20 and
found it quite charming. Your elegant
chopping to pieces of the audio tweakies
was very nice indeed. I even enjoyed the
writings of the others, many of whom I

know quite well....

* * *

[Six weeks later:]

I have now read the several issues of
The Audio Critic which I recently re­ceived. They are very interesting, and I
want to congratulate you on getting some
very good people to write for you. I am a
bit surprised that you and some letter
writers refer to my modest work so fre­quently but appreciate the interest I have
generated. There are many issues that
need deep thought, and I am delighted to
have these discussions take place, since
the light envoked (with some heat) even­tually seems to pry out the truth.

It is very annoying to have persons
at the extreme fringes of an issue use
one's writings to prove their point with
elements of these writings taken out of
context. Unfortunately, it happens only
too often. Shades of gray are too often
made black or white by fanatics.

I was pleased to see complete para-

graphs from my paper quoted in context
printed in Issue No. 21. Additionally,
your interpretation, printed on page 8, of
what my article said is quite accurate but
a bit more truncated than I would have
preferred. If you have the patience, I sup­ply for you herewith my own summary of
my work on acoustic polarity.

For perspective, my paper on polari-

ty was presented at an AES convention in

1991 and was published in the Journal of
the Audio Engineering Society, Vol. 42,
No. 4, 1994 April. Needless to say, this is
a professional Journal for which all arti­cles are extensively reviewed for quality
and accuracy. A version of this paper was
also printed in Audio magazine with mi­nor modifications to satisfy a different
(and much larger) audience. Neverthe­less, the essential contents of the several
versions of the paper are identical.

I strongly suggest that anyone inter­ested in this matter read the original
paper and not someone else's interpreta­tion of what it might say. Some of the lis­tening tests are fairy easy to duplicate,
and the paper is written in very simple,
not highly technical, terms.

These comments cover both the
work I did in 1991 and my more recent

experiences (1993-1994), which are indi­cated by brackets [...].

There are four points made:

1. It is clearly possible to show the
audibility of acoustic polarity inversion
with steady-state tones (electronically
generated) or quasi-steady-state tones
(produced on physical instruments but
with steady monotone playing tech­niques). These are boring, nonmusical
tones. The audibility of acoustic polarity
inversion for these very special cases has
been documented by several authors. [I
have repeated this listening experience
many times with both headphones and
various loudspeakers, and the experience
is so definitive that there is no question
about its existence.]

2. When real musical performance
material is used in such tests, it is very,
very difficult (nearly impossible) to hear
the effects of polarity inversion. Our
large group tests showed only very slight
positive results with loudspeakers in a
highly idealized and simplified listening
environment. [These listening tests have
been repeated with headphones and a
great variety of loudspeakers, and it has
been confirmed that it is very, very
difficult to hear polarity inversion. Nei-

ISSUE NO. 22 • WINTER 1994-95 3

pdf 4

ther I nor anyone I know, and trust, has
heard acoustic polarity inversion with ste­reo program material in a normal listen­ing environment.]

3. Because it was so easy to hear
polarity inversion with simple steady­state tones and so difficult to hear with
real music, a large part of the paper is de­voted to trying to determine the reasons
why this is the case. A major part of the
paper suggests, but does not firmly
define, these reasons. More work is re­quired to define this very subtile psycho­acoustic effect. [While I have continued
some work in this area, I have not found
a consistent, definitive cause/effect rela­tionship. However, it is clear to me that
the audibility of acoustic polarity inver­sion is dependent on both the acuity of
the listener and the nature of the program
material, and not highly dependent on the

transducers involved.]

4. The issue of the audibility of
acoustic polarity inversion is not a matter
of black and white but of a series of
shades of gray, seemingly dependent
upon the simplicity or complexity of the
program material being auditioned, to
some small extent upon similar factors of
complexity of the listening environment,
and to some extent upon the sensitivity of
the listener. [While I would like to see
standardization of polarity in recording
and reproduction, it seems to be a minor
issue compared to others that affect
sound reproduction much more strongly.]

Several final issues need to be laid
to rest. One is the question of the use of
digital recordings and digital program
material to do listening experiments. The
essential results described above have
been duplicated with real instruments,
microphones, and headphones in real
time without the use of any recording de­vices. Some fanatics suggest that only
they can hear things because of their
equipment. This is total nonsense.

Some have suggested that the quali­ty of the loudspeaker and/or measure­ment techniques used for the original
paper were somehow defective. (This has
been implied by some snide remarks by
C. Johnsen in The Audio Critic letters
column as well as in Audio magazine.)
Our experiments were set up with ex­treme care, using a large array of the
very best professional-level instrumenta­tion equipment in my Electroacoustics
Laboratory. We are totally comfortable
that we know how to use this equipment,
have designed a suitable loudspeaker, and

4

have analyzed the results in an entirely
professional manner.

Finally, all of our findings have been

confirmed with headphones of various

sorts and a number of quite diversely de-

signed loudspeakers. The design of the
headphones or loudspeakers, within rea­son, is in my experience irrelevant to re­vealing the phenomenon.

I have over the years discussed these
issues with many AES members, includ­ing Richard Heyser and Stan Lipshitz. I
believe that we all would prefer that the
industry took care with polarity conven­tions. But, they have not for the most
part. Polarity is simply not a high-priority
issue for most professionals and certainly
not highly important for the enjoyment of
reproduced sound.

Nevertheless, I am actively carrying
out additional experiments and I hope to
pursue the matter with the goal of
finding sources/causes of audibility of
acoustic polarity inversion and to specify
it more clearly in a scientific, responsible
manner.

That's it for the time being. I suppose
that heated debate will continue, just as it
does with the cable/interconnect issue.

Very sincerely,
R. A. Greiner
Professor
Fellow of the AES

Thank you for the kind words about
The Audio Critic. We not only get "some

very good people" to write articles for us
but also, as your example proves, some
very good people to write letters to the
Editor. Indeed, your letter dots the i's
and crosses the t's on the subject of pola­rity for all rational audiophiles. Let the
tweaks read it and weep.

—Ed.

The Audio Critic:

I love you curmudgeons. You're
even occasionally correct—but then, so
are John [Atkinson] and Harry [Pearson].

I have no doubts as to your superior
theoretical and technical qualifications; I
also have no doubt that much subjective
reviewing necessarily utilizes poor meth­odology; however, why do you have
problems accepting the idea that some au-

dible differences are either difficult to
measure using conventional parameters,
or are the result of phenomena not yet
fully understood?

Howard Cowan
Woodland Hills, CA

/ have no problem whatsoever with

the ideas you state as long as those "au­dible differences" are indeed audible.
What I have a problem with is the state­ment that "I can hear the difference"
when you are unable to prove to me un-

der controlled conditions that you can ac-

tually hear it. If there really is a provably
audible difference, the cause may or may
not be easy to determine—that's a totally
separate issue.

As for John and Harry, see "The

Doctor Zaius Syndrome " (page 10).

—Ed.

The Audio Critic:

I have some ideas about how I think
a modern music-reproducing system
ought to be be configured to minimize
hardware interactions with the music in­formation. I think from what I've read in
this magazine, and from some compo­nents already placed on the market (Me­ridian comes to mind), that some very tal­ented people are thinking along these
same lines, and I'm wondering why we
(meaning the industry and the hobbyists)
are not headed a bit more quickly in this
direction.

Before I explain, let me admit (as
you're wont to question this) that I have
no credentials other than a 30-some-year
interest in the hobby. I'm also a fairly re­cent convert from the music-to-justify­hardware group.

My concept of a system would have
it divided into two basic modules. One
I'll call the control module, the other the
speaker-system module—or actually mod­ules, as there would be several of these.
The control module would look very
much like a home computer system or
possibly a TV set, and might actually be
integrated with one of these—or both.

The purely electronic functions, such as
preamp, tuner, processor, etc., would be
installed as industry-standard plug-in
boards, with all of their switching and
control functions accessed as icons on the
screen—very similar to Macintosh, or

IBM with Windows, menus. Access to

the sound system might actually be a
menu option on your computer terminal.
The actual hands-on control would be a
mouse or infrared remote.

The control module would provide
ports with a standard multipin socket, for
inputs from purely mechanical program
sources such as an LP turntable [buggy
whip on the space shuttle?—Ed], CD
transport, cassette transport, etc. The

THE AUDIO CRITIC

pdf 5

module would operate entirely in the dig­ital domain, which would eliminate the
need for expensive hardware to maintain
a clean signal, and the output, or outputs,
to the speaker systems would be via
fiber-optic cables. System upgrades
would be accomplished by replacing or
adding boards.

The speaker systems would each
contain a speaker, or speakers, a band­width-limited amp with equalization (if

necessary), an electronic crossover, and a

DAC circuit for the fiber-optic inputs.

These systems would be tailored for their
particular function, such as subwoofer

(similar to the Velodynes), main left and

right, and surround-type speakers. I visu­alize the surround speakers as designed
much like track lighting on the ceiling,
spherical enclosures that would be aim­able and could contain their electronics in
small boxes which could flush-mount in
the wall or ceiling and be attached by
short cables to the spheres. A full system
would consist of two subwoofers limited
to 80 Hz, two mains for left and right (80
Hz and up), and four of the ceiling­mounted speakers, center front and back,
and back left and right.

Now tell me, where am I wrong in
this concept? And, if it's basically accu­rate, why aren't we already there? Disre­garding the tubes-and-LP crowd, I think
we're hung up in oldthink. I, for one, find
the stack of black chassis and their tangle
of wires an eyesore and probably unnec­essary.

Hartley Anderson
Waco, TX

To quote that song from the big­band era, "I'll Buy That Dream." As you
point out, bits and pieces of the dream ex­ist already: powered subwoofers are the
rule rather than the exception; powered

full-range speakers are still the exception

but no longer a great rarity; Marantz
showed a computer front end as early as

1991; I could go on. It hasn't all come to-

gether, though; the demand isn't there;

"separates" are still the audiophile

norm.

Your basic concept is very much in
line with my own thinking, but I'll go
even further: A/D conversion should take
place right out of the microphone preamp
and the signal kept in the digital domain
throughout the recording, editing, mas­tering, duplicating, broadcasting, domes­tic playback, etc., processes, right up to
the D/A conversion just before the ampli-

ISSUE NO. 22 • WINTER 1994-95

fication stage of each separately powered

speaker channel—and that includes digi-

tal filters for all crossovers. (Maybe you,
too, had that in mind but didn 't quite say

it.) The speaker deployment should prob­ably follow the Lexicon model in its full­est form: front left, center, and rear, sub­woofer(s), side left and right, rear left
and right. But these are details. You've
got the main idea right—and you will see

it happen. The question is, when? Some
observers feel that the audio consumer
will continue to resist the idea of the sep­aration of amplifiers and speakers.

—Ed.

The Audio Critic:

...My question specifically has to do
with whether a pair of subwoofers is bet­ter than one subwoofer. This question is
generated by a recent article by John F.
Sehring, which appeared in Audio maga­zine (February 1994). In that article, Mr.
Sehring gives a number of reasons why
stereo subwoofing is superior. I wonder
whether The Audio Critic has an opinion
in this regard. I also realize that the an­swer might depend partly upon a number
of variables (placement, room size, room
acoustics, etc.), and therefore no univer­sal answer or rules of thumb may obtain.
However, any opinion at all would be
helpful.

I also have another question regard­ing the use of built-in amplifiers in sub­woofers. Some recent literature which I

received from VMPS suggested that
built-in amplifiers are a bad idea because

subwoofer vibrations will eventually sim­ply rattle them apart, as it were. Does this
turn out to be the case? Do the electronics

in powered subwoofers self-destruct after

a relatively short life span?

Please keep up the good work; your
magazine is a delight.

Sincerely,
David R. Reich
Auburn, NY

/ have always been of the opinion

that a pair of stereo subwoofers is prefer-

able to a single mono (L + R matrixed)

subwoofer—see Issue No. 16, page 16—
but Tom Nousaine, who has studied the
subject in considerable depth, vigorously
disagrees. His findings are documented
in a forthcoming article in the January
1995 issue of Stereo Review. This looks
like one of the few legitimate controver-

sies in audio (unlike the nonsense about

blind tests, tubes, etc.), and I am quite

open to all arguments. But, as in other
debates about all but the most obvious
audio phenomena, a number of reliable
practitioners have to be able to repeat

the same tests and obtain the same re­sults. Maybe Tom Nousaine needs to
broaden his statistical base before com-

ing to a sweeping conclusion; maybe not.
(For one thing, he is not into classical
music; as I once told him, it's a case of

"Pop Goes the Weasel.")

The VMPS caveat sounds like sour
grapes to me, since they make and sell
only passive subwoofers. Why don't built­in crossover networks, whose large com­ponents and large boards are much more
prone to vibration than amplifier parts,

fall apart untimely? Why don't radios in
jeeps fall apart? Needless to say, a cer-

tain amount of care and competence in
construction and placement must be as­sumed in all such instances. Audio hypo­chondria is, of course, a proven market­ing platform.

—Ed.

The Audio Critic:

...I...noted that David Rich made
reference to the Marantz CD-63 in his
footnote on page 16 of Issue No. 21.
However, I was disappointed to see that
he (Editor?) refers to the unit as "Philips­designed." Ordinarily, such a reference
would be taken as a compliment. Indeed,
in the Journal of the Audio Engineering
Society, I noted some years ago in their
regular review of audio patents that the
reviewer referred to Philips in the follow­ing way: "In many ways, Philips is the
audio equivalent of Mercedes-Benz. If
there is an esoteric way of doing things,
this is how Philips will do them." As I
seem to have misplaced the particular is­sue, I cannot say that the above is an ex­act quote, but it surely is very, very close
to the original comment.

In the case of the Marantz CD-63,
the model is in fact wholly designed
within the Marantz organization. Specif­ically, our chief CD designer, Mr. Yo­shiyuki Tanaka, is the gentleman respon­sible for the CD-63 and most of our other
CD players. He has a strong technical
background in digital audio as well as
analog circuit design, including power
supplies, and is very familiar with a wide
variety of available devices, CD mecha
nisms, and the like. I have in the past for­warded to him articles of interest, most of
which have come from the pages of your
magazine.

5

pdf 6

I am always pleased to see mention
of our gear in your magazine, and at Ma­rantz we are always quite proud of the
Marantz association within the giant Phi­lips concern, but I would ask that in the
future products submitted for review such

as the CD-63 be referred to as Marantz­designed, as this is the reality. Philips per
se had nothing to do with the CD-63 de­sign; however, we happily acknowledge
the fact that the CD-63 employs a number
of Philips-originated components, such as
the CDM12 mechanism, TDA1301 servo,
SAA7345 decoder, etc., that are also
found in Philips-branded models, as well
as models from other firms, such as Au-

dio Research....

As always, I appreciate your interest

in Marantz gear.

Best regards,
David Birch-Jones
Marketing Manager
Marantz America, Inc.
Roselle, IL

You'11 find that the review of the Ma-

rantz CD-63/63SE in this issue, essential­ly an updated leftover from Issue No. 21,
has been annotated to reflect your input.

That purely Marantz, non-Philips engi­neering seems to come up with products
that offer very solid performance per dol­lar. As for the articles you send to Mr.
Tanaka, I can understand why they are

from this publication, not from the digital

cloud-cuckoo-land of...well, you know.

—Ed.

The Audio Critic:

Having just completed reading the
review of the Magneplanar MG-1.5/QR, I
believe some important issues are raised.
Audiophiles know well that assessing
loudspeaker performance is difficult be­cause it is to a large extent subjective.
Every loudspeaker design is a compro­mise, and no particular transducer tech­nology has exclusive rights to accuracy.
Given your statements at the outset of the
review, some would question your "de­tached objectivity" in this case. It is clear
that you hold planar magnetic loudspeak­er technology in low regard and this ap­pears to have predestined your conclu­sions.

One of the generic criticisms of
Magneplanar designs made in the review
is the lack of low bass capacity. Your
nearfield measurements of the MG-1.5/
QR indicate its response extends to ap­proximately 40 Hz. The farfield response

6

in my listening room is significantly
flatter than you suggest (±4 dB), and
there is usable bass down to approximate­ly 30 Hz. Static distortion in the bass is
low (did you even bother to measure it?).
The conclusion that the MG-1.5/QR has
"not enough bass" doesn't follow, partic­ularly when compared to other loud­speakers available at the same price.

The other major criticism presented
involves driver ringing. This has been
noted by others in the past and appears
to be related to the physical nature of the
driver. It has been described in a variety
of uniformly driven nonrigid drivers,
which include ribbon and electrostatic
systems in addition to Magneplanar driv­ers. The audibility of the effect has not
been established, to my knowledge, as it
relates to a farfield listening position.
Transducers with significant energy stor­age typically emphasize and smear con­sonants in the spoken voice and have
poor square wave response, neither of
which is true of this speaker.

As for the "not enough focus, too
much coloration" that is reported in the
review, I can only comment that in my

experience the MG-1.5/QR is an extreme-

ly revealing transducer that allows a
wealth of musical detail to emerge. A

common characteristic of neutral trans­ducers is that each recording played

sounds different. This is exactly how

transparent the MG-1.5/QR sounds. Per­haps the Editor heard what he thought he
measured rather than the reverse.

None of the many conventional

loudspeakers that I am aware of in its
price range produces as wide, deep or
stable an image as the MG-1.5/QR. These
aspects of performance are not even dis­cussed in the review. The LEDR tracks
from the first Chesky test disc provide an
effective and quick method of evaluating
the spacial characteristics of loudspeakers
in conjunction with the surrounding envi­ronment. Few loudspeakers can generate
substantial "height" with the vertical sig­nal on the disc or provide an apparent
soundstage wider than the stereo pair.

With the MG-1.5/QR, the signal extends

to the ceiling in the vertical test and be­yond the width of the stereo pair in the
horizontal test.

The rather cursory nature of your
loudpeaker reviews has concerned me
since I first subcribed to your journal.

Perhaps it is time for the Editor to apply

the same rigorous standards to loudspeak-

er reviews as Dr. Rich does with the vari-

ous electronic components he discusses.
May I suggest that in the future more of
the test results and discussion of the lis­tening conditions (especially speaker
placement) be provided. It would also be
useful to have the manufacturers respond
to issues raised in the reviews. These
changes would allow the reader to better
assess the relative merits of each design
and reach his own conclusions.

Yours truly,
Dr. Douglas M. Hughes
Rochester, MN

I'm not surprised that the Minnesota
audio mafia finds staunch supporters at
the Mayo Clinic (or am I misinterpreting

your prefix and your address?), but you

happen to be mistaken on most of the

points you bring up. Not all of them,

though.

You 're right when you say that loud-

speaker evaluation is highly subjective,
although I try to back up my subjective

opinions with objectively verifiable evi-

dence. Furthermore, my subjectivity has
been refined over the years through expo­sure to literally hundreds of speakers—/
have some very good reference points of
subjective comparison.

You're also right in observing that

my loudspeaker reviews are not quite as
rigorous as David Rich's reviews of elec­tronic components, but there are good
reasons for that. A typical electronic sig­nal path in audio (such as, say, the left
channel of a power amplifier) has one in-

put and one output. It's relatively simple

and straightforward to examine the I/O
relationship. A speaker, on the other
hand, has one input and n outputs. Which

of the latter do we examine? Where in
space is the valid, or "official," output of
a loudspeaker system? How many points

in space are sufficient to give us an accu-

rate picture of the total output? These are
nagging questions of measurement meth­odology, and then there are the endless
other questions such as the physical
difficulty of ABX comparisons of speakers

(see Issue No. 20, page 39) and the vari­ous biases introduced by the reviewer's
accustomed listening room, etc. Rigorous
standards? We're working on them. Don
Keele is perhaps the most rigorous loud­speaker reviewer of us all, but then I do a

few tests that he doesn't. Amplifier-like

certainty in speaker testing we don't
have—and will not have soon.

Now then, here are the points where

you are wrong. (1) I do not "hold planar

THE AUDIO CRITIC

pdf 7

magnetic loudspeaker technology in low
regard," au contraire, my MG-1.5/QR re-
view begins with "Yes, I have a soft
spot," etc., and later I enthuse over the

upper bass and lower midrange of the
Tympani IVa. (2) The lower bass of the

MG-1.5/QR is defined by the fundamental

resonance of 44 Hz; it may be that your

particular room happens to offer some

reinforcement down to 30 Hz, but the de-

signer can't count on that in my particu-

lar room. (3) My frequency response
measurement was anechoic (via MLS)
and therefore closer to reality than any
in-room measurement. I don't know how
you arrived at your ±4 dB in-room figure,
but it's irrelevant; no speaker designer
deliberately makes the anechoic response
nice and jagged in the hope that the room
will homogenize it. (4) The (true) ribbon
tweeter of the Tympani IVa doesn't ring,
so your generalization is incorrect. (5)
The audible effects of ringing depend on
the frequency but they are very real; hav­ing been deeply involved in speaker de­sign as well as testing, I can only say I
wish you were right. (6) The MG-1.5/QR
does have poor square-wave response;

furthermore, ringing doesn't necessarily

affect the square-wave response unless

located near the square-wave fundamen­tal or its odd harmonics. (7) Revealing­ness is a relative quality—revealing com­pared to what other speaker? (8) I listen
before I measure. (9) I did comment on
the excellent height and width of the
soundstage. (10) Just as a single exam-

ple, the ACI (Audio Concepts, Inc.) G3

speaker, reviewed in Issue No. 19, beats
the MG-1.5/QR on bass and just about
everything else, at less than two thirds
the price.

As for manufacturers' comments, we
publish every word of them unedited—if
they write us. I don't believe in letting
them preview the reviews, however; it
makes for fruitless preemptive hassles.

Final thought: your Magneplanar
MG-1.5/QR is every bit as good, or bad,
after my review as it was before it. You
are quite certain that it's a great speaker,
so why is it important to you that it
should be blessed by The Audio Critic ?

—Ed.

The Audio Critic:

Most, if not all, of the articles ap­pearing in The Audio Critic are informa­tive, so I would like to suggest that some
knowledgeable individual write a short

article on what qualities make something

ISSUE NO. 22 • WINTER 1994-95

sound live. When one walks into a dining
room or bar and hears a piano, it is easy
to tell with the first chord if the music is
live or is being reproduced, and some­times even the make of the piano. It is
one of the reasons I have stopped going
to most concerts, as they seem to feel
they must use speakers, and therefore I'm
forced to listen to a loudspeaker (not nec­essarily a good speaker) and not the in­strument itself. If I go to a concert I want
to hear the actual horn or piano or what­ever. If I want to hear speakers, I can stay
home, as I've better speakers than they
use and far better sound.

I happen to be a collector of jazz
from the 1920s to the 1950s and still buy
78 rpm records when I find something in
very good condition that I want. I collect
these old records because many of these
recordings have not found their way onto
LPs or CDs. The bonus is that the music
on these 78 rpm records has more of a
live sound than LPs and much more than
CDs. I think Doug Sax touched on this
once. I truly believe, and hope, that I'm
not being influenced by the scratch and
noise common to these old records.

My cassette deck, a Nakamichi
680ZX, has a plasma display in place of a
VU meter or LEDs as level indicators.
When I copy a CD or even a 33-rpm LP
record onto a cassette, it is easy to see the
lighted portion of the display move from
left to right on drum rim shots or struck

piano notes (anything percussive). By
that, I mean one can easily see the lighted

portion travel from say -40 to 0 dB. Al­though fast, it's easy to see it move up

the scale with an increased level. When I
copy a 78-rpm record onto a cassette, it's
different. The music going from -40 dB

to 0 dB will result in the entire display,

up to 0 dB, instantly being lit. It is so fast
that one cannot see any movement of the

lighted display; it just appears. This in­stant change in level also happens if I

record the grand piano in our great room.
All of this is not as noticeable on LED
displays and completely obscured on VU

meters, as both are slow in comparison

to respond. Maybe that is why there are

so few plasma displays.

This leads me to believe that the at­tack time of a sound seems to have a di­rect bearing on how live it appears, and a
live sound is what we are all striving for.
Even when played through a 5 or 8 kHz
lowpass filter, 78-rpm records have a live
sound. To focus my question, why
doesn't a CD sound as live as a 78-rpm

record? By eliminating the compressors
and limiters and minimizing the active

stages and feedback, surely we should be

able to archive CDs that sound as live as
the old 78-rpm records of sixty years ago.
The added benefit of having no noise
would be wonderful.

I have every issue of The Audio

Critic, having read all of them at least
twice. Keep up the good work, as there
is so much #*%!# being shoveled out

there....

Yours truly,
Thomas F. Burroughs
Prescott, AZ

Aren't you the Tom Burroughs who
was selling Klipschorns in New York City
circa 1951? (I was very, very young then,
of course, and so were you.) If so, here's
some advice from one geezer to another:

Finagle yourself an invitation to a
live recording session (of a reasonably
competent label, I should add). Listen to
the direct sound of the musicians. That's
live, right? Now step into the monitor
room. (I'm assuming something a little
better than a telephone booth, and decent
monitor speakers.) What you now hear is
the CD sound (via the same l's and 0's
as will appear on the CD). How "live " is
it? Do you now feel that 78-rpm shellac
would sound more nearly like the live
musicians? I seriously doubt that you
would feel that after such an exercise. I
think you have lapsed into some kind of
technostalgia (to coin a word).

Of course, a superb 78-rpm record­ing from 1947, pressed on vinyl (which
they started to use around then) may
sound more "live" than an indifferent LP

from 1951, which in turn may sound bet-

ter than a botched CD from 1984 (they
had a few of those). But the best 78s ver­sus the best LPs versus the best CDs? Get
out!

I have no idea what's with your lev­el indicator—it could be any number of
things, including overload—but it isn't
attack time you're measuring. The lead­ing edge of a dynamic peak is determined
by the high frequencies, and the 78-rpm
shellac medium was certainly not superi­or in bandwidth to LP and CD. As for
compression, yes, it can upset the apple­cart, but good CDs aren't compressed.

And, by the way, feedback (correctly
applied feedback) is not the bad guy.
That's a 1970s notion, meanwhile laid to
rest by some of the best minds in the engi­neering world. Gotta keep up with the

7

pdf 8

times, old-timer. (Oops, what if you're

not that Tom Burroughs? What a burn...)

—Ed.

The Audio Critic:

I am a recent subscriber and I am
unbelievably happy to have found The
Audio Critic. I assembled a good system
for the first time in the last year, so I as­sume that I am the kind of person that
other magazines whine about needing to
recruit to save the "High End." Please al­low me to give you my perspective as an
inquisitive novice.

After reading enough analog drivel
every month, I actually started to wonder
if I had made a mistake by selling my LP
collection 4 years ago, so I went down to
my favorite hi-fi dealer and compared a
CD player with a much higher-priced
turntable/cartridge combo. The CD front
end was so vastly superior in detail, dy­namics, noise, and overall quality that I
remembered instantly why I smiled the
first time I heard a good CD system. I
also thought it was pretty darn musical
too, whatever that may be.

I am constantly bombarded with ar-

ticles recommending unbelievably expen-

sive wires, interconnects and, most re-

cently, magic wooden disks. I can't hear
a difference in controlled blind tests and
neither can the people that recommend
them, but if you don't agree with them
you are some kind of uncultured ignora­mus. Isn't it convenient that the differ­ences are supposed to be unmeasurable
things such as dynamic bloom and liquid-

ity.

I guess some people like distortion
in their music and that's why they love
those outrageously expensive vacuum­tube amps. I think that the real reason is
that a lot of people went over to Grand­pa's house as a kid, and his stereo glowed
in the dark. Now that they have six-figure
incomes they'll be damned if theirs isn't
going to glow too! It's a good thing we
have those East Bloc 1930s economies to
supply us with 1930s-technology vacuum
tubes.

When I entered the High End I had
no inkling that it was a fantasy world in­habited by mystics, romantics, and char­latans. The High End is hurting and at­tracts almost no women (another frequent
lament) because it is dominated by re­viewers, retailers, and manufacturers that
are either greedy or foolish and have lost
touch with reality. I don't think a lot of
them realize how bizarre it looks from
the outside.

8

The Audio Critic is like a breath of

fresh air. I'm glad that logic, reason, and
facts have a proponent in the audio
world. Keep up the good work!

Darren Leite
Scottsdale, AZ

Everything you say is right on the
money, but you don't ask the sixty-four­dollar question:

Is the truth bad for business?

My answer is that, in the high-end

audio world, the truth is probably bad for
business this week and next month but
very good for business over the next ten
years. B.S. has a limited shelf life; given
sufficient time, most consumers tend to
switch to the truth. The tweako/weirdo
high-end promoters, however, don't think
that far ahead.

Thank you for your kind words.

—Ed.

The Audio Critic:

I would like to know if you have an

explanation for the fact that, in my sys­tem, the Parasound HCA-2200II sounds
noticeably better through its balanced in­puts. I have a Parasound P/LD-1500 driv­ing it, with 24' lengths of Straight Wire
Flexconnect (unbalanced) or Canare Star­Quad (blanced). The levels are matched
to within about 0.2 dB, and out of 10
trials, single-blind (I was unaware of the

choices, but the switcher was), the differ­ence was immediately apparent each
time.

Yours truly,
Rob Bertrando
Reno, NV

As David Rich's review in Issue No.
21 clearly explained, the simplistic input
buffer circuit makes the balanced-input
distortion of the Parasound HCA-2200II
more than an order of magnitude worse
than through the unbalanced input. The
distortion is probably still below the
threshold of audibility. Parasound has
acknowledged that there was some kind
of minor foul-up in the production ver­sion of the balanced input circuit, and we
were supposed to get a letter from John
Curl that would clarify the matter. We
are still waiting.

"Sounds noticeably better" is of
course a purely subjective opinion and
unprovable. "The difference was immedi­ately apparent" is, on the other hand, a

provable statement and probably true in
your case.

Here are the possibilities that I see:

(1) The balanced-input distortion was

just above the threshold of audibility in

your unit, and you liked it. (2) The level
matching wasn't good enough ("about"

0.2 dB could have been 0.3 dB, which is
often perceptible), and you liked the loud­er or the softer choice. (3) The tweako
Straight Wire interconnect, which I'm not

familiar with, may introduce a rolloff or

other marginally audible inaccuracy, and

you liked it.

One thing is certain: the I/O rela­tionship is more linear when the unbal­anced input of the HCA-2200II is used.

—Ed.

The Audio Critic:

I am writing to respond to Mark S.
Willliamson's letter, as well as the excel­lent article on clock jitter by Robert W.
Adams, in Issue No. 21. First, Mr. Wil­liamson states, "My only regret is that
some of the information is a little techni-

cal for those 'laymen' who are not part of
the engineering kingdom. I would be
grateful if you could dilute some of the
techno-lingo from time to time."

Williamson has been complemen­tary to The Audio Critic as am I; howev­er, the idea of catering to a less technical
(or nontechnical) readership sends shiv­ers up my spine. Dilute? To what end?
I'm not trying to attack this man. I just
don't want you to do what he says. I have
been reading audio publications for 25
years, with a craving for the detailed arti­cles and technical subjects that appear in
every issue of The Audio Critic. There
are many audio publications that are al­ready diluted for those who have no
stomach for details. If I want serious,
scientific, technical analysis of audio is­sues, there are only two choices: TAC at
$24 per year, or pay hundreds for a pro­fessional journal such as that of the Au­dio Engineering Society (AES). I worry
that there will be continuous pressure on
TAC to broaden its appeal by deleting the
important details and writing for the mar­keters. I must admit to you that I am a
practicing electronics engineer with a
B.S.E.E. under my belt, so it figures that
the technical details would be much to my
liking. I do not wish to demean Mr. Wil­liamson's comments at all, and am
pleased that he has joined the thinking
among us. The physics of the universe is
a complicated thing, and we as mortals
must use difficult technical language to
describe it with any accuracy at all.

THE AUDIO CRITIC

pdf 9

Robert W. Adams's article on clock
jitter was simply outstanding, providing a
great analysis of the specifics of clock­jitter effects on digital-to-analog conver-

sion. This is a good example of what I am
talking about. Mr. Adams is in a position

(as few are) to write this article. We need
just these kinds of articles in order to
really understand what is going on, and
what is important. David Rich has written
several detailed articles that spoke to me
as a designer and engineer, sending me

scurrying for my semiconductor data
books and reviewing circuit theory. Ex­cellent. I have been waiting many years
for this kind of reading from the audio
press. Reading TAC is actually challeng­ing for me. I get through a Stereo Review
in about 15 minutes, but TAC takes
weeks to fully absorb.

Diluted summary: If you change

anything I'm going to get really mad.

Clark Oden
Project Engineer
Frontier Engineering, Inc.
Oklahoma City, OK

Just how technical should a respon­sible audiophile journal be? It's a
difficult question. If we oversimplify, we
become superficial—and there are al­ready plenty of other superficial audio

magazines. If we speak mainly to the E.E.
sensibility, we lose the vast majority of
our readers. We must strike a delicate
balance. As I've stated before, The Audio
Critic is not "My First Book of Electrici­ty. " If you don't know what impedance is,
or what a FET is, you'11 have to find out

elsewhere. On the other hand, we aren't
the "Journal of the AES, Junior," either.
We try to keep the math to an absolute

minimum (although David Rich always
wants to sneak in more and more). I want
the reader who doesn't understand, say,
20% of what we publish to understand

the remaining 80% perfectly. That way
the 100% understanding can be expected

to come eventually. Our basic conclu­sions must be crystal clear to everyone—
and I think they are.

One thing we can do is to break out
the highly technical stuff in sidebars, so it
doesn't slow down the nontechnical read­er of the main article. Sometimes the arti­cle is written in such a way that it's very
difficult separate the rough from the
smooth, but as you know we try to do it
when we can.

Don't worry, we don't intend to "di­lute " our technical accountability. And I
agree completely with your beautiful sen­tence about the physics of the universe.

—Ed

The Audio Critic:

.. .Keep up the tweak bashing!

Seriously, I have been working on
two processes relating to digital audio

recently—sample-rate conversion and

noise-shaped dithering. In an effort to
find out what is really bearable vs. what
is measurable, I became sucked into the

audiophile-tweak world—I was con-

vinced my sound system wasn't good

enough because I couldn't hear character­istics I could measure. After all, these
golden ears doing reviews can obviously
hear the differences—right?

I found your magazine in time to be
rescued from a fate worse than death—
financial and otherwise. Thanks.

Regards,
Bruce Hemingway
Hemingway Consulting
dB Technologies, Inc.
Seattle, WA

It's useful to know what to say first,
right off the bat, to those golden ears who
claim to hear the differences you can't.

You say, "No, you can't hear that. You're

just telling me you can but you'll never

be able to prove it." That immediately
steers the discussion in the right direction
without allowing it to go off on some
highfalutin, abstract, pseudoscientific, psy­chobabble tangent—which is what I find
worse than death.

—Ed.

The Audio Critic:

To the little person "in the smallest
room of [his] house" [Issue No. 21, page
4]. The Editor gave you credit for writ­ing, but we know it was copying, don't
we? You forgot the quotation marks on

"/ am... and ...behind me." Then you
forgot to give the composer Max Reger
credit for the quote.

You could have signed your name.

We would have understood the X.

Ron Garber
La Porte, IN

What an erudite subscriber! What a
sucker of an Editor! What a scummy
anonymous letter writer!

I looked it up, and you're absolutely
right of course. Here is what Max Reger
wrote to the Munich critic Rudolph Louis
in response to the latter's review in the
Münchener Neueste Nachrichten, Febru­ary

7,

1906:

Ich sitze in dem kleinsten Zimmer in
meinem Hause. Ich habe Ihre Kritik vor
mir. Im nächsten Augenblick wird sie
hinter mir sein.

Translation: "I am sitting in the
smallest room in my house. I have your
review before me. In a moment it will be
behind me."

One must understand that in Europe
in 1906 the use of newspaper for hygienic
purposes was quite common. As for the
quip, Max Reger (1) obviously made it up
himself and (2) signed his name.

—Ed.

Coming:

A review in depth of the $19,000 Snell Acoustics Type A loudspeaker system,

along with other interesting speakers.

Reviews of high-quality surround-sound processors and preamplifiers, from

Lexicon, Marantz, B&K Components, and others.

The long-promised survey of FM tuners and indoor antennas (really!).
Still more reviews of power amplifiers and preamplifiers (they keep coming).

Further evaluation of perceptual coding technologies and hardware.

Some big surprises (you'll never guess).

ISSUE NO. 22 • WINTER 1994-95

9

pdf 10

Paradoxes and Ironies of the Audio World:

The Doctor Zaius Syndrome

By Peter Aczel

Editor and Publisher

When the truth is so terrible that admitting it would surely make
the whole system crumble, ape logic demands denial and coverup.

Have you ever seen that marvelous 1967 science­fiction movie The Planet of the Apes? If you have, you
will recall that it depicts a planet of the future where
English-speaking anthropoid apes are the rulers and
humans are speechless beasts of burden, enslaved by the
apes and despised as a totally inferior species. The apes
have horses and guns but no real technology. Doctor
Zaius, the subtle and highly articulate orangutan who is
this society's "Minister of Science and Defender of the
Faith" (he is played by the great Maurice Evans), knows
something the other apes do not: that humans in a past
era possessed not only speech but superior technology,
flying machines, powerful weapons, and so forth, all of
which served only to bring about their eventual downfall
and reduce them to their present condition. Doctor Zaius
fervently believes that any knowledge of this truth about
humans would totally destabilize the society of apes and

result in the end of their world. The ape dogma he fanati­cally protects, even though he knows better, is a blatant
denial and coverup of the actual history of the vanished
human civilization and a paean to the eternal superiority
of the ape.

I won't give away the rest of the plot to those of
our readers who haven't seen the movie and may want to,
but doesn't Doctor Zaius resemble certain key figures in
the high-end audio community? He knows the truth but
it's bad for the establishment. The system would come
crashing down if the truth were revealed. To pick an ob­vious example, consider John Atkinson, the subtle and
highly articulate editor of Stereophile. Don't you think he
knows? Of course he knows. But if he admitted that
$3000-a-pair speaker cable is a shameless rip-off or that
a $7000 amplifier sounds no different from a $1400 one,
the edifice of high-end audio would begin to totter—or so
he thinks (and may quite possibly be right). Consequent­ly, he spouts convoluted scriptural arguments and episte­mological sophistries, just like Doctor Zaius, in order to

pervert the obvious, uncomplicated, devastating truth.

There is a perfect illustration of this process in the

August 1994 issue of Stereophile, where Zaius-Atkinson
once again bashes blind listening tests in an "As We See
It" editorial. Such tests are of course considered extreme-

10

ly threatening by a publication that reports night-and-day
differences in sound which absolutely nobody can hear
when the levels are matched and the brand names con­cealed. He brings up all kinds of intricate flaws and draw­backs that may very well exist in some blind tests but
turns his back on the large number of blind tests in which
all of his objections have been anticipated and eliminated
and which nevertheless yield a no-difference result every
time. He knows very well, for example, that no one has
ever, ever proved a consistently audible difference be­tween two amplifiers having high input impedance, low
output impedance, and low distortion, when operated at
matched levels and not clipped—but like Doctor Zaius
he conceals that knowledge. He'd rather collect rare case

histories of screwed-up blind tests than deal with the vast
body of correctly managed blind tests that undermine the
Stereophile agenda. (Just for the record, I'll state for the
nth time that there are only two unbreakable rules in
blind testing: matched levels and no peeking at the name­plates. To eliminate "stress," take a week or a month for
each test, send everybody else out of the room, operate
the switch yourself at all times, switch only twice a
day—whatever. The results will still be the same.)

A hard-nosed insight by the Weasel.

Our columnist Tom Nousaine (a.k.a. the Weasel),
in a recent conversation with me, stated his belief that
any longtime audio reviewer who has tested hundreds of
different audio components over the years knows exactly
what the truth is about soundalikes because it is utterly
impossible to escape that truth after so much hands-on
experience. It asserts itself loud and clear, again and
again. Therefore, he argued, the audio journalists who in­variably report important sonic differences are most like­ly a bunch of hypocrites, i.e., exhibit the Doctor Zaius
Syndrome. I was strongly inclined to agree with him, but
then I said, "Well, what about Bob Harley?" We agreed
that Harley could be an exception. He may very well be
sincere because he just doesn't get it, not even after all
these years. Larry Archibald, on the other hand, is smart
and tough and definitely knows the truth, we felt. He is
probably the biggest Doctor Zaius of them all, ready to

THE AUDIO CRITIC

pdf 11

make a monkey of any insufficiently enlightened audio­phile. At the risk of offending against the principle of de
mortuis nil nisi bonum, I'm willing to venture the opinion
that even the late Bert Whyte and Len Feldman, regard­less of their other important contributions, did a Doctor
Zaius number on certain audio issues rather than face the
wrath of the tweaks and the accusation of heresy. As for
Harry Pearson and company, who knows? Are astrolo­gers, shamans, and witch doctors sincere or hypocritical?
As long as they don't try to usurp scientific arguments,
what difference does it make? And if they do try, they're
pathetically ineffective anyway.

A serious credibility gap.

In the same issue of Stereophile as the John Atkin­son blind-test-bashing editorial, Larry Archibald views
with alarm the low bit-rate coding scene in an open letter
to Pioneer. He wants them to hold off on the implementa­tion of the Dolby AC-3 coding standard for LaserDisc
because it may not be the highest-quality solution soni­cally. In other words, he suddenly doffs his orangutan
suit and shows concern for something that may actually
be true, i.e., audible.

Well, you blew it, Larry baby. You went ape—or
cried wolf, to mix my animal metaphors—so many times
about low-credibility tweako matters that on the Pioneer
level of big-money decision making you are no longer
taken seriously even when you may have a perfectly le­gitimate, nontweako argument. That's quite obvious
from the two replies by Pioneer executives printed in the
September issue, both of which basically tell you to re­lax, tweak boy, take it easy, and let the real experts get
on with their work—at least that's the way I read them.
You may conceivably end up being right, and Pioneer
wrong, about Dolby AC-3, but you're clearly wasting
your breath. (See also Tom Nousaine's column in this is­sue for a somewhat different perspective.)

By the way...

As I recently noted with a poignant sense of recog­nition, Stereophile's visual leitmotiv for blind testing is
the familiar three apes with hands on eyes (get it?), ears,
and mouth. Now you know why. It isn't just an art direc­tor's passing fancy. Doctor Zaius can feel right at home.

Why do I even bother to tell you all this?

All of our readers who have been with us for more
than just one or two issues are aware of my enormous
frustration on the subject of scientific truth in audio. The
very idea of a Doctor Zaius Syndrome, even it's only a
parody, suggests the existence of antiscience in audio as
a tradition, not just a momentary aberration—and a tradi­tion it is, going back to the early 1970s, at the very least.
In the late '40s and throughout the '50s and '60s, whatev­er the most highly qualified and experienced engineers
said about audio was the accepted truth. Then came post-

ISSUE NO. 22 • WINTER 1994-95

modern irrationalism, post-Watergate anomie, fortune
tellers in high places, pyramid power, Jesus-haired
record-store clerks as self-proclaimed audio experts, un­tutored high-end journals, pooh-poohing of engineering
societies, derision of degreed academics—the B.S. era of
audio (and I don't mean Bachelor of Science). Today, the
melancholy truth is that tweako cultism has become
mainstream audio, at least above a certain price range,
and engineering facts are regarded as disturbingly radical
or at least eccentric. The scientific audio community has
been marginalized.

I despair at this point of a journalistic solution.
Even if The Audio Critic increased its circulation by a
factor of 50 overnight—I'm being deliberately absurd—
it might still be too late for the message. The cultists
have been too deeply indoctrinated and too long. The
pimply-faced kid in the Bon Jovi T-shirt who tried to sell
you AudioQuest Sorbothane Feet (the bigger kind) in
your local audio salon is not going to change his belief
system. Not in this antirationalist age and culture.

I can think of only one effective remedy. Many
years ago, long before our younger readers became inter­ested in audio, the Federal Trade Commission put an end
to fraudulent power-output claims in amplifiers. Today,
the power-output specification must take the form of
"200 watts rms into 8 ohms from 20 Hz to 20 kHz at less
than 0.25% total harmonic distortion." Before then, the
same amplifier could have claimed 800 watts because it
could produce that for 2 milliseconds at 1 kHz into 2
ohms with 10% distortion. What if the FTC suddenly be­came interested in audio cable advertising, for example?

That chattering sound you hear comes from the teeth of
cable vendors at the mere mention of the possibility. And
that low, rumbling sound you hear is Doctor Zaius growl­ing, "That's heresy!"

Anyone out there whose nephew or brother-in-law

is a young, crusading, Ralph-Nader-like employee of the
FTC? Get him interested!

* * *

P.S. Long after the above was written, just before

press time, I received a PR release from one of Stereo-
phile's flacks, hyping the magazine's willingness to tell
the "absolute truth" about a product even at the risk of
losing the advertising support of the manufacturer. The
latest editorial by Larry Archibald-Zaius simultaneously
proclaims the same lofty principle. The Velodyne
brouhaha is used in both instances as proof: they panned
the DF-661 speaker; Velodyne canceled all its ads; see
how incorruptible they are. Hey, you can't buy off the
Defenders of the High End Faith with a few ads when
they face the deadly threat of a midpriced super speaker!

"I [don't] see any reason...why a magazine

couldn't have both principles and commercial success,"
the PR release quotes Archibald-Zaius. "I've never had
even a second thought on the subject."

The monkey doth protest too much, methinks. •

11

pdf 12

Loudspeakers Are

Getting Better and Better

By Peter Aczel

Editor and Publisher

The proof is in the recent designs that nudge the state of the art,
particularly in bass reproduction, but also in minimizing distortion
over the full range.

I can't say it often enough: if you already own a
fairly decent home music system, nothing can sig­nificantly change the quality of your audio life except
new and better loudspeakers. They are so much more im-

portant than preamps, power amps, CD players, etc.
What happens in most audiophile households, however,
is that the main speakers are firmly embedded in the liv­ing-room décor, so that any major change is subject to
vigorous spousal objection, especially if larger speakers
are contemplated. The typical audiophile then satisfies
his lust for shiny new equipment by buying, say, a new
preamplifier and persuading himself that the sound is
now much better, when in effect it hasn't changed the
least bit. It's a syndrome that depresses the hell out of me.

(Incidentally, I was recently exposed to a pair of

rather large loudspeakers packaged in a novel way that
could conceivably overcome spousal objection, even
though at first glance the speakers actually appear to be
larger than they are. The system, not yet sold anywhere,
is called the Applied Acoustics Model 10A and is the
brainchild of Jim Suhre, a Raytheon rocket scientist (real­ly!), and Vic Kalilec, an electronics engineer, both from
Tennessee. What they showed me was a monumental
floor-to-ceiling wall system, the most salient feature of
which is its gigantic, seamless, curved tambour door.
Open the door and all your electronic equipment is in
there, including your large-screen TV. Close the door
and the whole shebang looks structural, not like audio
equipment. Jim Suhre claims that the curved tambour
acts as an ultrasophisticated dispersion device. I have my
reservations about that but can report that the sound had
exceptionally even spectral balance, from the lowest to
the highest frequencies. The speakers, when their floor­to-ceiling grille is taken away, are revealed to be only
chest high; I'd say they have the overall impact of B&W
Matrix 803's or something along those lines. The drivers
and network appear to be of the highest quality. The
price was still up in the air when I looked; those who go
for this sort of thing will no doubt be able to afford it.)

12

ACI "Spirit"

Audio Concepts, Inc., 901 South 4th Street, La Crosse, WI

54601. "Spirit" floor-standing 2-way loudspeaker system,
$499.00 the pair (direct from ACI). Tested samples on loan

from manufacturer.

"In this world nothing can be said to be certain, ex­cept death and taxes," Ben Franklin wrote. That may
have been true in 1789, but today I would add: "and good
value from ACI." The combination of direct marketing,

conscientious design, and just plain common sense have
made Mike Dzurko's trademark synonymous with "more
speaker for your dollar." That's certainly true of a pair of
Spirits, which satisfy all the basic audiophile demands,
except for the deepest bass, at the totally unexpected
price of $499.

The speaker is a 32" high box with a footprint of
less than a square foot, housing an 8" woofer with poly­propylene cone and a 1" aluminum-dome tweeter. The
woofer is aperiodically loaded with four small holes in
the back of the cabinet near the floor; the tweeter has a
plastic dispersion plug and is surrounded with felt. The
left and right speakers are mirror-imaged. The oak veneer
of my samples was of good quality.

The impedance curve of the Spirit shows the box to
be tuned to 50 Hz and the minimum impedance of the
system to be 6½ ohms (8 ohms nominal). Total impe­dance variation is between that minimum and 33 ohms in
magnitude and ±45° in phase. Any decent amplifier
should be able to drive such a load.

The quasi-anechoic (MLS) frequency response was
interesting in that both the woofer and tweeter were quite
flat, within ±2 dB or so, but the transition between them
in the 2 to 3.5 kHz range was not smooth, showing vari­ous irregularities of the order of 7 dB, depending on the
angle of measurement. Since all these irregularities were
basically minus (i.e., not peaks) and in the most sensitive
range of the ear, they may have acted as inadvertent zip-

THE AUDIO CRITIC

pdf 13

piness suppressors. At this price, you can't expect the
crossover network to be too sophisticated; it appears to
be second-order, with out-of-phase wiring of the drivers.
The tweeter has a double resonance at 17 kHz and 24
kHz; otherwise it's quite smooth and remarkably clean in
response to tone bursts; there is no ringing at any fre­quency. The same is true of the woofer cone.

The nearfield response of the woofer shows the f

3

(-3 dB point) to be 50 Hz, with only a 12 dB per octave
rolloff below that frequency. That means useful response
down to 35 Hz or so. Can't ask for much more than that.

The sound of the Spirit is essentially neutral. That's
a simple statement but not a simple achievement. More
than a few extremely costly speakers don't sound neutral.
Transparency is good but not superb (what did you ex­pect?). Dynamic range is very good. I didn't measure the
distortion because it's obvious from the drivers that it
can't be either very low or very high, and the process is
time-consuming; the sound, let me assure you, is quite
clean at high levels. All in all, this is a highly acceptable,
far from puny-sounding, musically pleasing little loud­speaker. I was demoing a very high-end speaker to a
friend, and then switched to the Spirits. The difference
was obvious but not very dramatic. Hey, for $499?

Bag End ELF Systems
S10E-C and S18E-C

(continued from Issue No. 21)

Modular Sound Systems, Inc., P.O. Box 488, Barrington, IL

60011. Voice: [phone]. Fax: [phone]. ELF-1
two-channel dual integrator electronics, $2460.00. S10E-C
black-carpet enclosure with single 10" woofer, $234.00 each.
S18E-C black-carpet enclosure with single 18" woofer, $658.00
each. Tested samples on loan from manufacturer.

I see this equipment in a slightly different perspec­tive now that I have measured it and evaluated it at con­siderably greater length. I don't take back my original
statement that "I have never heard bass like this in my
listening room," but the reason for that was not the ELF
technology. It was the combination of the air-moving ca­pability of two 18" drivers, flat response all the way
down into the 20-to-10-Hz octave, excellent damping,
and little or no dynamic compression—a combination I
hadn't previously experienced, as a total package, in my
listening setup.

I now believe that good conventional technology
could achieve the same results—even if it rarely does, for
various reasons. (See also David Rich's sidebar on the
subject and the Velodyne Servo F-1500R review below.)
Since an ELF-1 unit plus two S18E sub woofers plus two
high-powered amplifier channels could run into $5000 or
more, the question is whether or not the ELF approach
yields any substantive benefits in a domestic sound sys-

ISSUE NO. 22 • WINTER 1994-95

tern (as distinct from professional applications—see be­low) that are not obtainable by simpler means for less
money. My present feeling is that, from the stay-at-home
audiophile's point of view, the ELF system is "a solution
in search of a problem." That doesn't make it sound less
good than I said, but the I've-got-to-have-it factor is gone
because what makes it sound good isn't its uniqueness.

My measurements showed the S10E and S18E to
be almost identical in small-signal frequency response
when driven via the ELF-1 (with all switches down, the
most wide-open setting). The deviation from absolute
flatness is ±0.5 dB down to 20 Hz, dropping to -4 dB at

10 Hz. The Bag End literature shows curves for the S18E
that indicate -2 dB response at 10 Hz, which is more
consistent with an f3 (-3 dB frequency) of 8 Hz, the
claimed system cutoff. I see no reason to make a federal
case out of the discrepancy; the measured f3 of 12 Hz is
good enough for me. Of course, as the signal is in­creased, the air-moving and power-handling capability of
the S18E quickly passes that of the S10E; on the other
hand, multiple Sl0E's in a cluster could keep up with the
S18E in all respects.

The f3 also moves up, inevitably, as the level is in­creased; either the driver or the amplifier (remember, it's
boosted 12 dB per octave), or both, will reach a limit in
linear output capability. The ELF concealment circuit
deals with this very neatly (again, see sidebar); it is prob­ably the cleverest and most original element of the total
system. You can crank the volume to any level you wish;
at some point when, say, the bass drum is thwacked, the
concealment threshold lights come on, but you hear no
distortion and are unaware of compression; it's smooth
as silk. You could argue, of course, that this is needed
only because the system has the inherent weakness of be­ing based on tremendous electronic boost—one compli­cated mechanism to correct the side effects of another
complicated mechanism.

I measured the distortion of the Bag End subwoof­ers with the ELF concealment threshold set to leave the
circuit inactive at the levels tested (all switches down).
That way I was measuring the true electroacoustic accu­racy of the equalized transducers. Needless to say, the
amplifier was at all times operating well below clipping.

On the whole, the distortion figures were not im­pressive. For example, at 30 Hz, as I gradually increased
the 1-meter SPL from 80 dB to 95 dB, the distortion of
the S18E as measured very close to the cone went from
2% up to 4%. At 40 Hz, the distortion over the same SPL
range varied between 1.1% and 1.7%. At 20 Hz, I mea­sured 2.6% (80 dB) to 10% (95 dB). Compare that with
the Velodyne F-1500's distortion figures (see below) and
you'll begin to understand the design priorities of each.
The S10E has a very similar distortion profile down to 30
Hz but gets worse at 20 Hz, as you'd expect. Overall,
even the bargain-priced Hsu Research HRSW10 sub­woofer (Issue No. 19, pp. 21-22) beats the Bag Ends on

13

pdf 14

14 THE AUDIO CRITIC

What the Bag End ELF System Does and Doesn't

By David A. Rich, Ph.D.

Contributing Technical Editor

OK, here we are stuck in a sidebar

because of all the equations I want to use.
But do not worry; it really is not so com­plicated. The first thing we need is the
transfer function of a sealed-box-loaded

speaker:

The amplitude and phase of a loud­speaker's output are related to its input
signal through the system function that
governs its operation. For a sealed-box
system, this turns out to be a second­order differential equation. The expression
above uses the Laplace transform, which
relates time functions to frequency­dependent functions. What it is going to
allow us to do is to manipulate the differ­ential equations using just algebra. (Yes,
I realize that those who are sophisticated
in mathematics have just gone into clip­ping over this oversimplified statement.)

The expression above is a second-order

highpass filter. Above the frequency w

0

the output signal level is the same as the
input. Below w0 the signal output de-

clines at 12 dB per octave. This w0 is the

-3 dB cutoff frequency in radians per
second (Hz x 2π) that you already know

about. The Q in the equation is the Q you
have seen mentioned in this magazine
many times. The value of Q determines
the shape of the response near w0. If Q is
above 0.707, the response will have a
peak before rolloff. Now let us assume
that you want to put in series with the
loudspeaker an electronic equalizer that
would change the value of w0 to a lower
value w1. Physically the equalizer boosts
the input signal to the speaker between
the range from w0 down to w1 The trans­fer function of an equalizer that will do
this is given below.

An electronic equalizer that generat­ed the required response was a product
from Allison Acoustics, called the Elec­tronic Subwoofer. This device came out
in 1978. It is no longer available, perhaps

because it only worked with sealed loud­speakers that had a Q of 0.707 and it re­quired the user to know the value of w0.
The other problem with the Electronic
Subwoofer was that it could boost power
into the speaker by more than 10 times,
depending on the ratio of w0 to w1 This
required a powerful amplifier and a loud­speaker with low distortion characteris­tics below its cutoff frequency w

0

The ELF Concealment system used
by Bag End neatly solves the last prob­lem by making the value of w0 dependent
on the power level. The value of w1 is in­creased as the power level of the signal
becomes higher in the range below w0.
This prevents the amplifier from clipping
and prevents the subwoofer from distort­ing or bottoming out. The method by
which w1 is moved is proprietary, but I
would expect it is similar to the methods
used by Dolby Laboratories in the Dolby
noise-reduction systems. Now w1 can be
set to have a bottom limit at a very low
frequency. In the case of the Bag End
ELF systems this is around 10 Hz. The
result is very extended low-frequency re­sponse and also very low phase shift at
the higher frequencies where the sub­woofer will be crossed over. This makes

it easier to cross the speaker over without
amplitude variations in the crossover re­gion due to subwoofer phase shift. Be­cause w1 can be set very low, the Bag
End ELF system adds an additional sec­ond-order lowpass filter fixed at 10 Hz in
order to prevent subsonic overload. (The
Allison Electronic Subwoofer also had a
subsonic filter.)

The Bag End ELF system works a
little differently than the basic configura­tion I started this explanation with. The
modification involves setting the sub­woofer's w0 to a very high value of, say,
440 rad/sec (70 Hz). This modification
permits reduced size or improved effi­ciency (which limits cone movement) or
a negotiable combination of both. Since
the equalizer is going to introduce sig-

nificant boost, this is clearly an important
feature, but the advantage is somewhat
reduced by the fact that the maximum
boost at low frequencies (w0/w1) is now
higher, since w0 is larger.

The equation to equalize the Bag

End woofer is given below.

Original Electronic
loudspeaker equalizer
response

New Lowpass filter
loudspeaker to be used for
response the crossover

As can be seen, in the Bag End ELF
approach an additional term appears on
the right-hand side of this equation. This

is a second-order lowpass filter with a
cutoff frequency of w0 (C. 70 Hz for the
Bag End subwoofer) and a Q of 0.707 (that
of the Bag End subwoofer). The equalizer
has a Q = 0.5 because it is formed by cas­cading two first-order equalizers.

The ELF electronics contain a
matching second-order highpass filter to
be cascaded with the main loudspeaker.
The cutoff frequency of this filter can be
varied from 50 Hz to 205 Hz in 5 Hz in­crements. If the main speaker had a cut­off well below 70 Hz—but in that case
why do you need a subwoofer?—then the
highpass filter would be set to 70 Hz for
flat frequency response. But we have a
problem, since this requires that the w0 of
the subwoofer, which is based on me­chanical parameters, should match the
electrical parameters of the highpass
filter. Any mismatches and frequency re­sponse errors will occur around the cross­over frequencies. The same problem ex­ists with the Allison Electronic Subwoofer,
but here the frequency aberrations will
occur at the cutoff frequency of the main
loudspeaker, typically around 40 Hz. The
ear will be less sensitive to this than in
the region of 70 Hz.

But a bigger problem is that the
highpass crossover of 70 Hz is not going
to be correct for many kinds of main
speakers, especially not for small moni-

Original
loudspeaker
response

Electronic
equalizer

New loudspeaker
response if Q = 0.707

pdf 15

distortion. To say something nicer, the damping of these
sub woofers is superb; low-frequency tone bursts in the
30 to 70 Hz range excite virtually no spurious output, re­gardless of the number of cycles in the burst. (That's
what the untutored would call "fast.")

As for impedance, both the S18E and S10E are es­sentially 6-ohm units, but the fundamental resonance,
which determines the ELF lowpass "corner" frequency,

is slightly different for each: 62 Hz for the S18E, 67 Hz
for the S10E, at least in my samples. Sample-to-sample
variations can be assumed; in a general discussion of the
ELF system, 70 Hz would be a safe round number to use.

Since the ELF-1 unit acts as a front end just before

the power amplifiers, I measured it as if it were a line­level preamplifier. At both ELF Out and HF Out in each
channel, the THD + N versus level curves appeared to be
noise-dominated, as in any decent preamp. The minima
at both outputs and at most frequencies were in the -81
to -87 dB range, with soft clipping before the 10 V abso­lute maximum output. That's par for the course. The 20
kHz distortion profile at HF Out, however, wasn't very
good. Even with the most favorable settings, the 20 kHz
distortion was at its minimum of -74 dB at only 650 mV
out, then climbed steeply to -60 dB at 4 V out and worse
beyond that. That's a classic case of dynamic distortion.
Since no circuit schematic was available, only a block di­agram, I didn't involve David Rich in diagnosing the
possible causes of this, but it doesn't really matter. We
aren't dealing with audible levels of distortion here but
only engineering refinements. Still, why not do it 100%
right? Channel separation at HF Out was 45 dB at 20
kHz, 66 dB at 1 kHz, and 97 dB at 20 Hz. Again, that's
OK, but just. The ELF-1 doesn't scale new heights as an
analog front end.

My present conclusion about the Bag End ELF ap­proach, now that all the tests and analytical comparisons
are behind us, is that it probably still works best in pro­fessional sound, where the whole thing started in the first
place. A rock group, for example, could set up large clus-

ters of the highly portable and cost-effective S10E and
possibly get away with using only a single expensive
ELF-1 to drive all the amplifiers. That makes sense. For
residential audio this is not the system of choice as I now
see it. Of course, if you've already bought it, you've got
one hell of a subwoofer setup, capable of stupendous
bass, but you could have obtained equal or possibly even
superior results by simpler and less costly means.

Velodyne DF-661

(continued from Issue No. 21)

Velodyne Acoustics, Inc., 1070 Commercial Street, Suite 101,
San Jose, CA 95112. Voice: [phone]. Fax: (408) 436-

7276. DF-661 "Distortion-Free Full-Range Loudspeaker,"
$1695.00 the pair in black vinyl, $2245.00 the pair in rose­wood. Tested samples on loan from manufacturer.

After several months of exposure to this radically
new and different speaker system, and after having com­pleted the usual laboratory measurements (and more), I
must reiterate my initial impression that we are dealing
with a very important, groundbreaking design here. It has
flaws, but its virtues are infinitely greater. There may be
a problem, however, with that evaluation. It is possible,
though by no means certain, that our review pair was
measurably and audibly superior to the samples tested by
other reviewers (John Atkinson for Stereophile, Julian
Hirsch for Stereo Review, David Moran for the Boston
Audio Society, et al.). David Hall, president of Velodyne
and designer of the DF-661, told me that there had been
some early QC problems with the aluminum-cone driv­ers; conceivably we lucked out and got a perfect pair.
Now, an engineering design is as good as the most per-

fect available sample, but a commercial product is only
as good as the least perfect sample that leaves the factory,
and that's something I can't possibly pass judgment on.

The deployment of drivers on the front baffle of the

tors, since they may already be rolling off
by then. And, most subwoofers should be
crossed over with a slope of at least 18
dB or 24 dB per octave. Higher-order
crossover slopes limit the range over
which the speakers interfere with each
other and reduce the amount of low­frequency energy sent to the main speak­er. Given the unchangeable 12 dB per oc­tave slopes of the ELF crossover, the
ideal match for the subwoofer would be a
main speaker with flat response down to
at least 35 Hz, so that in the 70 Hz cross­over region the highpass filter alone
would control the main speaker's low­end rolloff.

Both the Bag End ELF woofer and
the Allison Electronic Subwoofer are ex­amples of open-loop correction systems.
A closed-loop correction system using
feedback is clearly a more sophisticated
way to extend low-frequency response
and reduce distortion. The problem with
a closed-loop system is the transducer
(motion sensor) required to convert the
acoustic signal back to electrical energy.
Any errors in the transducer's transfer
response will be reflected in the speaker's
closed-loop response.

It is unfortunate that the good basic
design work by the engineers of the Bag
End ELF system is tarnished by the liter-

ature supplied by the company. The liter­ature makes claims for the system that
are not true, and more often than not mis­states or oversimplifies how the system
works. A good example of the oversimp­lification is the use of the concept of an
integrator to explain how the equalizer in
the subwoofer path works. That's not
how it works, as explained above. An ex­ample of complete inaccuracy is the
claim that the signal delay in the Bag End
ELF system is short and fixed because in­tegrators have fixed delay. As can be seen
from the above, the delay is not any dif­ferent from that of any other loudspeaker
with the same low cutoff frequency.

ISSUE NO. 22 • WINTER 1994-95 15

pdf 16

DF-661—from top to bottom: 6" midrange driver, 1"
dome tweeter, 6" woofer—requires aiming the measuring
microphone at the tweeter for correctly summed on-axis
and off-axis measurements. The quasi-anechoic MLS
technique yielded much flatter response on axis, both at 1
meter and 2 meters, than reported by others. The tweeter
(made by LPG but labeled by Velodyne) appeared to be
almost amplifier-flat above 7 kHz, right up to the ultra-

sonic resonance. Below 4 kHz, the response is again ex­tremely flat, within ±2 dB down to 300 Hz. Only between
4 kHz and 7 kHz is there a larger deviation, namely a 4
to 5 dB dip centering approximately on the 5 kHz cross­over frequency. This dip becomes a deep and wide notch,
of the order of 6 to 8 dB, in the off-axis area of radiation.
A nonoptimal crossover network is probably the culprit,
possibly aggravated (and certainly not helped) by diffrac­tion from the sharp edges of the cabinet. The crossover
schematic shows that the tweeter is brought in with a
very steep (24 dB per octave) highpass filter, ostensibly
to restrict the moving system to its linear range, whereas
the midrange driver is rolled off on top with a 12 dB per
octave filter. There may be a mismatch between the two

filters. Another possibility is that the midrange driver is

simply too large to have good response up to 5 kHz over
a reasonable solid angle. Even so, the DF-661 still has a

better frequency-response profile than the majority of

speakers, at least on the basis of our samples. Except...

Except, yes, for the bass. There isn't much of it.
The vented-box system appears to be tuned to something
between 60 and 62 Hz, but it is a very peculiar kind of
tuning, further confused by a 250 µF capacitor in series
with the woofer. The purpose of the capacitor is to re­strict the excursion of the driver and keep the distortion
very low, but it also makes the design parameters of the
bass system hard to determine with standard measure­ments. In any event, this is at best a weakish 60 Hz box,
meaning that a subwoofer is needed for full-range perfor­mance, and that makes the funny bass more or less moot,
at least in my book.

The impedance curve of the DF-661 is quite rea­sonable above 100 Hz, fluctuating between 3.2 and 8.5
ohms in magnitude, and remaining within ±30° in phase
(the latter only above 200 Hz). At the lower frequencies
that huge capacitor takes over; the magnitude goes way
up and the phase goes highly negative—what else did
you expect?

So why am I so enthusiastic about the DF-661? Be­cause, to my ears, it sounds more transparent, more accu­rate in rendering instrumental and vocal timbre, more
lifelike in overall tonality than any other speaker I was
able to compare it with, either side by side or from mem­ory. The reason for that is almost certainly the ultralow
distortion, which was David Hall's main design goal. In
fact, he explained to me that he had paid only routine at­tention to other parameters and focused almost entirely
on distortion "because I wanted to prove a point." I find

16

his point proven because the DF-661 is merely a decent
but not great speaker in all other respects and yet it sets a
new standard in transparency. It must be the low distor­tion; there's nothing else to account for it.

How low? I measured the woofer and the midrange
driver separately, each driven with a continuous signal to
a midband level of 95 to 96 dB at an axial distance of 1
meter. (That's loud!) The distortion reading was in the
nearfield, right off the diaphragm. The midrange driver
did not exceed 0.2% THD + N at any frequency between
800 Hz and 4 kHz, remaining well below 0.1% over most
of that range and dipping as low as 0.04%. That's more
like an amplifier than an electroacoustic transducer. The
woofer averaged 0.1% in the 125 to 800 Hz range, dip­ping as low as 0.03%, but climbing below 125 Hz to a

peak of 1.3% at 63 Hz. Pretty damn impressive—unless

you think, like some nitwits, that distortion matters only
in the electronic signal path but not the speaker. Well,
my ears tell me that it matters a great deal.

Please note, however, that with my priorities I rep­resent one kind of listener; as I wrote in the last issue, "I
am basically a tonality/balance/transparency man." You
may be an imaging man, or a front-to-back depth man (or
woman—sorry), in which case the DF-661 may not be
your cup of tea. The soundstage it presents is a little flat
in comparison with some speakers (e.g., the Win SM-8,
reviewed below). The reason may again be the crossover
problem or diffraction or both; indeed, the deep crossover
notch off axis may be part of the syndrome. As I said, the
design is not without flaws. The network, by the way,
drives the midrange out of phase with woofer and tweet­er; the second-order (12 dB per octave) slopes of the
midrange bandpass filter appear to be the reason why.

The drivers themselves, in addition to producing
extremely low harmonic distortion, are also very clean in
terms of freedom from ringing. I made more tone-burst
measurements, at more frequencies, than I ordinarily do
and found only negligible storage. Whatever weaknesses
the design may have are all related to frequency response.
(Of course, as I said before, I can only talk about the par­ticular pair I tested.)

Bottom line: I played one of my favorite Julianne
Baird tracks (on Dorian) through the DF-661, then
switched to a speaker I had previously thought of as quite
wonderful. It sounded disturbingly veiled by comparison.
(I won't identify the speaker because the same thing
would have happened with any number of others.) The
more I played DF-661, the more addicted I became to its
tonality/transparency characteristics and began to find

other, in some ways better, speakers unsatisfactory. What
I currently listen to most often—just for enjoyment that
is, when I'm not testing—is the Velodyne DF-661, ex­tended on the bottom with the Velodyne Servo F-1500R
subwoofer. (A separate review of the latter follows.)

Final thoughts: I have been told that the unfavor-

able review in the June 1994 issue of Stereophile pretty

THE AUDIO CRITIC

pdf 17

much shut down the sales of the DF-661. Now, I consid­er Stereophile quite capable of "cooking" their measure­ments and reporting worse sound than they actually
heard, because ultrahigh-end performance out of a lowly
upper-medium-priced speaker simply cannot be allowed
—it's against their religion. (Remember the number they
did on the Carver "Amazing"?) Such ugly thoughts must
be quickly reconsidered, however, in the light of all the
other not-so-enthusiastic reviews. It's hard to believe
they're all just politicking. My recently calibrated Brüel
& Kjær 4133 microphone and Audio Precision "System
One Dual Domain" are too dumb to lie, and the evidence
of my ears was corroborated by several staff members, so
I must stick to my guns; even so, as I pointed out, our
samples may conceivably have been better than what
everybody else got. More than that I cannot say.

Velodyne Servo F-1500R

Velodyne Acoustics, Inc., 1070 Commercial Street, Suite 101,
San Jose, CA 95112. Voice: [phone]. Fax: (408) 436-

7276. Servo F-1500R servo-controlled powered subwoofer
with remote control, $1595.00 each (3190.00 the pair). Tested
samples on loan from manufacturer.

In Issue No. 16 I made it quite clear that I have the
highest respect for Velodyne's motional-feedback ap­proach to bass transducer design. Now that I have tested
their latest-and-greatest model—not just a single unit but
a pair—I feel even more strongly that they're doing it
right. Their distortion figures are so much lower than
anyone else's that I find myself unable to favor designs
with higher distortion regardless of other possible advan­tages. For example, in the case of the F-1500R the small­signal bass-cutoff frequency (f3) is 16 Hz as against the
Bag End S18E's 12 Hz. That's not enough of a differ­ence in low-frequency extension, at least in my opinion,
to opt for the far costlier Bag End ELF solution with its
considerably higher distortion. Remember, distortion is
the difference between the input and the output, i.e., the
index of accuracy—and accuracy is what high fidelity is
all about.

I bring up the Bag End ELF system specifically be­cause it was my top choice (see the review above) before
the Velodynes arrived. A pair of Bag End S18E's will
move slightly more air than a pair of F-1500R's, which
are "only" 15-inch woofers, but the subjectively per­ceived heft, ease, and lifelikeness on the bottom end are
very similar in a comparison of the two designs, and the

Velodyne then moves ahead of the Bag End on other

counts. The far lower distortion is only one of these; an­other is the convenience of the built-in 250-watt power
amplifier; still another the variable lowpass filter, adjusta­ble from a nominal 40 to 100 Hz, allowing the flexibility
to match various highpass contours for the satellite
speaker. With the ELF system you are married to a single

ISSUE NO. 22 • WINTER 1994-95

crossover frequency, namely the fundamental resonance
of the woofer (typically 60 to 70 Hz). All in all, I now
consider the Velodyne to be a more desirable subwoofer
for domestic music systems—but that doesn't make the
ELF system less impressive than I originally reported.

Measuring the distortion of the F-1500R requires
awareness of its compressor/limiter circuit, which holds
the output level below a certain distortion limit, regard­less of the input level. When the limit is reached, in­creased input will not result in increased output. That in
itself is a good thing—what's the point of higher output
if it's distorted?—but it restricted my full-bandwidth
measurements to a maximum SPL of 95 dB at 1 meter
(normalized to 40 Hz). At that level—which is damn
loud!—there appears to be no limiting at any frequency.
THD + N at 95 dB hovered between 0.12% and 0.2%
from 100 Hz down to 28 Hz; at 20 Hz it rose to 0.6%; at
16 Hz it was 0.8%. At 90 dB the 20 Hz distortion was
only 0.24% but above 30 Hz the THD + N matched quite
closely the 95 dB profile. Going lower in SPL actually in­creased the measured percentage of distortion slightly,
probably because of the irreducible amount of ambient
noise picked up by the microphone. Overall, this is sensa­tionally low distortion, unequaled by any woofer design
known to me. Just for the record, Velodyne provides the
following distortion spec, which I didn't verify but find

compatible with my measurements: less than 0.7% at 25
Hz at 104 dB SPL. Would it be possible to obtain com­parable distortion figures without motional feedback?
David Hall told me yes, with the same transducer design
techniques he used in the DF-661; so far, however, he
hasn't applied those techniques to a big woofer, and late­ly he has expressed doubts about the feasibility of it. As
for flatness of frequency response, I measured ±0.25 dB
over the unit's intended range—but you would expect
that of any decent design.

It should be noted that the F-1500R is a forward­firing 15" subwoofer, unlike the ULD-15 Series II re­viewed in Issue No. 16, which has a similar 15" driver
firing downward. A forward-firing cone has only mass; a
downward-firing cone has both mass and weight. The
weight (i.e., the pull of gravity) can cause stress on the
suspension and eventually affect proper centering, al­though techniques exist to counteract that problem. Also,
the straight-ahead wave-launch is more predictable and
controllable than floor loading. All in all, I'm more com­fortable with the forward-firing configuration, and so are
most speaker designers. The R suffix stands for remote
control, the latest wrinkle added by Velodyne. The slim,
handheld remote unit controls power on/off, volume
down, volume up, volume restored to manual setting, and
mute. To me this is a very minor convenience, but some
will consider it to be totally cool, and they are entitled to
their gizmos, playthings, and sitting habits.

Does the F-1500R have a weakness? A single­cycle 25 Hz tone burst into the amplifier input results in

21

pdf 18

1½ cycles of acoustical output, indicating a very slight
problem with damping at that worst-case frequency. At
higher frequencies I found no such (mini)hangover. No
big deal, in any event.

In my old ULD-15 review referred to above I ex-

pressed my wish for a pair of subwoofers instead of a sin-

gle matrixed (L + R) unit and gave reasons. Our colum-

nist Tom Nousaine, for one, disputes those reasons and
has some good arguments to the contrary, as well as ex­perimental data. Others, including David Hall, side with
me. This is a legitimate controversy (unlike so many stu­pid tweako hassles in the high-end press) and merits full­er treatment, perhaps even a complete article. Meanwhile
I'm as happy as the proverbial pig in excreta with my,
count them, two F-1500R's. Matrix, schmatrix, I refuse
to part with either one of them.

Win SM-8

Win Research Group, Inc., 7320 Hollister Avenue, Goleta, CA

93117. Voice: [phone]. Fax: [phone]. SM-8 Stu­dio Monitor, $5000.00 the pair, including stands. Tested sam-

ples on loan from manufacturer.

This is a derivative of the marvelous Win SM-10
reviewed in Issue No. 17—not better, not less good, but
an alternative design with somewhat different trade­offs—and the reader is referred to that review for (1) my
comments on Dr. Sao Zaw Win as an audio designer, (2)
the mathematical credentials behind the Win speaker­design efforts, and (3) a complete discussion of the black

lacquer cabinet, which is identical in both models.

The immediate raison d'être of the SM-8 is that the
SM-10 is no longer being made, at least until a legal has­sle between Win Research and Matsushita (Panasonic)

has been resolved. (The SM-10's diaphragms, magnets,

and voice coils were Win designs; the frame, however,
was based on Matsushita tooling, which Win had pur­chased exclusive rights to—but Matsushita apparently
thought otherwise. The controversy is too convoluted for

further comment here.) The SM-8 adheres to the same
rigorous design philosophy as the SM-10 but instead of
the latter's coaxial configuration it uses a conventional

woofer/tweeter layout. As a result, coherence in the
nearfield is not as perfect, but there are also advantages,
as we shall see.

The new 8" woofer of the SM-8, covering the

range up to 2 kHz, is considered by Sao Win himself to
represent his best work to date. It has a flat diaphragm
made of molded balsa-wood pulp, a completely new idea.
Flexing is the major problem with all speaker diaphragms
above a certain size, and this solution is claimed to be a
breakthrough in that regard. Other diaphragm materials

represent various kinds of compromise between stiffness,
mass, and internal damping; balsa combines a very high
stiffness-to-mass ratio with good self-damping character-

ISSUE NO. 22 • WINTER 1994-95

istics and is claimed by Sao Win to be a no-compromise
substance for the purpose. The balsa is pulped, and the
slurry is poured into an intricate gridlike mold, designed
with computer simulation techniques and finite element
analysis to obtain the maximum possible structural
strength in the resulting flat membrane. It isn't your usual
speaker cone, that's for sure.

The tweeter is a highly customized inverted-dome
Accuton, using the magnet structure of their 3½ driver
but the diaphragm of their 1" unit. Special doping of the

ceramic dome in accordance with Sao Win's preferences

is part of the OEM custom specifications. Again, it's not

just a tweeter out of the catalog.

The crossover network is in a separate, very hand­some case with a transparent cover. The fully visible cir-

cuit board is a work of art. The network is a modified
4th-order Linkwitz-Riley, computer-optimized to include
the acoustical rolloff characteristics of the drivers. It's
the right choice in my opinion, especially since the 2 kHz
crossover frequency is quite close to the woofer's upper
and the tweeter's lower limit for best performance. The
design of a two-way system almost inevitably stretches
those two limits in search of the best possible handshake
between the two drivers. (That's why there are three-way
systems.) Steep crossover slopes are a big help under
such circumstances.

The integral speaker stand of the SM-8 is a definite
improvement over the SM-10 stand. Made of a heavy al­loy chosen for exceptional acoustical deadness, this black
tubular stand is very easy to mount and tilts the speaker
backward at a slight angle. The benefits of the latter fea­ture are debatable, but of course it does offer a wee bit of
delay compensation and looks high-tech.

As I said, the vented cabinet is identical to that of
the SM-10; the costly finite element analysis that had
been originally invested in that structure had to be saved,
so the woofer parameters were adjusted to those of the

enclosure rather than vice versa. My raving endorsement

of this gorgeous box can be found on page 13 of Issue

No. 17.

Since there are separate input terminals for the two
drivers to accommodate the external crossover network, I
was able to measure the "naked" woofer and tweeter. The
woofer appears to have flat response (within approxi­mately ±2 dB) and no anomalies over its assigned range,
except for a slight rise just before it starts rolling off very
steeply a 2 kHz. That rolloff point is an Achilles' heel; a
2 kHz tone burst excites significant ringing, not evident
at any other frequency. The crossover notch completely
hides this weakness when the network is connected. On
the bottom end, the vented box is tuned to 41 Hz and the
maximum output of the vent is at 53 to 54 Hz; that's
deeper bass than is obtainable with the SM-10 and is
about all that can be squeezed out of any 8" driver in an
enclosure of 0.75 cubic foot internal volume without seri­ously compromising efficiency. With just a little bit of

23

pdf 19

generosity this can be called a 40-Hz box. (You can't
deny that computer optimization has greatly helped the
design of vented systems.)

The tweeter, without the network, exhibits a some­what puzzling response profile. On axis, up to 7 kHz, it's
flat within ±1 dB. Between 8.5 and 20 kHz it's even
flatter, ±0.5 dB or at worst ±0.75 dB. Between 7 and 8.5
kHz, however, there's a 4 dB step downward. Very
strange. Off axis, still without the network, the step is
"homogenized" into a smooth 3-dB-per-octave rolloff
above 5 kHz (as measured at 30°). Also strange, since the
20 kHz response is then barely down from the on-axis
level. But wait. Inserting the crossover network changes
the picture.

The network has a tweeter level control, which I
left in the 0 (all the way up) position for my measure­ments. The on-axis response of the complete system is
virtually identical whether the microphone is aimed at
the tweeter or at the midpoint between the woofer and

tweeter centers. That response is a smooth downward
slope for a total drop of 6 dB from 2 to 20 kHz. Going
30° off axis barely changes this measurement, at least up
to 12 kHz. Between 12 and 20 kHz there is an additional
drop of 8 dB, but that has little effect on the subjectively
perceived high-frequency energy in the off-axis area. So
what we have here, instead of the usual flat response on
axis and significantly rolled-off response off axis, is an
almost invariant, gently sloping response over a large sol­id angle. For all I know, the true power response of the
SM-8 is flatter than that of a more typical "flat" speaker.
With my 1-meter MLS measurements, all I can be sure of
is the anechoic response at a limited number of micro­phone positions; I don't have an averaged power re­sponse. Certainly, the SM-8 doesn't sound attenuated on
top; if anything, it may sound a little bright on some pro­gram material with the tweeter level set to 0, in which
case the -1 setting should be tried.

Tone bursts reveal no anomalies in the drivers oth-

er than the woofer's minor problem at 2 kHz noted
above. The bottom end of the system could be slightly
better damped, but then the bass would be leaner and the
quasi-full-range impact of the speaker diminished—it's a
tradeoff. Distortion is extremely low in the tweeter; in
the crucial 2 to 5 kHz range, at a 1-meter SPL of 95 dB,
THD + N stays between 0.1 and 0.2%. That's almost as
good—at some frequencies just as good— as the distor­tion performance of the Velodyne 6" midrange driver
over the same range, proving that the 1" tweeter is not
being "stretched" excessively at its lower frequencies.
The woofer is also impresively low in distortion but not
quite as low as the Velodyne 6" low-frequency driver. In
the range from 50 Hz to 2 kHz, the Win woofer fluctuates
between 0.25% and 1.2% THD + N at a 1-meter SPL of

95 dB. Ignoring the widest swings, the distortion at that
hefty level can be said to stay mostly within 0.45% to

0.75%. That still makes the SM-8 the silver medalist in

distortion competition, I think. The impedance curve of

the system is classic and unproblematic; nominal impe-

dance is 6.5 ohms.

That brings me to the evaluation of the sound (since
you can't stand the suspense any longer). I didn't fall in
love with the SM-8 as I did with the SM-10, even though
the SM-8 can be driven harder, is cleaner at high levels,
has deeper bass, and images just precisely. The SM-10
was no longer available, so there was no possibility of a
side-by side comparison, but somehow the exquisite
beauty of the string sound, the silky, delicate, finely
etched quality I remembered, seemed to be absent, at
least to some degree. That's my subjective impression;
not everyone will agree with me. According to Sao Win,
the mica-and-alumina tweeter of the SM-10 is even bet­ter than the Accuton custom unit but will not extend
down below 2 kHz as is required in the SM-8 design. He
is working on a solution and the change will be made—if
and when. The SM-8, having been designed to the same
listening standards as the SM-10, sounds very similar of
course, and some will probably prefer it for the aforesaid
advantages. As for the coaxial versus up-and-down
configuration, the latter permits larger magnet structures
and generally leaves more elbowroom for massaging the
driver designs, so the SM-8 trades away perfect coher­ence for increased large-signal capability. I don't hear
anything negative in the sound of the SM-8 that I could
specifically attribute to the absence of coaxial geometry.

If you have read the review of the Velodyne DF-661
above, you can imagine that I wanted a shootout between
it and the SM-8. As it turned out, the two were almost

perfectly matched in their farfield volume level without

any attenuators. Two associates and I spent at least two
hours each switching back and forth between the two
speakers, playing all kinds of program material. There
was enough of an instantly audible difference to obviate

blind testing but not enough to result in a quick and
definite preference. The Win has better bass and images
far better. The Velodyne excels the Win in transparency

and naturalness of timbre. That was the consensus, but
we all agreed that the two are close in quality. I contin-

ued the comparison for several days without significant

new insights to add. Visually there's no comparison; the

Velodyne looks just fine but the Win is a stunning show­piece. On the other hand, the Velodyne costs less than

half as much, and that makes it the easy winner on a per-

dollar basis. It's a nice dilemma, when you think about it.

Note: None of the above applies to a few pairs of

early SM-8's with very low serial numbers, which had a

different crossover network and different tweeter. •

24

THE AUDIO CRITIC

pdf 20

Good Things

Are Still Happening in

Analog Electronics

By Peter Aczel

Editor and Publisher

&

David A. Rich, Ph.D.

Contributing Technical Editor

Preamplifiers and power amplifiers are getting to be so uniformly

good that reviewers will soon have relatively little to review.

I laugh my head off these days reading the reviews

of preamps and power amps in the tweako magazines.

Soundstaging, front-to-back depth, midrange transparen­cy, brightness, dullness—all the pluses and minuses of
microphones, recording methods, speaker systems, and
listening rooms are attributed to these poor electronic sig­nal paths, which have demonstrably no sound of their
own unless very poorly designed or downright defective
or overdriven. They still need competent reviewing, of
course, because poor design is always an outside possi­bility, distortion usually an index of engineering sophisti­cation, adequate power into difficult loads often a need,
build quality always a measure of value per dollar, ergo-

nomics invariably a livableness factor, etc., etc.—I have
been over this same ground a number of times.

The large accumulation of well-documented dou­ble-blind listening tests offers overwhelming proof that
fanatically intense "golden-ear" comparisons of this type
of equipment are a total waste of time. The may sell mag­azines to the wishful-thinking nonprofessional but yield
no genuine consumer benefits. We still listen very care­fully to each preamp and power amp we review because
it's a simple and direct way to spot anything unexpected,
but please don't expect us to dwell lovingly on the upper­midrangeliquidity and other bathroom-literature themes.
The Audio Precision System One has the most golden ear
of any reviewer.

—Peter Aczel

Line-Level Preamplifier

Aragon 18k

(Reviewed by Peter Aczel)

Mondial Designs Limited, 20 Livingstone Avenue, Dobbs Ferry,

NY 10522. Voice: [phone]. Fax: [phone]. Ara-

ISSUE NO. 22 • WINTER 1994-95

gon 18k line-level preamplifier, $995.00. Tested sample on loan

from manufacturer.

I must refer the reader to David Rich's review of

the Acurus L10 in Issue No. 18 (page 21). The Aragon

18k is basically an improved version of that $595 "best
buy"—for a lot more money ($1.67 on the dollar, to be
precise). For even more money, you can specify the Pen­ny & Giles volume control, an extremely high-quality po­tentiometer.

What are the improvements? Current sources have
replaced the resistors to bias the differential pairs, for
even lower distortion. MOSFETs running higher bias
currents have replaced the JFETs, allowing full drive into
loads of lower impedance. The outputs are now direct­coupled, with a trim pot to cancel dc offset. Power rails
are up by 2 V to ±20 V. The beefed-up power supply is
now an outboard unit in a vaguely pyramidal (New
Age?) metal box. The input and output jacks are no long­er mounted on the PC board but are solidly affixed to the
rear panel of the main chassis—a costlier but more reli­able solution. Cosmetically the Aragon 18k also makes
more of statement, although I find the snazzy, deeply
grooved control knobs to be dirt-catchers as well as eye­catchers. The tape-monitor outputs are still paralleled,
allowing a worst-case scenario of destructive, amplified
oscillations (unlikely, yes, but not impossible).

The THD + N measurements show that the various
design improvements have had only a minor effect. The
distortion is very low indeed, but it was only 5 dB or so
higher in the Acurus L10. The Aragon 18k reaches a
minimum distortion of -98 dB at approximately 8 V out,
with both a 20 Hz and a 1 kHz input. With a 20 kHz in­put some dynamic distortion is evident at output levels
over 3.5 V, which is pretty academic. Minimum 20 kHz

25

pdf 21

distortion is in the -90 to -93 dB range. The distortion
curves are completely noise-dominated; the noise floor it­self is about average for high-quality line-level preamps.
A low-impedance load (600 ohms) has no effect on the
maximum output level or minimum distortion. In channel
separation the 18k surpasses the L10 by about 9 dB but
only at the lowest frequencies, where the 18k bottoms
out at 104 dB. At 20 kHz both models are around 73 dB.

It seems to me that Mondial Designs should be
compared to one of those great Continental hotels where
even the cheapest room is quite luxurious because they
simply don't have the mentality to do it any other way. If
you pay more you get more, but you don't need to pay
more. The Aragon 18k is one of the better rooms.

Line-Level Preamplifier

Bryston BP20

(Reviewed by Peter Aczel)

Bryston Ltd., P.O. Box 2170, 677 Neal Drive, Peterborough,

Ont., Canada K9J 7Y4.Voice: [phone]. Fax: (705) 742-

0882. BP20 line-level preamplifier, $1395.00. Tested sample on
loan from manufacturer.

David Rich, as an engineer, is one of the biggest
fans of Chris Russell as an engineer, so David should
have written this review of Chris's latest-and-greatest
preamp. Because of various intramural errors—all of
them my fault—David never had a chance to do so, and
therefore, since "the buck stops here," I am closing up
the breach.

The BP20 replaces the 11B/12B series in the Brys­ton line and represents a complete rethinking on the part

of Chris Russell of just how a preamp should go together.
The basic Bryston gain-stage topology, using operational

amplifiers made up of discrete components and ±24 V
power supplies, has not changed (at least as far as I can
see), but the physical layout and construction of the unit
are new and different. The main chassis is pancake-flat
(barely 1½" high); the power-supply transformer is a
separate external unit with an "umbilical cord." The PC
boards have been redesigned, the signal paths simplified,
and a ground plane incorporated. Each of the two chan­nels has two XLR balanced inputs and one XLR bal­anced output, five unbalanced inputs and two paralleled
unbalanced outputs, plus a tape loop. That provides a lot
of interface flexibility. There is no record selector switch
(a debatable signal-path simplification), only a toggle
switch for Tape/Source selection. A balance control with
12 o'clock detent is next to the continuously variable vol­ume control. The construction details are absolutely
beautiful; parts quality is high.

Two favorable characteristics became apparent in
the course of my measurements: (1) there was no "better"
channel, the two being absolutely identical in distortion
and noise, and (2) the balanced signal path was just as

ISSUE NO. 22 • WINTER 1994-95

low in distortion and noise as the unbalanced. This uni­formity was new in my experience. On the other hand,
the balance control at the detent was off by 0.7 dB at uni­ty gain with 2 V out (XLR in/out).

The THD + N versus level measurements yielded
outstanding results. With either unbalanced in/out or
XLR in/out, the 20 Hz and 1 kHz distortion dropped to a
minimum of -97 dB shortly before the clipping level.
The 20 kHz curve showed a small amount of dynamic
distortion but only above 4 V and 8 V out (unbalanced
and balanced, respectively), so it is of no significance.
Maximum undistorted output was approximately 14 V
unbalanced and 28 V balanced—I think that should be
sufficient for any application, don't you? The curves
were absolutely linear (i.e., noise-dominated); the noise
floor appeared to be no better and no worse than I had
seen in other topnotch preamplifiers, give or take a
couple of dB. Channel separation, one of the stumbling
blocks with a compact chassis and a balance control, was
much improved over the 11B. Under worst-case condi­tions (balanced was worse than unbalanced, unity gain
worse than full gain), the separation was 70 dB or better
(up to 85 dB) at nearly all frequencies. Messing with the
balance control made things worse at the highest frequen­cies; 54 dB at 20 kHz was the most screwed-up reading I
was able to obtain that way. This is still highly accept­able performance.

The Bryston BP20 is only the second preamplifier
we have tested—the first was the Krell KRC-2—that will
accept both balanced and unbalanced inputs and deliver
both balanced and unbalanced outputs. If that's what you
need—and in certain highly elaborate audio systems you
almost surely will—the Bryston outperforms the Krell in
most respects at 38 cents on the dollar (unless you feel
the Krell's CMOS remote control alone justifies the other
62 cents). As a purely unbalanced front end the Bryston
has some rivals, though no indisputable superiors, and its
look-and-feel make it a delight to use in any event.

Mono Power Amplifier

Marantz MA500

(Reviewed by Peter Aczel)

Marantz America, Inc., 440 Medinah Road, Roselle, IL 60172-

2330. Voice: [phone]. Fax: [phone]. Model
MA500 mono power amplifier, $299.00. Tested samples on loan

from manufacturer.

This is so simple, it's simply beautiful. A single
channel of "minimalist" low-distortion amplification
without frills, on a deep and narrow chassis with a small
frontal area, highly suitable for multichannel clustering
in a small space, at a surprisingly low price. If this is
what crass commercialism in audio engineering is all
about, I want more of it.

27

pdf 22

The Marantz engineers went for what appears to be
a genuinely smart performance vs. cost tradeoff here. The
heart of the amplifier is the µPC 1342V chip, which incor­porates all the voltage gain stages within a single IC. We
have very little data on this highly "streamlined" chip.
What we know is that it must see the full ±60 V power
supply so it can drive the discrete output devices, which

have no voltage gain. One generally wouldn't expect un-

usually high performance from a high-voltage monolithic

chip, but the measurements of the MA500 are quite re­assuring in this respect. The output stage of the amplifier
uses dual paralleled discrete transistors. The coupling ca­pacitors in the signal path are electrolytics (what else did
you expect at this price?).

The dc rails are filtered by 10,000 µF capacitors.
Overcurrent protection is almost identical to the system
used in the considerably more costly Rotel RB-990BX.
Overcurrent is sensed by a bipolar device, which is con­nected to a custom protection chip. The protection chip
drives a relay, which is in series with the speaker termi­nal. In addition to overcurrent, the presence of a dc volt­age at the output will also cause the relay to open, as will
turn-on and turn-off. This level of protection is very rare
at this price.

The MA500 is rated at 125 watts into 8 ohms and

180 watts into 4 ohms. Two MA500 units can be bridged
to create a very high-powered amplifier channel rated at
450 watts into 8 ohms. THX certification is part of the

package, guaranteeing among other things that 3.2 ohms

is a viable load with continuous power. Any number of
the amplifiers can be cascaded via their remote-control
bus jacks and turned on/off simultaneously with the re­mote control of a Marantz front end (recent models
only). All these little quality touches contradict the
"cheap" image that might be suggested by the MA500's

price. And, believe me, it doesn't look cheap, either.

My THD + N tests indicated that the specs are con­servative. Clipping into 8 ohms takes place at 140 watts
and into 4 ohms at just below 200 watts. The 20 Hz dis-

ortion dips to a minimum of -94 dB into 8 ohms and -91
dB into 4 ohms. The 1 kHz distortion is just 3 or 4 dB
higher. The minima occur shortly before clipping, and
the curves appear to be noise-dominated, with the noise
floor itself unusually low. The 20 kHz curves into both 8
ohms and 4 ohms show some dynamic distortion, but
even at clipping the THD stays below -70 dB and at re­duced power levels is considerably lower (at 20 watts,
for example, I measured -78 dB into 8 ohms and -73 dB
into 4 ohms). These results are quite a bit superior to the
performance obtainable with some very high-end power

amplifiers, showing what is possible at minimal cost

when real engineers are in charge. (Don't misunderstand
or misquote me; this is still no Boulder or Bryston, but
it's respectable.)

The PowerCube measurements showed very good

dynamic power into 8, 4, and 2 ohms, regardless of the

28

reactance of the load. At 0°, the intermittent 1 kHz tone
burst maxed out at 40 V into 8 ohms, 36 V into 4 ohms,
and 28 V at 2 ohms. At ±30° and ±60°, the readings were
slightly higher, as they should be. The dynamic head­room at 8 ohms came to approximately 0.9 dB. I was un­able to obtain reliable readings into 1 ohm because the

protection circuit started to get nervous at that point. I'm

not going to hold that against the MA500; such crazy
low-impedance loads are strictly for expensive amplifiers
with huge power supplies. At least the protection was

there.

As you can probably gather from all of the above,

this is my favorite "tourist-class" power amp so far.

Stereo Power Amplifier

PSE Studio IV

(Reviewed by David Rich)

Professional Systems Engineering, Inc., 9755 Hamilton Road,
Eden Prairie, MN 55344-3424. Phone: [phone]. Studio
IV stereo power amplifier, $995.00. Tested sample on loan from

manufacturer.

The performance of a smaller power amplifier can

often be superior to that of a larger amp because the drive

requirements to the output device are reduced, thereby re-

ducing predriver complexity. In addition, since the break­down VCE of the output devices is lower, it is possible to
find a device with a higher unity-gain current frequency.
Unfortunately, state-of-the-art designs are often reserved
for high-power, higher-priced amplifiers.

The PSE Studio IV is a state-of-the-art design with
a specified output of 100 watts per channel into 8 ohms.
A two-transistor current source, biased by a zener-diode­based voltage reference, powers the input differential
pair. The input differential pair is formed with degenerat­ed JFETs. The differential pair is cascoded with bipolar
devices and terminates to a resistive load. The second
stage is a push-pull stage with resistive load. The total re­turn loop gain is 40 dB, making this a moderate-feedback
design. On the upper supply rail, the second gain stage is
driven from an emitter follower. This important stage is
often eliminated in El Cheapo designs. On the lower sup­ply rail, the second gain stage is driven by a common­source amplifier degenerated to a gain of 1. Differential­to-single-ended conversion thus happens in the second
gain stage. The second gain stage drives a MOSFET
source follower, which in turn drives dual paralleled out­put stages. Biasing of the composite MOSFET/bipolar
stage is difficult because of the differing temperature
coefficients of the MOSFET and bipolar devices. (See
the Parasound HCA-2200II review in Issue No. 21.) In
the Studio IV, the biasing device is formed with an LED.
The temperature coefficient of the LED apparently
matches the temperature coefficient of the composite out­put stage. The output stage is current-limited by a bipo-

THE AUDIO CRITIC

pdf 23

lar-based circuit that is similar to the standard con­figuration but much more complex. This complexity is
introduced in an attempt to prevent the circuit from acti­vating into reactive loads when the current drain may be
high but the voltage across the transistor is low enough
so that the output device is still in the safe area of opera­tion. A single channel uses 20 transistors and 12 diodes.

A single toroid and pair of large 27,000 µF filter
caps form the unregulated supply rail for both channels.
A dual mono design would have added to the cost of the
unit, driving it beyond the $1000 price point. Nominal
supply voltage is 50 V. The voltage gain stages are on a
separate regulated 55 V supply. Higher supply voltage is
required here as explained in the Parasound HCA-2200II
review. Separate transformer windings and bridge rec­tifiers are used for the regulated rails. The regulators are
formed with LM324 integrated regulators. The use of a
regulated supply dramatically improves rejection of pow­er-supply noise and is an important aspect of this design
that many amplifiers costing much more than the Studio
IV fail to include. With such a regulated supply the im­provement in channel separation that a dual mono design
could effect would be small. The measured channel sep­aration was 68 dB at 20 kHz, 90 dB at 1 kHz, and 117 dB
at 20 Hz. No coupling capacitors are used in the circuit.
DC is nulled by a trim pot.

Construction is excellent. A large heat sink forms
the front of the unit. Double-sided circuit boards are par­allel to the heat sink to minimize wiring to the output de­vices. Some tweako signs in the construction include
multiple-paralleled stranded wires used for power supply
and output wiring, and a plastic top for the chassis.

Distortion would be expected to be higher than in
the more expensive competition, owing to the lack of a
cascoded second gain stage and the use of a simple,
open-loop, MOSFET-based predriver stage. Measure­ments confirmed this. Into an 8-ohm resistive load, THD
+ N reached a minimum of -86 dB at 40 watts and then

rose with a soft clipping profile to -78 dB at 83 watts. At

the "official" 100-watt maximum output level the distor­tion was -58 dB (0.13%). These figures were almost the
same at all frequencies; dynamic distortion was negligi­ble. Into a 4-ohm resisitive load, minimum distortion was

-83 dB, still at 40 watts, then rose with a soft clipping
profile to -70 dB at 180 watts. Dynamic distortion was a
little more evident into 4 ohms, especially in one chan­nel, where the 20 kHz curve deviated from the 1 kHz
curve by a worst-case margin of 7 dB. (The MOSFET/
bipolar output stage may have a bias-tracking problem
that causes this variation.) The inherent noise floor of the
amplifier was significantly lower than that of some presti­gious high-end units.

The PowerCube system measured a dynamic out­put voltage of 33 V (136 watts) into 8 ohms. That repre­sents a dynamic headroom of approximately 1 dB, as ref­erenced to the steady-state output of 110 watts at 1%

ISSUE NO. 22 • WINTER 1994-95

distortion. The PowerCube looked outstandingly good—
almost ideal—with both resisitive and highly reactive

loads at 8 ohms and 4 ohms, but at 2 ohms and 1 ohm the
output dropped drastically into reactive loads. This indi­cates that the protection circuit is coming on despite its

complexity. Into pure resistance the output was still good

at 2 ohms and 1 ohm, down 15.5% and 41.4% respective­ly. Peak current output was 37 amps.

Given the complexity of the circuit design and the

relatively modest power output of this amplifier, one

would think it had the potential to perform at the Boulder
or Bryston level. The PSE Studio IV doesn't quite live up
to that potential. Do not get me wrong; it performs as
well as other amplifiers we have recommended. Unfortu-

nately, one such amplifier is the Marantz MA500 (see the
Editor's review above), which costs significantly less and
offers slightly more power into an 8-ohm load. For the
additional $397, the American-designed and manufac­tured PSE gives you better build quality and better power
supply rejection, owing to its regulated supplies. For
many this will be enough to justify its cost, and I would
not discourage them from purchasing this unit. But we
might think of the PSE Studio IV as work in progress be­cause, with small modifications to the protection circuit­ry, it would draw an excellent PowerCube, and changing
the output-stage predriver from MOSFETs to bipolars
could probably bring the distortion down by about 10 dB.
The Studio IV would then be able to challenge directly
its more expensive competitors.

Line-Level Preamplifier

Rotel RHA-10

(continued from Issue No. 21)
(Reviewed by Peter Aczel)

Rotel of America, P.O. Box 8, North Reading, MA 01864-0008.

Voice: [phone]. Fax: [phone]. RHA-10 Stereo
Active Controller, $1799.90. Tested sample on loan from manu-

facturer.

This is supposed to be the high-end predecessor of
the modestly priced RC-980BX, and there are some re­semblances but they are generally unflattering to the
RHA-10. Unlike the cheaper unit, the RHA-10 uses dis­crete transistors in the gain block. Complementary bipo­lar differential pairs are biased by single-transistor cur­rent sources and resistively loaded. The second gain
stage is a push-pull degenerated bipolar circuit, which
drives a push-pull emitter follower. The second stage is
not resistively loaded. Referring to the generic active
amplifier stage explained on page 18 of Issue No. 18, the
coupling capacitors C1 and C3 are electrolytics, just as in
the cheaper Rotel, only bigger (25 µF instead of 10 µF).
The more important C2 is not used, again as in the cheap­er unit.

As for the power supply, each channel has its own

29

pdf 24

bridge rectifier and open-loop regulator. The pass transis­tor regulator is of a two-transistor composite design and
is set to the regulated voltage by a zener diode biased by
a current source. Using the same zener reference for both
channels here is a peculiarly chintzy feature in an expen­sive preamp, more understandable in the similarly
configured RC-980BX. The power transformer, on the
other hand, is large for a preamp.

As in the cheaper unit, the output stages are muted
by relays on power-up. Unlike the cheaper unit, the
RHA-10 has its tape monitor outputs buffered by AD712
op-amps. The record selector is correctly designed to pre-

vent self-looping of a tape recorder, which could lead to
destructive amplified oscillations. The input and record
selector switches, as well as the volume control, are
sealed units. The volume control is a motorized pot, of
good quality but not on the Penny & Giles level, control­lable with the remote that also controls the Rotel RHT-10
FM stereo tuner. The PC boards are double-sided, but the
RCA jacks are mounted directly on the board, another
chintzy touch in this price range. The sheet metal is cos­metically more attractive than the cheaper unit's but not
really of superior quality. The screws are soft and too
easily strippable.

Although the basic engineering of the RHA-10 is
quite ordinary and in some details downright parsimoni­ous, careful execution has resulted in extraordinarily high
performance. Would you believe -100 dB THD + N at 7
V out? At any frequeny, including 20 kHz, with hardly a
trace of dynamic distortion? That's the best we've mea­sured so far. The noise floor, however, is not lower than
in other high-quality preamps. Maximum output before
clipping is just below 14 V. A low-impedance (600 ohm)
load reduces maximum output to 12 V and causes some
dynamic distortion above 2.5 V out. Channel separation
is excellent; at unity gain it ranges from 105 dB at the
lower frequencies to 94 dB at 20 kHz. That't partly due

to the dual volume control in lieu of a separate balance

control.

Overall, the Rotel RHA-10 is a damn good unit,

but I just don't see $1800's worth of preamp under the
cover. At $900 or even $1000 I'd be wild about it.

Stereo Power Amplifier

Rotel RHB-10

(continued from Issue No. 21)

(Reviewed by Peter Aczel)

Rotel of America, P.O. Box 8, North Reading, MA 01864-0008.

Voice: [phone]. Fax: [phone]. RHB-10 stereo

power amplifier, $2699.90. Tested sample on loan from manu-

facturer.

After further examination, reflection, tests, and
headscratching, I am not at all enthusiastic about this unit
and don't feel like giving it much space here. Not that it

30

isn't an excellent power amplifier. It is. But it costs 2½
times as much as the Rotel RB-990BX and is virtually
the same amplifier! The circuit is almost exactly the
same, except for an emitter follower after the differential
pair and before the second gain stage. The output-stage
transistors are of a different type. The heat sinks are larg­er but not much. There are two power transformers, not a
single big one, and more full-wave rectifiers and filter
caps—but still no regulated supplies! The PC boards are
still single-sided—hey, at $2700? Protection is better,
with relays in series with the outputs. The sheet metal is

prettier but not greatly superior. I just don't see the

$1600's worth of extras. Yes, the power output is a little

higher (210 watts per channel into 8 ohms, 420 watts into
4 ohms) and the THD + N lower (minimum around -95

dB, with plenty of dynamic distortion at 20 kHz). That's
not enough to light my fire at this price. Sorry.

Passive Control Unit

Rotel RHC-10

(Reviewed by Peter Aczel)

Rotel of America, P.O. Box 8, North Reading, MA 01864-0008.

Voice: [phone]. Fax: [phone]. RHC-10 Passive

Controller, $999.90. Tested sample on loan from manufacturer.

As Gertrude Stein might have said, had she been an
audiophile, "a passive front end is a passive front end is a
passive front end." What's there to review about it? Well,
maybe parts quality and construction. Crosstalk, yes. But
harmonic distortion? Or, especially, the sound? Only the
untutored cultists at the tweako magazines indulge in that
kind of rubbish. A passive front end is a bunch of jacks,
wires, switches, potentiometers, and suchlike. Might as
well talk about the audio performance of the light switch
in your wall. (Of course, the tweaks do.)

All right, the parts quality in the RHC-10 is high.

The construction is beautiful. The crosstalk is remarkably

low. Under absolute worst-case conditions, I measured
better than 60 dB channel separation even at 20 kHz, im-

proving by several orders of magnitude depending on fre-

quency, attenuator setting, and source impedance. You
can say that the unit is transparent.

The Input switch has 5 positions, corresponding to
5 pairs of unbalanced input jacks; the Rec Out switch has

4 positions plus Off. The outputs are Main Out and
Record Out, unbalanced. The continuously variable high­quality attenuator has concentric L and R knobs, adjust-

able separately for balance and friction-locked for volume.

I don't see much point in passive front ends be­cause well-engineered and correctly buffered active units
are equally transparent and more versatile, but if the idea
appeals to you and you think $1000 is chicken feed, this is
a very fine passive control, recommended without reser-

vations. Remember, though, that it has no gain and a
10k output impedance. Use only short interconnects.

THE AUDIO CRITIC

pdf 25

Stereo Power Amplifier

Sunfire (more a preview than a review)

(By Peter Aczel and David Rich)

Sunfire Corporation, P.O. Box 1589, Snohomish, WA 98290.
Sunfire "Load Invariant High Fidelity Stereo Power Ampli-

fier, " $2175.00. Tested sample on loan from manufacturer.

Here is another instance of being able to report an
exciting new development thanks to our habitual lateness.
We were able to have Bob Carver's superduper hot-off­the-production-line Sunfire amplifier for about 24 hours
and run a few tests on it before (finally!) going to press.

First, let's get all the identities straight. Sunfire
Corporation is the new, small, high-end company Bob
Carver formed after he had left Carver Corporation.
(That breakup was widely reported in the business press,
so I won't go into its Byzantine intrigues and details. Bob

is still the largest, though not majority, stockholder of

Carver Corporation, but without a job there.) The Carver

company also has a new amplifier, called Lightstar and in
many ways similar, but not identical, to the Sunfire. It

was started by Bob before he left and finished by others.

We have been promised a sample of the Lightstar, but
this preview is of the Sunfire, i.e., the all-Bob-Carver
execution of the same basic design concept. (Lightstar,

Sunfire... Guess who made up those names?)

The amplifier is very handsome in a sleek, black,
understated way; much more High End in appearance
than any of the older Carver models; beautifully con­structed with high-quality parts—and a joule meter as the
pilot light! It has two outputs per channel, one with very
low output impedance (voltage source) and one with a
huge 1-ohm series resistor (current source). The latter

simulates a typical tube-amplifier output, yielding a soft­ening of the highs and a slight plumping of the lower fre­quencies when feeding a typical speaker load (see Issue
No. 16, page 55). I don't believe that such "processing"
belongs in a power amp; in fact I don't believe in tube
amplifiers with highish output impedances; but of course
you don't have to use that output.

I had time to make only a few measurements and I
want to discuss the results only in a very general way,
until we receive our "official" review sample and can test
it unhurriedly. The Sunfire is a 700-watt-per-channel ste­reo power amplifier. That's a continuous-power measure­ment into 4 ohms, just below clipping, at any frequency.
Is that enough muscle for you? With intermittent 1 kHz
tone bursts as the input, it can deliver over 2000 watts
per channel into 1 ohm! What's more, it doesn't care
whether the load is resistive, capacitive, or inductive. It

just keeps pumping almost constant voltage into any

phase angle. The exception is 1 ohm at -60° and +60°,
where the output voltage drops considerably—but no
loudspeaker presents that kind of load. Into 1 ohm at

-30° and +30° the output voltage is still remarkably high.

ISSUE NO. 22 • WINTER 1994-95

The amplifier is labeled "load-invariant" on the fascia,
and that's a fair statement, at least at the present stage of
our investigations. Our usual THD + N measurements in­clude harmonics and noise up to 80 kHz; in the case of
the Sunfire this test is skewed by the high-frequency
switching noise. Reducing the measurement bandwidth
to 22 kHz yields dramatically better readings, and with
an A-weighting filter there is another huge improvement.
A detailed discussion of the Sunfire's measured perfor­mance will have to wait until we receive a sample we can
keep for a while. Meanwhile, my initial impression is
that this is an amplifier of great originality and unique ca-

pabilities.

—Peter Aczel

To borrow from NPR's "Classical Countdown,"

this amplifier is genuinely interesting. The problem that

Bob Carver is addressing is the power dissipation in a
power amplifier that occurs as the amplifier drives the
loudspeaker. As we explained in Issue No. 20, this power
dissipation is the result of the amplifier's finite conver­sion efficiency. One method to improve conversion
efficiency is to use class D (switching) power amplifier
design. These circuits are very efficient because the un­regulated power rails are often switched off from the
load. The problem with switching amplifiers is that the
switching clock (100 kHz and above) can leak into the
output, violating FCC regulations. In addition, the design
of a low-distortion PWM amplifier can be very difficult,
as has been demonstrated in numerous AES preprints.

What Sunfire does is to use a switching amplifier to
generate the power supplies for a conventional amplifier.
The switching power supplies are set to be ±7 V away
from the output signal level. The amplifier itself is a con­ventional design which can be made to achieve as good
performance as desired using standard techniques.
Switching noise is attenuated by the power-supply rejec­tion of the amplifier. Distortion in the power-supply rails
has almost no effect on the amplifier, even if it is 10%.
So we have almost the efficiency of a class D amplifier
but without the performance problems associated with
that technology.

The extra pair of tracking/switching power supplies
adds a lot of electronics to the amplifier but does not add
to the cost. The reason is that the large chassis and heat
sinks normally required to dissipate the heat of a high­power amplifier designed to drive low-impedance loads
is eliminated. While it may come as a surprise to many of
you, it turns out that sheet metal costs a lot more than
electronics, so in the end Sunfire can produce the amp at
the same cost as a comparable conventional design, or at
even lower cost. What is gained is that with its very high
efficiency the amplifier can drive very low-impedance
loads continuously. Even the largest of the conventional
designs cannot drive loads of 2 ohms and below continu-

(continued on page 49)

31

pdf 26

Catching Up on the

Digital Scene

By Peter Aczel

Editor and Publisher

&

David A. Rich, Ph.D.

Contributing Technical Editor

Some of these reviews were left out of the last issue; some are
new; all are still of current interest.

We were hoping to bring you in this issue a defini­tive article by Steve Norsworthy on the pros and cons of
delta-sigma D/A converters, explaining idle channel
tones, the proper use of dither, etc., but Steve got totally
tied up in projects of higher priority (to him, that is). The
article is coming, though, and when it comes it should
have the same kind of impact as Bob Adams's jitter arti-

cle in the last issue. (The general reaction to the latter

was a sigh of relief that an authoritative, professional, yet
relatively simple explanation of a tweako-cultist-ridden
subject was at last available.)

Compact Disc Player

Denon DCD-2700

(Reviewed by David Rich)

Denon America, Inc., P.O. Box 5370, Parsippany, NJ 07054-

5370. Voice: [phone]. Fax: [phone]. DCD-2700
compact disc player with remote control, $1200.00. Tested
sample on loan from manufacturer.

Editors Note: Since this review was written, the

DCD-2700 has been succeeded by the DCD-3000, at the

same price. No service manual for the newer player, or
even a circuit schematic, was available from Denon as of

press time, only the operating instructions, from which

the differences appear to be relatively unimportant, with­out probable bearing on the major points made by Dr.
Rich. The transport has been moved to the center of the
chassis and is of a slightly different design; digital inputs
have been added, allowing the use of the newer player as
an outboard D/A processor (but would you want to do
that?); balanced analog outputs have been added; the

pitch control feature has been eliminated; the weight of

the player has been reduced by 13%. Since the dollar
buys fewer yen today than a year ago, I don't expect the
newer player to be more expensively made; au contraire.

The centerpiece of the design, which is the Alpha System
processor, has not changed, and that alone makes the re­view that follows still fully relevant.

* * *

This CD player is for all practical purposes De­non's top of the line. The company does produce a five­figure, two-piece transport/processor setup, but it is sold
in very limited distribution. The DCD-2700 is loaded
with all the standard features found at this price point, in­cluding a motorized volume control. The most unusual
feature from the user's point of view is the pitch control
function. [Eliminated in the DCD-3000, as noted
above.—Ed.] A pair of buttons on the unit are pressed to
raise or lower the playing speed. The unit also has a peak
search feature for setting the tape-deck record level. Peak
search scans the CD for the maximum signal level on the
disc and plays a few seconds on either side of this point.
Going to a specific index point is more difficult on this
unit than most. (How strange, since Denon CDs use in­dex points more than anybody else. Denon indexes all
important subsections of a composition). You have to
press the index button and then punch in the track and in­dex numbers from the keypad.

The most surprising thing about the operation of
this CD player is how long it takes to access tracks. It
takes twice as long as one would expect for a unit in this
price class with a linear drive transport. Indeed, access
time is comparable to that of the Marantz CD-63 and
Harman Kardon HD7725 CD players, which use the El
Cheapo gear-driven laser mechanism. To find out what's
going on we have to remove the cover. The first thing
you see is a structural reinforcement bar that runs from
the front to the back of the unit. It is held in place by fric­tion only. Remove it and the reason for the player's slow
tracking becomes apparent. The laser transport mecha-

nism is a low-end gear-driven unit, just like those in the

Harman Kardon and Marantz players!! This type of cost

ISSUE NO. 22 • WINTER 1994-95

33

pdf 27

cutting is not what the purchaser of a $1200 unit should

expect. The only interesting thing about the mechanism

is that the laser pickup (Sony KSS-240A) has the RF
amp built in. This makes the laser pickup less susceptible

to electrostatic damage. The servo circuit uses an analog
technology instead of the newer (and probably cheaper to
manufacture but not necessarily better) DSP-based servo
circuit used in the Sony and Marantz CD players. [The
DCD-3000 has a digital tracking servo system.—Ed]

The transport is not the only sign of cost cutting.
The PC board is a single-sided affair with lots of jump­ers. A single smallish power transformer is used. It is
similar in size to the one in the Marantz CD-63. Two
full-wave rectifiers are used in the power supply, one for
the digital circuits and one for the analog. Unregulated
filter capacitors are only 2200 µF for the analog side (by

comparison, the Harman Kardon HD7725 uses 4700 µF
caps on the analog side). The rated working voltage for
all the electrolytics in the analog supply, including these
capacitors, is a surprisingly high 50 V (all is not cost
cutting on this unit). A single regulator (7805) and two
overcurrent protection devices are all the supply regula­tion that's on the digital side. On the analog side a pair of
regulators, 7812 and 7912 (±12 V supplies for $1200—
how come Marantz does it for $399?), are all that's used.
The regulators supply both channels. All regulators are
mounted on the chassis for dissipation of heat. The
DACs are subregulated with 78L05 and 79L05 regulators
(the L is for low current—translated: low cost). Separate
regulators are used for the left and right channels on the

DAC. Spectral plots of the player's line output showed

significant hum components of 60 Hz and 180 Hz, peak­ing at approximately -115 dB relative to full scale. CD
players in this price range usually do better, given their
more robust power supplies. On the other hand, channel

separation was excellent, 120 dB at 1 kHz and 100 dB at
20 kHz, so clearly signals are not crosstalking through
the power supply.

The digital filter chip used in this unit is the NPC
SM5845AF. Here Denon spares no expense. Many high­end manufacturers of five-figure units claim the much
less expensive SM5813 is just perfect, but NPC apparent­ly disagrees since they produced this more complex chip.
Denon states that the SM5845AF is essentially the
SM5842AP with an added feature discussed below. The
SM5842AP has a data path which is very wide at 24 bits;
coefficient word length is 26 bits, and the accumulator is
32 bits wide. The number of taps in the FIR filter was not
specified. Stopband rejection is excellent at -117 dB, ex­ceeding even Sony's current best effort, the CXD2567.
The passband ripple of the SM5842 is ±0.00002 dB (here
Sony wins by 0.00001 dB). The chip has a built-in dither
generator to decorrelate the signal from the truncation
error that occurs when the accumulator output is truncat­ed to 20 bits. It also has digital de-emphasis. Strangely,
we measured a peak of 0.08 dB at 4kHz when measuring

34

the DCD-2700's frequency response in the de-emphasis
mode. Even stranger are some 0.02 dB ripples in the
passband frequency response of the DCD-2700 from 2

kHz to 13 kHz. These measurements make me wonder if

something has gone wrong in the evolution of the

SM5842AP into the SM5845AF.

The SM5845AF adds digital circuitry which at­tempts to estimate the value of a quantized data point at a
resolution below the quantization level. This can only be
accomplished by making some estimate of the type of
signal that was digitized (because Dr. Shannon has a lit­tle theorem forming the basis of information theory that
says you cannot take a quantized PCM signal at the Ny­quist rate and extract any more information from it).
When you play undithered low-level signals through the
Denon, the outputs are much more continuous in appear­ance, and distortion of the signal is lower, than the theo-

retical minimum for an undithered signal. But a quan­tized sine wave should have distortion. When dither is
not used, the quantization error is correlated with the sig­nal and this causes the distortion. It is for this reason

that all recordings are done with dither before the A/D.

Dither added at the recording site decorrelates the quanti­zation error, and the distortion is eliminated. Noise, not
distortion, occurs instead. Dither also allows us to record
signals at levels below an LSB. The Denon digital filter
produces a low-distortion undithered sine wave, but how
can it improve things with dithered music signals? To do
this it would have to reduce the noise level of the signal.
Since the noise is decorrelated with the signal and the
filter chip has no way of estimating the statistics of the
dither or the music, I do not see how this can work. It
looks to me as if the whole thing were designed to make
undithered sine waves look good at hi-fi-show demos,
but Denon's information on this technique—they call it
Alpha System Processing—is very sketchy, and a small

possibility exists that the Denon system could be of some

value. Audio Alchemy is claiming to have a device
which performs a similar function, so one must be very

careful not to dismiss all this without more information.

The DACs are of the Burr-Brown PCM1702J type.
The PCM 1702 is a device which is a redesigned PCM63
in a BiCMOS process. It is a smaller die that fits into a

16-pin DIP package or an even smaller surface-mount

package. The pins eliminated include the pins for pot ad-

justments, which were virtually impossible to adjust cor-

rectly. The spec sheets for both devices are similar, but
the THD for the K grade of the smaller type is 2 dB
worse at -20dB and dynamic range is no longer given a
worst-case spec. Minimum idle-channel SNR is 6 dB
worse at a still excellent 110 dB, and the PSRR spec is
not given. Worst-case gain linearity is not specified and
the typical number has changed from ±0.3 dB to ±0.5 dB
at -90 dB. On the DCD-2700 we measured a gain linear­ity error of +0.25 dB at an input signal level of -80 dB.
This then rose to +0.75 dB at -90 dB and +1.25 dB at

THE AUDIO CRITIC

pdf 28

-100 dB. The CBS test disc monotonicity test looked ex­cellent.

Two DACs are used per channel in a balanced
configuration. The balanced signals come from the
SM5845AF digital filter (a feature not found on the
SM5842AP). This approach is less effective than when
used with 1-bit coders because the 1-bit coders' outputs
are highly repeatable from chip to chip since they are al­most totally digital. The analog R-2R ladder-based DAC
chips should not exhibit significantly correlated nonline­arities (see Issue No. 15). I would rather have a single

pair of K-grade DACs.

The ICs in the analog path are of the NEC
µPC4570 type, one chip per DAC (4 total). The
µPC4570 is also used for the balanced-to-single-ended
converter and the buffer for the GIC filter that drives the

output terminals. Although the µPC4570 is a dual device,

only a single op-amp is used; the other op-amp is not

used (don't look at me, I don't know why). The NEC

µPC4570 is said to be an NE5532 equivalent in the NEC

data sheet and the specifications are similar, but the parts
are not the same. The principal specification difference
relates to the output stage: the NE5532 is designed to
drive 600 ohms; the µPC4570 can drive only 2000 ohms.
The µPC4570 uses a simple 2-gain-stage topology; the
NE5532 uses a more complex 3-gain-stage topology.
The simpler circuit and smaller output stage make possi­ble a smaller die and a lower-cost part. For $1200 I
would expect a state-of-the-art op-amp such as the
AD797, but instead you get a cost-reduced NE5532.

But the worst is yet to come.

The generalized impedance converter uses an
NJN4558 from the New Japan Radio Company (JRC).
From the data sheet: "...combining the features of the
NJM741 with the close parameter matching and tracking
of a dual device." Yes, folks, this is a dual version of the
25-year-old 741!! Slew rate, for those of you who forgot
(normally this thing is not used in audio even in the
cheapest of the cheap because it is not fast enough), is 1
V/µs and the gain-bandwidth product is 2 MHz. What
happens when you put a slow op-amp in a GIC filter with
a 30 kHz passband edge frequency? You get Q enhance­ment that results in the peaking of the filter's response.

[Sedra and Brackett. Filter Theory and Design: Active
and Passive, Chapter 9.8.] That's just what we measured.
The frequency response of the unit stays in a ±0.05 dB
window (those strange ripples discussed above) until 13
kHz, then it rises up to +0.35 dB at 20 kHz. I ran a com­puter simulation of the circuit, using a macro model of
the NJN4558, just to make sure that the op-amp was the
cause of the peaking. The simulation results correspond­ed closely to the measured results, but substituting an
ideal op-amp in the simulation reduced the peaking by
only 0.2 dB. It turns out that the remaining peaking is
caused by the use of 5% carbon resistors throughout the

analog signal path, instead of 1% metal film resistors that

ISSUE NO. 22 • WINTER 1994-95

we expect to find in a four-figure-priced unit. Since 5%
resistors are available in a much more limited number of
resistor values, the exact values required for the filter are
unavailable, and the values actually used cause the re­maining peaking. All I can say is, the person who de­signed this stage must have come straight from the de­sign of a $100 portable CD player, and he/she/it (I hate
this PC stuff) did not change his/her/its design methodol­ogy for this $1200 unit.

Full-scale THD + N was substandard at high fre­quencies. It measured -95 dB at 500 Hz and then it start­ed to rise. Measuring -90 dB at 3 kHz, it rose to a maxi­mum of -63dB (!!) at 13 kHz. Dropping the input level
to -24 dB resulted in distortion readings of 1 to 2 dB
above theoretical limits. This indicates the distortion is
coming form the analog stages and is not unexpected,
given the analysis of the analog stage above.

The dc blocking capacitor at the output of the ana­log stage consists of a series-parallel quad of electrolyt­ics. Each series combination has the plus terminal of the
capacitors tied together. This is an attempt to mimic a
nonpolar capacitor. One set of capacitors is 10 µF, the
other is 470 µF. One sign of quality design is the muting
circuit, which uses a relay that shorts the output to
ground.

Summing it up:

This unit is overpriced, given the parts inside.
Some interesting aspects of the design are the balanced
DAC and the digital filter, but the analog section is a
poor performer (if I were Corey Greenberg, I would
have a more colorful way of putting this). Denon has pro­duced products that performed well and represented good
value in the past. I hope the DCD-2700 is an aberration.

Outboard D/A Converter with Transport

DPA (Deltec Precision Audio)
PDM 2 and Tl

(Reviewed by Peter Aczel)

Audiophile Imports, 2012-B Main Street, Cross Plains, WI

53528. Voice: [phone]. Fax: [phone]. DPA
(Deltec Precision Audio) PDM 2 Bitstream D/A converter,
$4395.00; T1 transport with remote control, $1595.00. Tested
samples on loan from distributor.

This U.K. company, which sometimes signs itself
as "dpa digital ltd" and has its headquarters in Wales,
makes ultrahigh-end digital processors. The costly equip­ment reviewed here is not even their top-of-the-line at
this point, although it was at the time we tested it and is
still being sold at the above prices, which are—believe it
or not—reduced from the original. The designer of the
circuits and managing director of the company is Robert
Watts.

The juxtaposition of the PDM 2 D/A converter to

35

pdf 29

the Tl transport is a somewhat schizoid one because the
two-chassis PDM 2 is exceedingly high-tech, whereas
the Tl is definitely not. To dispose of the Tl first, it's a
cheap plastic Philips player with what David Rich called
a "Hollywood false front"—namely a machined alumi­num front panel and a metal outer skin with rounded edg-

es and a high-gloss black finish. (The two PDM 2 units

are similarly packaged, only much smaller.) The trans­port doesn't have the all the sophisticated disc-control fa­cilities one expects these days in a high-priced player,
but the interface with the PDM 2 is as sophisticated as it
gets. DPA actually had to chop off some of the original
Philips plastic to make room for the interface board and
its fair-sized transformer.

The rest of the system is incredibly advanced in
some respects and old hat in others. The analog and digi­tal sections of the PDM 2 are in two entirely separate
boxes. The Philips DAC 7 (TDA1547) Bitstream chip is
used as the D/A converter, in conjunction with the Phil­ips SAA7350 Bitstream PDM device, but the implemen­tation is much more elaborate than in any other processor
we have tested. Most of the components are surface­mounted, a technology almost never seen in low-volume

high-end audio equipment. It eliminates the leads, gets
rid of parasitic capacitance, allows a large number of
components in a small circuit-board area, and saves a tre­mendous amount of space. Only wave-soldered, double-

sided PC boards are used, and they're really packed.

Each channel has a custom hybrid circuit in the signal
path, another ultrahigh-tech feature hardly ever seen in
consumer electronics. (A hybrid circuit permits the mix­ing of discrete and IC technologies of every type, as well
as great flexibility in resistor and capacitor values, toler­ances, etc.) On the other hand, the S/PDIF decoder is the
mundane, old-tech Yamaha YM3623B, with no signs of
any auxiliary circuitry, and the digital filter is the similar­ly banal Yamaha YM3414. On the plus side again, the
switches are high-quality sealed units, the muting and de­emphasis switching is by relays (not solid-state switch­es), and both the analog and digital power supplies
appear to be quite elaborate and well-regulated.

A multiplicity of fiber-optic cables is used to con­nect the digital box to the analog box. These Toslink con­nections carry very high-speed signals transmitted at a
high oversampled rate, too fast for a single plastic optical

cable to transmit. This is basically an isolation technique,
designed to prevent intermodulation of the digital and an­alog signals. An additional fiber-optic cable sends the
DAC clock back to the CD transport, thereby assuring
clock synchronicity. DPA calls this the "Deltran Sync­Lock," although Sony used the same jitter-elimination
technique a good many years ago in their high-end CDP­Rl and DAS-Rl—but that doesn't make it a less good
solution.

On the laboratory bench the DPA system yielded

mixed results. Full-scale THD + N was unimpressive; at

36

the lower frequencies it hovered in the vicinity of -85
dB, rising to a maximum of-82 dB at 7 kHz. These were
the figures with an S/PDIF input; with a CD input from
the Tl they were worse by 2 to 4 dB. Reducing the input
signal level in order to get rid of gain-related analog dis-

tortion still did not improve the results sufficiently; the
best obtainable figure as normalized to full scale was -95
dB, still 3 dB short of the theoretical 16-bit minimum,
but at least the worsening at the higher frequencies was
gone. The DPA literature shows an FFT plot of a dith­ered 1 kHz tone at -60 dB to prove that the noise floor is
an astonishing -130 dB; I was able to duplicate this, but
the plot is deceptive—the distortion/noise is shown bin
by bin, not as an rms total. Other processors I have mea­sured can produce an equally good plot under the same
conditions. The noise floor as measured conventionally
with digital zero input is actually not so great; there are

power-supply bumps at 60 Hz and 180 Hz hovering

around the -100 dB level, and the noise reaches -97 dB
at 60 to 90 kHz. Repeating the FFT test at full scale or

even -10 dB with a dithered 7 kHz tone showed some

very peculiar noise-floor modulation effects; Steven

Norsworthy and David Rich are still debating the cause.
It can't possibly be a good thing. On the other hand, gain
linearity was of the utmost perfection, with zero error all
the way down to -100 dB (that's delta-sigma for you).
The full-scale frequency response of the PDM 2 drooped
to almost -1 dB at 20 kHz; I wonder if that's really nec­essary. At 10 kHz the response was -0.25 dB. A very
slight audible softening of the highs is therefore a possi­bility. I didn't agonize over it when I listened; the sound
was excellent, needless to say.

Speaking of Norsworthy and Rich, their recent
AES paper on idle channel tones in delta-sigma convert­ers involved proprietary test techniques that were tried on
the PDM 2 among many other processors, and it appears
that the complete separation of the analog and digital
electronics pays off handsomely in this particular respect.
The PDM 2 produced significantly smaller amounts of
idle channel tones than other processors using the same
Philips chip set. A detailed explanation of how and why
this happens will have to wait until the groundwork is
laid by our scheduled introductory article on the subject.

Bottom line: this DPA equipment incorporates
some very advanced technology, but on a black-box basis
a number of other designs perform better. The latest DPA
models, which are just out, may be a different story.

Compact Disc Player

Enlightened Audio Designs
CD-1000

(Reviewed by Peter Aczel)

Enlightened Audio Designs Corp., 300 West Lowe, Fairfield, IA

52556. Voice: [phone]. Fax: [phone]. CD-1000

THE AUDIO CRITIC

pdf 30

"Audiophile CD Player" with remote control, $1599.00. Tested

sample on loan from manufacturer.

I can make short shrift of this one because I have in
effect reviewed it before, though not in this form. The
transport uses the offbeat but well-designed Pioneer
"Stable Platter" upside-down drive mechanism, which I
reviewed as part of the Pioneer Elite PD-75 CD player in
Issue No. 18. The electronics are essentially the same as
in the Enlightened Audio Designs DSP-1000 Series II
outboard D/A processor, which I reviewed in Issue No.
20 and recommended as an excellent choice. (In fact, Da­vid Rich bought one.) I could end my review right there,
leaving you with the impression that this is a highly de­sirable CD player, but there's a little more to it than that.

For one thing, the PD-75 (which was originally in­troduced in 1991 and is no longer in the Pioneer line) had
index buttons on its remote control; the EAD CD-1000
does not. That alone would be sufficient for me, personal­ly, not to buy it even at half the price, but those who have
few or no classical CDs probably won't mind (rock-pop
CDs never—or hardly ever?—have their numbered
tracks further divided into index numbers). I am firmly
opposed to the use of the forward and backward scan but­tons for index hunting; it's a drag (and let's not even talk
about the CBS CD-I standard test disc, which is all in­dexes). Other than that, the transport of the CD-1000 is

just fine and dandy (see my comments in the aforesaid

PD-75 review).

I expected the standard measurements to show pret­ty much the same results as with the EAD DSP-1000 and

I was at least partly right. In the right channel, THD + N

at full scale (0 dB) was 5 dB worse than the theoretical

ideal (-98.08 dB) at nearly all frequencies. That was
clearly due to gain-related analog distortion because,

with the digital signal level reduced from 0 dB to -24

dB, the departure from perfection was only 1 dB, match­ing the performance of the DSP-1000. The left channel,
however, was another story. At full scale, the excess dis-

tortion was 11 dB at the lowest frequencies and never

better than 6 dB at any frequency. Even at -24 dB, there

was still approximately 3 dB excess distortion at all fre-

quencies—much worse than the other channel. I suspect-

ed low-level hum components (60 Hz and multiples), but
spectrum analysis revealed no such bumps, only a gener­ally elevated noise floor, especially between 90 and 500
Hz, about 10 dB higher than in the right channel. David

Rich suspects a noisy op-amp. I'll give EAD the benefit

of the doubt and assume that other samples of the CD­1000 do not have this fault, but QC should have spotted
it and weeded it out before the unit was released.

In all other respects the CD-1000 proved to be
flawless on the lab bench. At full scale, the frequency re­sponse rolled off 0.05 dB at 10 kHz and 0.15 dB at 20
kHz. De-emphasis error was 0.15 dB at 10 kHz and 0.35
dB at 16 kHz, but that included the frequency response

ISSUE NO. 22 • WINTER 1994-95

decline. Channel separation was 86 dB at 16 kHz and im­proved 6 dB per octave downward. Gain linearity error
down to the -100 dB level was virtually zero (remember,
the DAC here is the 20-bit Burr-Brown PCM63P-K, as in
the DSP-1000), but the low-level noise in the left channel
was clearly worse in this test, also. Harmonic distortion
of a dithered 997 Hz tone at -90.31 dB was just about un­detectable above the noise floor.

If it had not been for the low-level noise problem
in the left channel and the lack of index buttons on the re­mote control, I would have considered the EAD CD-

1000 to be a close runner-up to the Sony CDP-X707ES,
at a saving of $401, as my favorite "Audiophile CD
Player" (to use the phrase inscribed on the EAD's front
panel). As it is, I cannot give it that ranking. But, if EAD
came out with a CD-1000 Series II...

Compact Disc Player

Marantz CD-63 and CD-63SE

(Reviewed by David Rich)

Marantz USA, 1150 Feehanville Drive, Mount Prospect, IL

60056. Voice: [phone]. Fax: [phone]. Model
CD-63 compact disc player with remote control, $399.00. Mod­el CD-63SE compact disc player with remote control, $499.00.
Tested samples on loan from manufacturer.

Editor's Note: Although this review was written in

the spring of 1994, the CD-63 is still in the Marantz line
and will remain therefor some time. Meanwhile, Marantz
has also added the CD-63 Special Edition to the line. The
SE version is listed at $100 more and is basically a mar­keting idea, designed to attract those who like the idea of

"tweaked" versions of standard components. At least the
tweaking here is done by the designers of the original
unit, not by some third-party cultist operating from his
basement. The "improvements" include a much heavier

bottom plate for the chassis (the SE weighs 37% more
than the plain CD-63), gold-plated jacks, a number of

fancier capacitors, a higher-grade power transformer,

additional shielding, etc. All very nice, but I could mea­sure no improvements of any kind in the actual output of
the SE. The two models are indistinguishable in perfor­mance—and that includes the small power-supply bumps
rising from the noise floor. Thus every word of Dr. Rich's

review is applicable to both models.

* * *

The marketing guys at Denon and Harman Kardon,
having read my reviews of their CD players, are no
doubt saying something like this: "That Dr. Rich is some
ivory-tower fruitcake. If he knew anything about the cost
of producing a CD player, he would understand that you
can't produce a CD player the way he wants to at our
price point." To these marketing people I answer: Ma­rantz CD-63. Now do not get me wrong, the Marantz
CD-63 is not a Sony CDP-X707ES for one fifth the

41

pdf 31

price—far from it. But its parts/price ratio is much better
than we have seen in the Denon and Harman Kardon
units. You even get a coax digital output in addition to
the Toslink. This is very rare at this price point. The unit
certainly does not look like it's from the bargain base-

ment, although the disc-search switches feel pretty

clunky. Only one important feature is missing but it's a

big one. The unit does not display indexes and it has no
provision to go to an index point on the disc. [The SE
version does not remedy this serious shortcoming.—Ed.]

The unit has an all-metal cabinet. Plastic is very
common at this price point. The power transformer is
about the same size as that of the Denon DCD-2700. The
CD-63 has the same number of regulators (4) and the
same analog supply rail (±12 V). Such large supply rails
are very rare at this price point. To be fair to the Denon,
it has larger filter capacitors and additional DAC subreg­ulators. But the Denon has a single-sided PC board,
whereas the Marantz uses a double-sided board so a
ground plane can be added. (There are no plated-through
holes on the Marantz board, so it is still full of jumpers.)

A digital signal-processor chip is used for the servo
functions (TDA1301). This is also the way the Sony
CDP-X707ES does it, but they have developed their own
chip (CXD2501Q). The actual transport is almost literal­ly a toy in comparison with the Sony's. It's almost all
plastic. According to the Philips literature, it's "designed
both for stationary use and for application in CD-Radio

Cassette Recorders....The low mass actuator is particu-

larly effective in low power and portable applications."
Low-profile design is again aimed at the portable market.
A cheap-looking rubber suspension looks as if it were not
going to work as well as a spring-loaded floating metal
subframe, such as used by Denon, Harman Kardon, and
Sony, but an additional plastic subassembly surrounding
the transport and apparently filled with a viscous sub­stance provides additional isolation. The build quality of
the transport and chassis is what justifies the Sony CDP­X707ES to be five times the price. The Sony would have
to last five times as long to recover the cost difference be­tween the two players. I do not think it is going to last
that long, but there is no question that you get what you

pay for in the Sony. The larger transports in the Denon

DCD-2700 and Harman Kardon HD7725, with their
bigger laser assemblies and metal subframes, look as if
they might be a little more reliable than the Marantz's
Philips transport, but they use a lot of plastic in places
most likely to go wrong (including the plastic gear as­semblies that move the laser), so on balance I would guess

the reliability to be similar to that of the Philips unit.

Recall that Marantz is a Philips subsidiary. Now

here is a surprise: Marantz does not use a Philips DAC;

instead they us an NPC SM5872BS. Readers may recall

that I found the Philips low-end DACs to be substandard

(Issue No. 15); apparently the Marantz engineers agree
with me. [See the explanation by David Birch-Jones of

42

Marantz in this issue's "Box 978." Marantz engineering
is basically all-Japanese and quite separate from Philips
engineering, although many Philips parts are used.—
Ed.] The SM5872BS is similar in many respects to other

MASH-type units; the oversampling rate is 32x, and the
resolution of the DAC is 18 bits. The principal difference
is that it has a 4th-order single-stage noise shaper with an

11-level multibit output. The multibit output goes into a
pulse-width modulator as in any other MASH-like part.
Note that the MASH acronym describes the architecture
of the digital delta-sigma modulator (MultistAge noise
SHaping). In a MASH system, multiple lower-order
modulators are interconnected to form a higher-order
modulator. This improves the stability of the modulator
over implementations in a single higher-order loop. The
direct design of higher-order loops is now much better
understood than it was when MASH chips were first in­troduced, and it appears that NPC has successfully over­come the design problems associated with a single-stage

4th-order modulator. I am not going to get involved with

the tradeoffs in the two approaches here, but the key

point is that it is not correct to call the SM5872BS a

MASH device because it has a single-stage architecture.

The key difference between this NPC unit and the
Philips Bitstream device is that in the NPC the digital
noise shaper generates a multibit code, which goes to a

PWM DAC, which then gets filtered by off-chip passive
RC and then active filters. (The Sony CXD2562 "pulse"
DAC does the same thing.) The Philips DAC generates a

higher-speed single-bit code which goes to an on-chip

switched-capacitor filter. The result is that the Philips
part has more noise, more distortion, and (as will be ex­plained in a future article) the potential to generate idle
channel tones. The gain linearity performance of the CD­63 is very good, the error increasing from -0.2 dB for a

-80 dB input signal to -0.7 dB for a -90 dB signal and
then remaining around -1 dB down to the -110 dB level.
THD + N with a -24 dB input signal (to remove the ef­fects of the analog-stage distortion) was 4 dB away from
theoretical levels at all frequencies. This could be the re­sult of noise at the output of the DAC or in the analog
stage, or it could be the result of clock jitter (see the Bob
Adams article in the last issue). As is typical for DACs
that use oversampling, the time-domain monotonicity test
was significantly nosier-looking than what we get with a
good multibit DAC, such as the one in the Denon DCD-

2700.
The SM5872BS also has the interpolating 8x digi-

tal filter built into the same die as the DAC. (A digital
filter is also built into the SAA7345 CD decoder chip of
the CD-63, but this digital filter is not used.) Specs from
the data sheet of the SM5872BS on the digital filter are
the same as for the SM5840D filter used in the Harman
Kardon HD7725 CD player—only 87 taps (other digital
filters have 200 to 300 taps), 19-bit data path, and a very
small 14-bit coefficient word length (other filters are in

THE AUDIO CRITIC

pdf 32

the range of 19 to 28 bits). Ripple is ±0.125dB (4 orders
of magnitude worse than the other filters), and stopband
attenuation is only -55 dB (3 orders of magnitude worse
than other filters). The result of the filter's presence can
be seen on the frequency response plots of the CD-63 in
the form of small ripples in the range of 2 kHz to 15 kHz.
On the $849 Harmon Kardon I found the use of such a
simple filter unacceptable. On this $399 CD player it
looks like a good compromise design solution. An IIR

digital filter is also included on the SM5872BS for the

de-emphasis filter. The de-emphasis is not very accurate;

the error at 4 kHz is -0.7 dB, and not less than ±0.2 dB at

any frequency. The SM5872BS also has a digital attenu­ator, which the CD-63 uses to vary the analog output as
well as the headphone output between 0 and -30 dB.

The Philips low-end DACs supplied only a single­ended output which resulted in even more distortion.
The NPC SM5872BS is balanced and requires an exter-

nal differential-to-single-ended converter. The differen­tial-to-single-ended conversion and the first active pole
of filtering is performed by an NJM2114D dual op-amp.
(I have no information on what is probably a 5532
clone.) The signal then goes to the other half of the
NJM2114D. This time the op-amp is configured as an in­verting second-order filter. The advantage over a Sallen­and-Key filter is that the input stage of the op-amp sees
no common-mode signal. The disadvantage is that it has
a much lower input impedance.

The filter is then followed by an 8-transistor dis­crete op-amp in a unity-gain configuration. The whole
circuit is encapsulated in a metal can just like the mega­buck stuff. The input stage is a JFET differential pair; the
output stage is a self-biased JFET follower. The differen­tial pair is current-source biased (Krell did not do this in
their Studio!). The two JFETs are a special low-noise de­vice mounted in one package. The op-amp is configured
with one gain stage, called a folded cascode. This struc-

Large Loudspeaker System with Subwoofers

Snell Acoustics Type A

(last-minute mini preview by the Editor)

Snell Acoustics, Inc., 143 Essex Street, Haverhill, MA 01832.

Voice: [phone]. Fax: [phone]. Type A Music'
Reference System, $18,999.00. Tested samples on load from
manufacturer.

This new flagship of Snell Acoustics was definitely
not scheduled to be reviewed in this issue; it came in
when all the other reviews were already set in type. If we
hadn't been late (yes, as usual), the issue would have
been printed and delivered, with no chance for these
comments.

ISSUE NO. 22 • WINTER 1994-95

ture has fast settling characteristics and wide bandwidth.
It has somewhat less gain than a two-gain-stage amplifier
but it has more than enough for a buffer. Overall, this to­pology looks as good as the best. (Yes, I did say the
whole thing sold for $399.) Now if they would just use
the same topology in the two pervious stages—because
the full-scale distortion performance is not so hot with
the NJM2114D's in the signal path. The best distortion

performance was -90 dB (8 dB off the theoretical lev-

els). Above 2 kHz it started to rise, reaching -81 dB at 10
kHz. To be fair we must point out that this beats both the

Denon DCD-2700 and Harman Kardon HD7725 (but the

Sony CDP-X707ES's performance crushes the Marantz).
The full-scale distortion of the CD-63 did not change at
all when the output was loaded down with 600 ohms.

Output capacitors are back-to-back (nonpolar) 220
µF electrolytics. Muting is by a pi network of bipolar
switches, not relays (you can't have everything for
$399). Each channel is driven by a separate control line

from the digital filter. It is possible in this arrangement to

shut down one channel in a channel separation test and
keep the other alive. You would get very good channel­separation test results this way (we measured a phony

133 dB at 16 kHz) but you pay for the extra components

to pull this "Julian-should-be-very-happy" trick.

The Marantz CD-63 presents readers of The Au-
dio Critic with an interesting dilemma. It performs well
enough to ensure that the audible signal at its output ter­minal is indistinguishable from that of even the best­performing CD player. It is built well enough so that it
should last several years under steady use. If you spend
more, you can get a better-measuring unit, a more relia­ble unit, a unit with more sophisticated technology, a unit
with better look-and-feel. The Sony CDP-X707ES is the
paradigm of such a unit. Is such a unit worth the five
times greater cost? You must answer that question for
yourself. •

Under the circumstances, I am still able to squeeze

in the following statement:

Of all the loudspeaker systems I have listened to in
my studio/laboratory, at my leisure, this is the one I like
best so far. That doesn't mean it's perfect—and I haven't
run any tests on it yet—but it appears to be more com­plete from the bottommost to the topmost frequencies,
more neutral, and more seamlessly integrated than any
other I am aware of. Of course, you could say that at
nineteen thousand dollars it had better be good, but in
high-end audio a lot of expensive equipment turns out to
be disappointing. This isn't one of those. (Then I must
hasten to add: unless you expect a miracle. It's still just a
loudspeaker.)

More, lots more, in Issue No. 23.

43

pdf 33

44 THE AUDIO CRITIC

In Your Ear

pdf 34

Large-Screen TV for Home Theater:

Is a 40" Direct-View Tube Big Enough?

By Peter Aczel

Editor and Publisher

The top-of-the-line Mitsubishi direct-view set is good enough to
make the choice between formats more difficult.

Our policy is to have at least one high-end video

review in each issue for the home-theater contingent, but

we ran out of space in the last issue and ended up with
nothing on the subject. We don't want that to happen
again because home theater has become the fastest-

growing sector of home-entertainment electronics,

whereas perfectionist two-channel stereo is on the wane
everywhere except in Asia. At this point, we who are in­terested in the latest and greatest power amplifier circuit
and such represent the nerdy rear guard of audio. We
have no intention to stray significantly from those reac­tionary interests but we must keep up with the times and
pay some attention to video, lest we should be perceived
as having the priorities of the Yokohama Triode Society.

40" Direct-View Color TV

Mitsubishi CS-40601

Mitsubishi Electronics America, Inc., 5665/5757 Plaza Drive,
P.O. Box 6007, Cypress, CA 90630-0007. Voice: (714) 220-

2500. Fax: [phone]. CS-40601 color TV set, $4999.00
(best street price, $3050.00). Tested sample on loan from manu-

facturer.

To be perfectly accurate, the model I tested was the
CS-40FX1, immediate predecessor of the more recent
CS-40601. The two models are identical, the only differ­ence being the inclusion of a closed-caption decoder in
the CS-40601, as required by the FCC since July 1993.
As there are no performance-related considerations in
closed captioning, I am taking the very small liberty of
running this review under the newer model number, just
to be more up-to-date. (As we go to press, the very latest
advertised model is the CS-40503; it appears to be pretty
much the identical set. TV model designations tend to
change several times a year.)

The 40" direct-view tube used in this Mitsubishi
design is the largest obtainable in a commercially distrib­uted TV set, at least to my knowledge. When you realize
that rear-projection TVs also come in screen sizes from
40" up (or at least they used to, before 45" became the
more common minimum), you begin to see just how
huge this tube is. It goes without saying that the picture is

ISSUE NO. 22 • WINTER 1994-95

sharper and brighter than with typical rear-projection
screens—no contest. A couple of issues ago I wrote that

"big is more important to me than sharp" but here the
question is no longer big vs. sharp but big-enough-and­sharp vs. bigger-but-less-sharp. I still like 55" screens,
but a 40" picture isn't exactly small, and the advantages
of the direct-view tube begin to offset the size difference.

The sharpness advantage could be deemed illusory,
since the number of scanning lines is independent of the
screen size, but the closer spacing of the lines on the
smaller screen always creates the impression of higher

definition, so that the 40" picture appears sharper than the

55" picture. The brightness advantage, however, is very

real—and huge. I have never seen a rear-projection set
that wasn't much more enjoyable to watch in a darkened
room, whereas it hardly makes a difference with the 40"
direct-view Mitsubishi. The net result of all that is the
perception of a "better" picture—better but smaller. If you
need to see a life-size linebacker blitzing the quarterback,
then you'll opt for the rear-projection screen. I'm not so
sure. I'm waiting for a 55" direct-view tube (fat chance).

One disadvantage of the big Mitsubishi is that it's

extremely heavy and, unlike typical rear-projections sets,
doesn't come in the kind of cabinet that can be equipped
with casters. Three of us out-of-shape weaklings were
barely able to take it out of its factory carton and place it
on the matching MB-40FX base. (That's not included in
the price, but you don't absolutely need it; you can use a
suitable table or anything else.) Once installed, the set is
very easy and convenient to use.

As usual, I ran all the standard tests available on

the Joe Kane laser videodisc A Video Standard (Refer­ence Recordings, 1989). Black level retention was not as
good as on a perfectly calibrated studio monitor, meaning
that the ability to hold black at black, regardless of the
picture content, was satisfactory but not perfect. Contrast
could be easily turned up to the point where the peak lin­ear capability of the set was exceeded, but with a little bit
of tickling a reasonably linear and highly pleasing con-

trast level was obtainable. Color performance via the S­video input was excellent; the best settings of Color and

Tint according to the test disc resulted in rich, brilliant,

highly pleasing and accurate colors on the screen. I

45

pdf 35

would rate the 40" Mitsubishi higher on color quality
than any rear-projection set I've seen. Nearly all the de­fault settings needed to be trimmed, however, for laser­disc perfection. On the other hand, the default settings
represented a better compromise on middling TV broad­casts, so I can't really fault Mitsubishi on this.

Geometry was equally excellent; I observed no dis-

tortion of circles, checkerboard patterns, etc. I was also
pleased to see that the picture wasn't overscanned. Con­vergence appeared to be right on the money. As for hori­zontal resolution, my test setup (as I've stated in the past)
is good only up to about 400 or at best 425 lines, and that
was no problem for the Mitsubishi. (The theoretical best
for NTSC broadcasts is 336 lines.) Happily, I saw no
fictitious ultrahigh-resolution specs for this set.

The two little outboard speakers provided with the
set must be affixed by the user to the hinges on either
side of the screen and wired to the back of the chassis.

The hinges permit only a limited range of aiming; the
speaker design is minimal, consisting of a tiny elliptical

woofer and cone tweeter in a vented plastic pod. The ste-

reo sound is surprisingly clean through these units when
driven by the set's very low-powered internal amplifiers,
but that's not what you want for home theater. The dinky
little two-way speakers built into the MB-40FX base
(they have floor-level dome tweeters!) are only a little
more advanced; I wasn't particularly interested in using
them. Instead, I hooked up the set to a complete Marantz
AV surround system terminating in the Atlantic Technol­ogy System 250 six-piece speaker array. That way the
sound was bigger and cleaner than in your average neigh­borhood movie theater, but I'm saving those reviews for

the next issue (except for the Marantz MA500 power

amp, which is featured in this issue). The 40" picture
proved to be big enough not be overwhelmed by the for­midable surround sound. There are no surround-sound
processor or control facilities built into the set itself; it is
assumed that the home-theater user will have the needed
AV equipment. There is, however, a separate output for a
powered mono subwoofer.

The control features and user conveniences of the
Mitsubishi are also outstanding. The microprocessor­controlled audio/video adjustments are about as complete
as I've seen on any set. The remote control, although far
from an ergonomic masterpiece in its layout, is extremely
versatile and effective. The menus that can be brought up

on the screen for permanent settings (clock, timer, nam-

ing channels, locking out channels, etc.—the list is end-

less) are very clear and easy to use. So is the owner's
manual, by the way. A single button will display on the
screen the status of most of the current settings. The PIP
(picture-in-picture) feature would be even in more attrac­tive if the set had a second tuner (so you could check out
a second program without an outside source, such as a
VCR). All in all, the set can be said to be nicely loaded,
as Chrysler would put it.

Cosmetically the Mitsubishi is in the high-tech

black-plastic-bubble idiom, which is not exactly an inte-

rior decorator's delight, but the front of the set is all tube

and virtually no plastic, so it will fit into a custom cabinet

quite readily. Electrostatically attracted dust is especially

visible on the black plastic surfaces. Bottom line: this is
the only game in town when it comes to a direct-view TV
of serious home-theater size—take or leave it. I'll take it.

Tech. Ed. of The Audio Critic to

Chair a Tutorial Session at a

Major Professional Conference

Dr. David A. Rich, Contributing Technical Editor of The Audio Critic, has been
invited to be chairman of the advanced tutorial session on audio at the coming

DSPx '95 conference (May 15-18, 1995, San Jose Convention Center, San Jose,

CA). DSPx is an exhibition and engineering conference for DSP and other
electronic disciplines. David Rich's session (May 17, 1:00 P.M.) will last two
hours and address the question: "Are there any design considerations in audio
that go beyond standard measurements?" Joining him as speakers on the panel
will be Richard Marsh of C.U.T.L., Inc., Sean Olive of Harmon International, Scott

Wurcer of Analog Devices, Inc., and possibly one other technologist.

For additional information contact Norien Hutchinson, Director, Industry Development, Reed Exhibi­tion Companies, 383 Main Avenue, Norwalk, CT 06851. Voice: [phone]. Fax: [phone].
E-mail: [e-mail]

46

THE AUDIO CRITIC

pdf 36

Editor's Note: Readers will pos-

sibly wonder whether this think piece

by Tom Nousaine reflects the philoso-

phy of the Editor and of this journal.

The answer is: not entirely, but some
of Tom's arguments are strong, and
we agree with him on more points
than not. That's sufficient reason to

print him.

* * *

I attended a forum on audio stan­dards for surround sound and HDTV
at the Late, Great, and Last Summer
CES in Chicago. The panel included
editors from a high-end audio and a
high-end video magazine, and repre­sentatives from manufacturers pio­neering new data-reduced multichan­nel formats. The audience seemed to
be mostly audio manufacturers and

journalists. The main question was:

"Will developing data reduction and
compression standards for surround
and HDTV formats compromise
sound quality?"

It is a legitimate question. One
which brings an in-your-throat re­sponse of "It better not!" or we'll be
stuck with less than optimal repro­duction for years into the future
("Look at FM radio!" cry the protest­ers), sacrificing musical enjoyment
for the shrinking audiophile popula­tion, and unknowingly deny the joys
of high-fidelity reproduction to mil­lions of unsuspecting Americans.
During the course of the discourse,
the High-End Community pushed for
an opportunity to review the formats,
inspecting for "transparency" before
standards are set.

The martyr in all of us wants to
stand and applaud. Yeah, let "Us"
march on Washington, demanding
Justice. But wait, Weaselthink de­mands analysis. What are the real im­plications of weak standards? How
will an opaque standard actually af­fect an outcome? A quick review will

ISSUE NO. 22 • WINTER 1994-95

reveal that chickenshit obstruction­ism could only get in the way of our

quest for good sound.

Let's go back a short 15 years.

In 1980 the Compact Disc was still

an idea. A survey of hi-fi consumers

would have shown that the Compact
Disc wasn't needed. We were all
happy with the quality of our best di­rect-cut LPs and 30-ips open-reel
tapes. Many of us would have made
the "you can't go from steak to ham­burger and back" transparency argu­ment.

How wrong those responses
would have been. Analog techniques
are so fundamentally different (and
inferior) to digital, we just had no
real way of evaluating the real bene­fits in advance. Instead, we reactive­ly clung to the positives of the Status
Quo, which, in fact, we knew to be
manifold as we watched it become
remarkably good with improvements
over the years. What possible New
Deal could be substantially better?

In addition to an innate inability
to envision the benefits of change, we
also harbor jaundiced views of
present standards. Standards are sim­ply consistent formats for interfacing
machines and controlling their per­formance; they lower costs and make
the machines easier to use in systems
with complementary machines. For
example, your Bryston works much
better because it has convenient hard­ware and electrical interfaces that al­low easy connection to other audio
machines.

To complicate matters, standards
are often developed before a technol­ogy is fully mature. Check out the
RCA jack/plug. Most of us fear-and­loath them, but at the time—and even
as ridiculous as they seem by today's
rules—they were a key element of
progress. Little development money
would have been put into an industry
that had no rational way of assuring
new products could be easily used by
the market.

Fundamentally, the RCA ap­proached perfection in many ways.
First, it did not materially impede
delivered performance. RCAs sound
like any other decent metal-to-metal
connector at audio frequencies. Im­portantly, they are among the easiest
to use of any type. They are also ac­ceptably reliable in field use. They

are inexpensive and therefore univer­sally available. A perfect standard.

Perfect at the time. Perfect in use.
Perfect over time. In retrospect you
could not have found a better way.

Further, our concept of "formats
and standards" has been shaped by
the hardware-based technology de­velopments of the past. American so-

ciety developed as we found ways of
making better machines and better
machines to make them. The automo­bile, the telephone, modern kitchen
appliances, hand tools, amplifiers,
LPs, even CDs are all good examples
of progress through better machines.

Hardware-based standards tend

to require massive financial commit­ments (changeouts of machines) when
revised, a la Compact Disc. Although
our entertainment values are perfor­mance-driven, they are value-based,
and we are psychologically averse to
throwing out perfectly good, hard­earned hardware. How many of us
hung on to our open-reel machines
years after we no longer used them? I
ditched my LPs pretty quickly but
my turntable lingered until a couple
of years ago.

On the other hand, we also regu-

larly change out existing equipment.
How many phono cartridges have
you swapped out in the name of bet­ter sound? How many amplifiers?
Preamplifiers? Turntables? We change
hardware all the time.

In fact, we crave acquisition of

new present-technology equipment
even when it represents, at best, a
marginal improvement in perfor­mance. Yet, we bristle at the sugges­tion that new technology will force
changeout of our current gear. Even
as we fabricate reasons to buy/trade
new present-technology hardware...
outboard DACs anyone??

But, no one purchased their CD

player or DAT at gunpoint. And how
many can't find replacement cartridg­es? Or can't buy open-reel tape?
Zero. None. On the other hand, how
many gleefully found all kinds of
previously thought-to-be-out-of-print
LPs that showed up when record
manufacturers started cleaning out
warehouses?

Fact is, nobody gets shoved out

when things change. No one is forced
to buy new. It doesn't happen. On
the other hand, current technology

47

pdf 37

suddenly gets light-years better and
much less expensive when new tech­nology shows up. Check out linear­tracking tone arms in the mid-'80s.
Everything, even the junk, gets bet­ter when technology improves. Our
subliminal fears are totally unfound­ed.

Ultimately there are only two

reasons for equipment turnover, any-

way—performance gains and cost

benefits. The movement to digital

was, by any rational standard, perfor-

mance-based, as are the new sur­round formats. On the whole, re­cordings today are frighteningly
better than the best available in 1980.
Sure, there are bad recordings. There
were bad ones in the old days too,
and we all complained about it pas­sionately.

But now we have spectacular

S/N, full bandwidth (2 Hz to 22 kHz),

flat response, 74-minute+ play time,

convenient no-touch access and play­back, smaller size, no wear, and sig­nificantly improved sound. Much,

much better performance.

CDs are also less expensive.
True, the "manufacturer's suggested
retail price" for new releases hasn't
changed in 12 years (given the high
inflation in the early '80s, that's a
price reduction), but the average out­the-door price has fallen signif­icantly. Players are now practically
free. Compact Disc pretty much
sums up the impact of technology
change—higher performance and
lower cost.

Better performance and lower

cost. Why do we recall the pre-CD

era as the good old days? Well, per­haps they were in some respects, but
mostly it's because our memory is
deficient. We regularly recast the
past in a positive light while compar­ing it to the present with all warts in
full itch.

It's the grass-is-greener syn­drome, where time heals all wounds
and every story gets better with each
retelling. Herculean feats become
even more epic; heroes get more he­roic; cars get faster; fish get bigger;
girls get better-looking; and your sys­tem seems better sounding in retro­spect then it ever was in reality.

For example, we all remember
our cars from the '60s as being in­credibly fast. Much faster than to-

48

day's smog-choked pretenders. Yet,

when a car-buff magazine recently
published a comparison of the fastest
Z28 Camaros from the '60s and '70s

with the new 1993 version, the new

car turned out better in every regard,
especially speed. The '93 is a full

second faster to 60 mph than the fast­est Camaro of yesterday. And, of
course, it will go 154, which is 20
mph faster than the fastest vintage Z.
It stops, accelerates, corners, and
rides better than the very best of yes­teryear. Yet, our memories recall oth­erwise.

Okay, so we shouldn't bemoan
the passing of old ways and ignore
newer and better ways. But, that still
doesn't answer the problem of get-

ting trapped with suboptimal format

standards. What happens when a
suboptimal standard gets used?

First, life goes on. The format
adapts, or dies. Check out the com­pact cassette, which started life as a
better dictaphone and ultimately
turned into a decent, if imperfect, re­cording format that blossomed into a
huge market even while saddled with
format-related deficiencies.

I never owned a cassette player
until 1988.1 still have my first one. It
isn't hi-fi. I didn't really want one.
Affected by the suboptimal format?
Sure, lots of manufacturers made a
hell of a lot of money selling cas­sette machines and dumped it into
the development of CD machines and
new recording artists. Good deal for
me.

I never owned an El Cassette, ei­ther. Did I suffer because someone
tried to develop an unneeded or sub­standard format? No way; I was nev­er forced to buy one.

How about DAT? Largely re­garded as a consumer flop, DAT has
quietly revolutionized on-location,
semipro, and amateur recording with
an order of magnitude improvement
in quality and cost. What was in it for
me? More and better recordings. Did
anyone force-feed the DAT on an un­suspecting public?

MD and DCC? Neither seems to
have any real consumer application.

If they fail, what have we lost? Have
you been forced to buy one? Have
they degraded anybody's stereo en-

joyment? Have they caused any re-

duction in recording quality? Or

made other equipment unavailable?

Would it have made sense to
"delay" their introduction because
they might not be transparent? Of
course not. Sony and Philips might
lose their butts on MD and DCC, but
we bear no risk. And in the meantime
Philips has demonstrated that a data­reduced format can be completely
transparent.

FM Stereo? If it were to be in-

vented today we would set higher

performance standards. But, when it
was developing, manufacturers strained
to reach the performance limits. Now
we have digital radio. Sure it will be
better. But it was impossible when
FM was a child. Should we have
blocked the introduction of FM Ste­reo until digital radio was here be­cause the standard was "subopti­mal"? Would that have been the
right thing to do? [Whoa! The debate

at the time focused on the pretty­damn-optimal Crosby system of mul­tiplexing versus the preempting of the
subcarrier for Muzak and similar

commercial purposes. The latter ap­proach prevailed. It was a political,
not a technological, compromise.—
Ed.]

There is less concern today than
ever because things change even fast­er now. Machines are becoming
"processors." Format standards of the
future will be software-based. You
won't have to buy a new machine.
Just an upgrade. And it will decode
all your old material or allow you to
run both versions.

Manufacturers won't need to
buy all new machines to make new
machines. Consumers won't have to
scuttle their old machines, either.
But, as a practical matter, they will
anyway. How many folks use their
original CD player? VCR? Laser disc

player? Ten short years ago I wrote

articles on a Kaypro CP-M machine
with 64 KB RAM. How many com-

puters have you owned or used in the
past 10 years?

Hardware changeout is part of
"higher performance." No, make that
a natural part of life. Yet, we tend to

have sweet, wet psychological
dreams about the next round of
present-technology products and
psychotic nightmares about Freddy

Krueger and Jack Nicholson secretly

developing all subsequent genera-

THE AUDIO CRITIC

pdf 38

tions.

Back to Chicago. I heard a
bunch of progress-paranoid obstruc­tionists, most of whom seemed genu­inely well-meaning, wearing Trans­parency and Consumer Protection T­shirts, encouraging us to write manu­facturers, urging that the High-End
Community be given an "opportuni­ty" to assess the transparency of new
formats before new standards are
adopted.

It was reminiscent of the knee-

jerk reaction to "Perfect Sound For-

ever." PSF, incidentally, turned out
to be pretty close to true on a practi­cal basis. Imagine if we had let the
High-End Community pass judge­ment on CD. 16 bits and 44.1 kHz
would never have been "good
enough" and we would still be vainly
trying to squeeze 16 Hz onto a two­channel plastic groove.

The High End won't concede
that two amplifiers have identical
performance even when the golden­est of ears can't tell them apart with
the nameplates off. Now they want
us to urge others to let them decide

when a data conservation scheme is

transparent? I don't think so.

Sure, we may go through a
couple generations of soft/hardware
incarnations before we have Perfect
Surround Forever. So what? Block­ing the market door, even wearing a
Purity and Transparency T-shirt,
will only serve to hinder R&D and
raise costs. I want manufacturers to
spend their time making money in
music and video. We should encour­age even half-developed standards
which will help companies pull "ear­ly" money out of the market, which
in turn will encourage more develop­ment.

I want High-Definition Video
and Higher-Definition Audio. It
should foster experimentation, en­courage new standards, and help new
products get to the market. The long­standing performance improvement
record of the electronics and comput­er industries, in the face of half­baked formats like Mac and DOS, is
ample evidence that progress cannot,
and will not, be hampered by emerg­ing interface, protocol, and operat-

ing-standard inadequacies.

The biggest mistake would be to

unwittingly join a groundswell of

well-meaning public sentiment with
its head up its ass. Inhibiting new
standards, even imperfect ones, will
curtail the quest for better sound. If
new format standards are truly inade­quate, they will die of their own ac­cord. If they are substandard but
mass-market acceptable, they will en­courage new development which
will, in turn, bring new, better sound
and lower costs. We win in either
case.

The High End really fears a re­prise of the Compact Disc—funda­mentally superior performance at
substantially lower cost. What if the
new digital surround sound blows
away existing two-channel formats,
including CD? What a shock to find
yourself two generations behind in
the quest for improved sound quali­ty! Blocking the doorway to progress
is the ultimate form of pissing
against the wind. Even if you are
wearing your Purely Transparent PC
T-shirt.

Sunfire (continued from page 31)

ously when operated at worst-case levels. (Class AB
amplifiers have maximum dissipation at about 0.41 of
full power into a resistive load. Different dissipation
peaks occur for different reactive loads.) The Sunfire
amp can do this because it is much more efficient.

You may ask why nobody else has done this before
now. Well, the degree of difficulty in designing an ampli­fier of this complexity requires a truly talented and crea­tive engineer. One thousand Wonder-Cap-substituting
nondegreed designers working for one thousand years
could not possibly design this thing. Of course, this is not
the only approach to the design of an efficient power
amplifier but it is unique in not requiring modifications to
the active amplifier electronics to function and hence im­poses no performance degradation on the design. Indeed,
a tracking power rail will hold the VCE of the output de­vices constant and reduce open-loop distortion in the
amplifier's output stage.

The amplifier we briefly tested was a very early
production unit. Construction was excellent, with good­quality parts on double-sided boards. A clamshell chassis
design gives the unit good structural rigidity and the
look of an even more expensive amplifier. Initial mea­surements showed that the amplifier is what it claims to
be—the most powerful amplifier we have tested, al­though the Parasound HCA-2200II came very close. We

ISSUE NO. 22 • WINTER 1994-95

did not test the Parasound under worst-case continuous­drive conditions, but based on its construction I would
expect it would go into thermal shutdown under condi­tions that would leave the Sunfire still operational. How­ever, as explained below, it would be possible to reverse
this situation because of a design simplification in the
production Sunfire.

Some switching noise was noted at the output, and
the protection circuits need some more design work to
prevent premature activation. In the version we examined
the switching power supplies were shared by both chan­nels. Consequently the maximum supply voltage required
by the "hotter" channel must be produced by the tracking
supplies. It is easy to imagine a set of test signals that
would cause the amplifier to have very poor efficiency,
since the right channel might need a high positive supply
rail and the left a high negative supply rail. Under such a
condition the maximum available supply voltage could
be across the output devices, not just the 14 V. Good
efficiency is maintained only if the left and right channels
are correlated. The distortion-reducing property of hav­ing the VCE of the output devices held constant is also
lost with totally uncorrelated left and right inputs. We did
not test this on the early production sample but will look
into it when the unit undergoes full testing.

—David Rich

49

pdf 39

Editor's Note: Just because there are no Harley Howlers discussed below, don't for a moment

assume that Stereophile's Consulting Technical Editor has stopped putting his jittery foot into his
digito-dilettantish mouth. Hardly a week goes by that some degreed technologist doesn't complain to
me about the outrageous Harleyfication of digital theory. But how many times can we repeat the
same depressing message? We must turn our attention to other sources of audio misinformation.

Et tu Forbes?

Forbes magazine is supposed to have a pretty good

research staff. If they did a piece on, let us say, the future

prospects of commuter airlines, they would go to the
world's top authorities on the subject and obtain a variety
of opinions and forecasts. Audio, however, is clearly not
worthy of such conscientious fact-finding. Anyone with
strong opinions is an authority, or so it would appear from
Toddi Gutner's article "Blasts from the Past" (7/4/94).

The article is about the resurgence of vacuum-tube
amplifiers in the audio market and starts with a quotation
from "Tom Havens, a Manhattan attorney and audio-

phile" to the effect that a tube amplifier "gives you the

feeling that you're in the same room with the performer."
Tom Havens? That's the authority available to Forbes?
Can't they call an E.E. professor or a psychoacoustics

Ph.D. at one of the major universities? Somebody like

Dick Greiner would have been very happy to set them
straight. Toddi Gutner then adds, editorially, that "even a
casual listener can easily discern" the difference a tube
amplifier makes because "music sounds warmer, more
natural and realistic."

A qualified source would have explained that the

difference, if at all audible, is not due to tube magic but
to higher second-harmonic distortion and higher output
impedance interacting with the speaker load, both of
which characteristics can be easily and cheaply duplicat­ed in any solid-state amplifier. (Ask Bob Carver, who has
done it often.)

The other "authority" quoted in the article is Wil-

liam Wright, co-owner of Cary Audio Design, one of the
companies that promote the single-ended triode idiocy.
Toddi Gutner would probably consult Dennis Hopper on
the medicinal benefits of cocaine. Forbes editors, repent!

Tim Smart is at it again in Business Week.

This is the same Tim Smart I castigated in Issue
No. 16 for "untutored technobabble" and relying on
tweako sources of information instead of advisors with
scientific credentials. I even wrote a separate letter to
Business Week and got a mealymouthed, all-purpose,
printed postcard for my pains. Well, Tim is still their
man for high-end audio, as witnessed by the "Personal
Business" section (10/10/94), "Electronics" subdivision,
under his byline. What do you think he writes about?
Yes, indeed, tubes! He is a little more careful—could it
have been my column or my letter?—but not enough. For
example: "What makers of tube equipment are seeking is
a more natural sound. Audio engineers speculate that
while all amplifiers distort to some degree, tubes do it in
a manner more gradual—hence less offensive to the
ear—than do transistors."

That's pure garbage. Audio engineers do no such
speculating; ignorant tweaks do. Any audio engineer
with a B.S.E.E. or better knows that a solid-state circuit
can be designed to distort just as gradually as any tube
circuit, if that's what the goal is (in most cases it isn't or
shouldn't be). Furthermore, a properly designed solid­state amplifier can have as little as 0.002% distortion,
which is not "some degree" but at least two orders of
magnitude below audibility, hence in no way "offensive
to the ear." Tim Smart also brings up vacuum-tube guitar
amplifiers but fails to grasp that they are popular exactly
because of their funky colorations/distortions near the
clipping point—they add a little something to the music.
An accurate amplifier, designed for reproduction only, is
not supposed to do that.

And who are Tim Smart's quoted "authorities"?

The marketing man of Audio Research. The president of

50 THE AUDIO CRITIC

Hip Boots

Wading through the Mire of Misinformation in the Audio Press

pdf 40

Valve Amplification Company. Sam Tellig/Gillet of Ste­reophile. Tweaks and suppliers to tweaks. When, oh

when, will a reputable business magazine quote Edward
Cherry or Bob Cordell on the subject of amplifiers? Or
Stanley Lipshitz or Dick Greiner or Floyd Toole or Mark
Davis on any number of other audio topics? Are their
university degrees and professional credentials standing
in the way? Maybe those highly paid editors at the busi­ness magazines don't know how to pick up the telephone
and find a genuine authority on a subject they don't have
to deal with every day. How sad.

Which part of The Absolute Sound d'ya read?

On page 7 of The Absolute Sound's annual Guide

to High End Audio Components, '93/'94 edition, Harry

Pearson writes:

Now, we who review for this publication are
all too aware that some true believers in

search of audio gospel will robotically go out

and buy whatever HP et al. find delectable. I,
for one, would just as soon have this sort
read Stereophile, which seems anxious for
this kind of "authority" over others. This
magazine is for people who think for them-

selves. I detest the notion that I am regarded,
by some, as an "expert" (none exist), and

"authority" (none exist), a "golden ear" when
I have never, repeat never, said or suggested
I wanted to be or was any of these things.
You'd be surprised, however, how many of
the mentally impaired describe us as "self­proclaimed experts or gurus."

Well, I must be mentally impaired because I keep

looking at the front cover of the Guide and I see six

words in big letters that could never, repeat never, have

been printed there. In a 24-point, bold, shadow-style,
Gothic typeface it says (or so it appears to my impaired
mind):

This is an audio journalist/editor we are supposed
to take seriously? And if you ever took him even half se­riously, no matter who you are and no matter how long
ago it was, don't you feel a little sheepish about it? The
man is an irresponsible blabbermouth and an embarrass­ment to the audio community.

Ken Kessler in Hi-Fi News & Record Review

Who is Ken Kessler? He is U.K.-based; he is a
tweako audio journalist (very engage); he writes well; his
main hangouts have been Hi-Fi News & Record Review,

Stereophile (recently dropped from the masthead), and
Audio (recently added to the masthead). HFN/RR and
Audio are currently the English and American representa-

tives, respectively, of the editorial philosophy that science

and antiscience deserve equal time. (Is that Politically
Correct or simply muddleheaded? Don't ask me.)

What caught my eye, long after publication, was a

letter to the Editor by Ken Kessler in the May 1994 issue

ISSUE NO. 22 • WINTER 1994-95

of HFN/RR. He responds to a Mr. Burmajster, who had

apparently disparaged subjective audio reviewing and as­serted the greater validity of measurements. Ken Kessler

writes: "the measurements of hi-fi components have little
if any bearing on the sounds they make" (italics his). He
adds that "unless we're talking about ludicrous anomalies,
like a speaker which has a 20 dB dip at 5 kHz or an amp
with 10% distortion, the hundredths-of-a-percentile mea­surements foisted on us (initially by the Japanese) do
nothing to tell us whether or not a product reproduces
sound convincingly and realistically." Well, Ken, how
about the ±1 dB or so frequency-response tailoring intro­duced by an amplifier output impedance of 1 ohm or so
(see our Issue No. 16, page 55) on which an entire multi­megabuck "tube sound" commerce is based? How about
an error of 0.3 dB in the de-emphasis curve of a CD
player that can make certain pre-emphasized CDs sound
too bright? And, in general, when you subjectively find
that a hi-fi component sounds like this or like that, there
exists no verifiable cause for the effect you perceive?

Further on, the letter challenges Mr. Burmajster as
follows: "if he can look at the measurements (manufac­turer's or self-measured) of five D/A converters and then
match them to the converters in a blindfold listening ses­sion, I will never again write a subjective, measurement­free review." The hypocrisy of that challenge leaves me
in a state of sputtering rage. Here is a subjective reviewer
who will unhesitatingly declare that A has better sound­staging, or less grain, or more bloom than B but never,
never feels the need to prove that he can actually tell A
and B apart "in a blindfold listening session"—and he
has the nerve to challenge someone else to have a perfect
score not in an A/B but an A/B/C/D/E blind test! You
know damn well, Ken, that neither one of you can tell
any two D/A converters apart in blind listening. That's

the whole point, man! Whether they have -97 dB or -85
dB distortion and noise, it's well below the threshold of
hearing, but I want to know those numbers—and a whole
slew of others— to find out which is more carefully engi-

neered and the better value. Do you have a better way of

evaluating them? Yes, I know, "by listening"—except

that you're unable to hear the difference unless you're al-

lowed to look at the nameplates! It's an unfunny farce.

There are certain basic realities in audio today, and
it's time for all audio journalists to face them. Present­day electronic signal paths are sufficiently clean to have
no distinguishable "sound" of their own. You have to

measure them, analyze the circuits, look at the quality of
parts and construction—that's how you evaluate them.

That great admonition "Why don't you just listen?"
works only with loudspeakers, earphones, and micro-

phones today. That makes the nontechnical, noncircuit-

reading, nonmeasuring reviewers who "just listen" some­thing of an underclass, but what do they care? They can
still get all the expensive toys they lust for on extended
free loan from the component manufacturers. •

51

pdf 41

Recorded Music

Editor's Note: David Ranada had a respite this time while I was catching up on my CD backlog, as
promised. (Not that I caught up totally—or ever will.) David will be back in the next issue, so that I

can fade into the background again (where I belong as a music critic, no?). As long as I was at it, I
threw in some very recent releases, so what we have here is a mishmash of the new and not-so-new.

Editor's Grab Bag of CDs,

New or Fairly Recent

By Peter Aczel

Editor and Publisher

The following two items deserved
a little more emphasis than the cap­sule reviews, I thought. (Note that in
all cases the year in parentheses after
the CD number is the year of record­ing, not the year of release.)

Fritz Reiner in "Living Stereo"

All of these Reiner/Chicago re­cordings from 1954—59 have been re­leased on various RCA Victor CDs
before, but the Living Stereo remas­tering process has given them a new
lease on life. It has to say "Living
Stereo" on the label, not Red Seal or
Gold Seal, to indicate this meticu­lously accurate recreation of what's
on the original analog master tapes
(most of them recorded by the great
Lewis Layton).

Richard Strauss: Also sprach Zarathus­tra, Op. 30; Ein Heldenleben, Op. 40.
09026-61494-2 (1954).
Brahms: Violin Concerto in D Major,
Op. 77. Tchaikovsky: Violin Concerto in

D Major, Op. 35. Jascha Heifetz, violin.

09026-61495-2 (1955 and 1957).

Bartók: Concerto for Orchestra; Music for

Strings, Percussion and Celesta; Hungarian
Sketches. 09026-61504-2 (1955 and 1958).
"The Reiner Sound" (shorter works by

52

Ravel, Liszt, Weber/Berlioz, Rachmani­noff). Byron Janis, piano. 09026-61250-

2 (1956,1957, 1959).
Mussorgsky: Pictures at an Exhibition
(orch. by Ravel); Night on Bald Moun­tain; "and Other Russian Showpieces."
09026-61958-2 (1957 and 1959).

There are two utterly remarkable
things about this series. One is the
quality of the recorded sound, which
is only a smidgen below the very best
we have today. Except for the some­what higher noise floor (tape hiss)
and infrequent distortion (mike over­load, tape saturation), these could be

1994 recordings (good 1994 record­ings, that is). For their time, they
were absolutely miraculous. Even
more miraculous is the playing of the
Chicago Symphony Orchestra. They
are still good today but not this good!
No orchestra today plays on this lev­el of virtuosity and synchronicity.
Maybe no orchestra ever did, except
possibly the New York Philharmonic
in the 1930s under Toscanini.

As for interpretation, I would
say the Strauss and Bartók recordings
have never been equaled; the others
are as good as any; and Heifetz in the
violin concertos is of course in a

class by himself. These CDs repre­sent the pinnacle of orchestral perfor­mance and also of audio restoration.

The Bayreuth/Barenboim Ring

In the last issue, I capsule-

reviewed the live Bayreuth (1991 and

1992) recordings by Teldec of Das

Rheingold and Die Walküre. Now the
Siegfried (4509-94193-2) and Die
Götterdämmerung (4509-94194-2) al-

bums are also out, and one can view
Barenboim's Ring in its totality. One
Wagnerite's assessment:

The sound is just right; for the
first time, the unique Bayreuth acous­tic, so perfectly suited to this music,
comes through with every detail in
place, both orchestra and singers in
crystal clear relief against each other.
That alone is worth the price of ad­mission. Barenboim, while neither a
Toscanini nor a Furtwängler, is a su­perior Wagnerian in his own right; he
is not mannered and is always in total
control. The Bayreuth Festival Or­chestra is simply magnificent. Only
the singers are so-so, except for Anne
Evans, an excellent Brünnhilde. John
Tomlinson as Wotan is fair to good.

I'd give this Ring an A minus.

THE AUDIO CRITIC

pdf 42

Athena Productions

This is a small label
from Staten Island, New
York, ostensibly trying to
obtain wider distribution.
They purchased the rights
to the recording below
from Sonora Records.

•

"Music for Violin and
Guitar." Arturo Delmoni,
violin; David Burgess, gui­tar. SACC102 (1991).

Single mike, Blumlein
pattern, no gain riding, cus­tom electronics, tweako en­gineer (Bob Katz, see Issue
No. 17, p. 45), high-end
audiophile artists, cryogen­ically processed violin
strings—get the picture? I
was ready to have a bit of
tweak-bashing fun with
this, but the sound is truly
gorgeous, featuring an ut­terly natural, sweet violin
tone and somewhat reti­cent but beautifully delin­eated guitar twangs. The
imaging is very stable. The
music is mostly transcrip-

tions of light classics (one

contemporary work: David
Leisner's violin/guitar so­nata); the playing is stylish
and highly respectable in

technique. A demo CD,

much as I hate to admit it.

Bainbridge

Here we are in Brad
Miller country, and that
means amplifier-clipping,
loudspeaker-busting sound
effects (although the man
is also quite capable of
making first-rate musical
recordings). The Colossus
digital processor and dis­crete multichannel master
tapes are ingredients of the
SPL fest in each instance.

•

"Sonic Booms 3." BCD

6289 (1992 and earlier).

Missiles, space shuttle
launches, supersonic jets,
and suchlike, alternating
with peaceful surf sounds,

etc. As good as this sort of
thing gets. Forget it with-

out a subwoofer, though.

"4449 Pinnacle! Daylight."

BCD 6295, Discs 1 and 2

(1991-92).

Strictly for aficionados
of classic steam locomo­tive sounds but the very
best of the genre. No, I
didn't listen to every one
of the 27 tracks on two
CDs, but I'm ready to ac­knowledge that Brad Miller
is the master of this. Once
again, a good subwoofer is
mandatory. (And I hope
you already own Beetho­ven's Ninth.)

Chesky Records

This label is one of the
apostles of "natural" sound
via minimalist recording

techniques, an approach
which in most cases yields
superior results—but not
always. (In jazz, Chesky's
long suit, it works without
fail.)

•

"O Magnum Mysterium."
Westminster Choir, Joseph
Flummerfelt, conductor;
Nancianne Parrella, or­ganist. CD83 (1992).

A potpourri of choral
music spanning five centu­ries (from Victoria and
Byrd through Mozart, Ver­di, Bruckner, Brahms, etc.,
to Stravinsky and Mes­siaen), very well sung by
the Westminster group and
chastely recorded by the
aforementioned Bob Katz
through tweako vacuum­tube electronics with audi­ble hiss. Church acoustics,
very good imaging.

CIM

The full name is The
Cleveland Institute of Mu­sic, which is a famous con­servatory, and what we

have here isn't really a CD
label but a one-off, in­house recording project.

•

George Szell: "Music by
George Szell." The Cleve­land Institute of Music Or­chestra, Carl Topilow and
Louis Lane, conductors;
The Cavani String Quar­tet. CIM release #2152
(1992).

Two orchestral works

and a piano quintet by the
great conductor, who was
an eclectic but highly
imaginative composer in
his early youth. Brahms,
Dvoralc, Mahler, Strauss—
you can hear them all in
these clever pieces com­posed between the ages of

14 and 24. The playing is

uniformly excellent in all
three works; the recording
of the orchestra is pleas­antly bright and punchy,
not at all harsh, and ex­tremely detailed and im­pactful (Judy Sherman was
the producer).

Delos

This remains one of

my favorite labels because
the highly intelligent lead­ership of Amelia Haygood
and the unique recording
skills of John Eargle al­most invariably translate
into classy repertory and
superb sound.

•

Ludwig van Beethoven:

The Complete Quartets,

Volumes I through VI. The

Orford String Quartet: An­drew Dawes and Kenneth
Perkins, violins; Terence
Helmer, viola; Denis Brott,
cello. DE 3031-36 (1984-

86).

This is a most unusual
situation—recordings dat­ing back 8 to 10 years, by a
now defunct string quartet,
recorded by a team (Marc
Aubort and Joanna Nick­renz) no longer associated
with Delos, and still being
released one CD at a time
(Volume VI only recently
this year), two quartets per
CD, with five quartets and
the Grosse Fuge yet to
come. But what a series!
The Orford was for more
than two decades the pride
and joy of Canada, as
good a string quartet as
any. I was weaned on the
Budapest and the Guarneri
but I enjoy the Orford
equally. Their playing is
totally secure and highly
nuanced; in the new Vol­ume VI, for example, the
monumental Opus 131 re­ceives a marvelously lucid
and musical performance.
The Aubort/Nickrenz digi­tal tapings are fully up to

1994 standards, as trans­parent as it gets. Even John
Eargle has to be satisfied,
although he would prob­ably have recorded the in­struments a little bit less
close. Matter of taste. As
for the delayed releases,
don't ask me why.

•

David Diamond: Vol. Ill
(Symphony No. 1; Violin

Concerto No. 2; The Enor­mous Room) and Vol. IV
(Symphony No. 8; Suite

from TOM; This Sacred

Ground). Seattle Sympho­ny and Chorale, Gerard
Schwarz, conductor; Ilkka
Talvi, violin (in the Con­certo). DE 3119 (1991-

92) and DE 3141 (1992-

94).

The recording of the
complete works of the age­less David Diamond con­tinues in Seattle under the
supervision of the compos­er. (Beethoven never had
that advantage.) Even the
most recent of these works
is more than thirty years
old; originally they were
not "modern" enough for
the critics but in the post­modern era they come off
as romantically inspired,
timeless art, masterfully
crafted with lots of orches­tral color, and perfectly
suited in sound to the
high-fidelity audio medi­um. John Eargle's highly
panoramic soundstage pro­vides the ideal framework.

You should listen to all
four volumes issued so far

and own at least one.

•

Walter Piston: The Incred­ible Flutist (Suite); Fanta-

sy for English Horn, Harp

& Strings; Suite for Or­chestra; Concerto for String

Quartet, Wind Instruments
& Percussion; Psalm and
Prayer of David. Juilliard

String Quartet; Seattle Sym­phony & Chorale, Gerard
Schwarz, conductor. DE

3126 (1991-92).

Walter Piston's orches-

tral music is exceptionally

"phonogenic"—that is to

say, CD-genic—quite in­dependently of its far from

negligible musical quali­ties. He was also one of
the insufficiently abstruse
mid-20th-century eclectic
composers now that much

more appreciated for his
then-reactionary compre­hensibility. This is part of

yet another Delos/Schwarz
complete-orchestral-works
project, the third of the

Piston series—and these
people appear to be capa­ble of chewing all that
they bite off! The Incredi-

ble Flutist is a delightful

1938 ballet, probably the

most attractive music on
this disc, but the other
pieces are also highly lis­tenable. The audio quality
is positively awesome, one
of John Eargle's most daz­zling efforts.

•

Sergey Prokofiev: Piano
Concerto No. 3 in C Ma-

jor, Op. 26. Aram Khacha-

turian: Piano Concerto,

1936. Dickran Atamian,

piano; Seattle Symphony,

Gerard Schwarz, conduc­tor. DE 3155 (1993).

The Khachaturian con­certo is a flashy, pop­appeal, highly Sovietized
piece in my opinion; the
Prokofiev concerto is an
early-20th-century master­piece. Atamian is a high­powered virtuoso, perhaps
without that final touch of
finesse but still very im­pressive. The curiosity
here is the audio: Delos
without John Eargle. Al
Swanson recorded the 20­bit master; no proprietary
technology is claimed for
the 20-to-16-bit conversion,
but the recording has a
huge dynamic range, and
the sound is still unmistak-

ably Earglian, possibly a

little closer and drier in the

Khachaturian.

•

Richard Strauss: Meta-

morphosen. Arthur Ho-

negger: Symphony No. 2.
Anton Webern: Langsam-

er Satz. Seattle Symphony
Strings, Gerard Schwarz,
conductor. DE3121 (1992-

93).

All of this is string mu-

sic, very competently but
not superbly played. The
Strauss is a late-in-life
masterpiece which should
be in every collection (but
maybe not in this interpre-

tation). The John Eargle
recording sounds rolled off
on top; could the Wadia
digital processor he used
be the reason? (It's not his
usual equipment.)

•

"Inaugural Recital: David
Higgs premieres the C. B.
Fisk Organ of the Meyer­son Symphony Center,
Dallas." DE 3148 (1993).

McDermott Hall at the
Meyerson Symphony Cen­ter in Dallas is famed for
its acoustics and has a
state-of-the-art system of
built-in chambers for tun­ing the reverberation time.
John Eargle, for one,
knows how to tune it, and
the new Opus 100 organ of
C. B. Fisk, Inc., is as good
as it gets. Add to that an
organist with the clean,
elucidating touch of David
Higgs and organ master­pieces by Bach, Liszt,
Franck, et al.—the combi­nation is awesome. Purists
note: one pair of omnis did
it all! I especially recom­mend this CD to audio­philes who are just starting
to get interested in organ
music.

Denon

Nippon Columbia con­tinues to make technically
superior recordings of im­portant artists on the rising
slope of their career, rather
than safe household names
"mailing in their perfor­mance." I always listen to
new Denon releases with a
sense of anticipation.

•

Johannes Brahms: Piano
Sonata No. 3 in F Minor,
Op. 5; Variations and
Fugue on a Theme by
Handel, Op. 24. Bruno­Leonardo Gelber, piano.
CO-75959 (1992).

The sonata, albeit quite
famous, isn't my cup of
tea (see Issue No. 17, pp.
58-59), but the Handel
variations are wonderful
music, beautifully crafted.
Gelber is a world-class vir­tuoso and has great affinity
for these pieces. He almost
sells me on the sonata, too.
The recording is close and
extremely dynamic, with
great in-your-room pres­ence. It is one of Denon's
new Mastersonic Series
with 20-bit processing—
very impressive.

•

Johannes Brahms: String

Quartets, Op. 51 (No. 1 in

C Minor, No. 2 in A Mi­nor). Carmina Quartet:
Matthias Enderle and Su­sanne Frank, violins;

Wendy Champney, viola;
Stephan Goerner, cello.

CO-75756 (1993).

Again, the peculiarly

ISSUE NO. 22 • WINTER 1994-95

53

pdf 43

strained, uptight Romanti­cism of these works repre­sents the least attractive

side of Brahms to me—but
that's just my own low­brow opinion. Besides,
Brahms the master crafts­man often peeks out from
under the sugary glaze.
Very slick, highly disci­plined playing and beauti­fully detailed, first-row
sound.

•

Claude Debussy: String
Quartet in G Minor, Op.

10. Maurice Ravel: String
Quartet in F Major. Car­mina Quartet: Matthias

Enderle & Susanne Frank,
violins; Wendy Champney,
viola; Stephan Goerner,
cello. CO-75164 (1992).

Two masterpieces, mas­terfully played. I don't
think you can get a better
digital recording of this
music. The Carmina players
offer style, precision, nu­ance, beautiful tone—the
whole bit. The recording is
again close, detailed, ex­tremely vivid. A great CD.

•

Joseph Haydn: Symphony
No. 94 in G Major ("The

Surprise"); Symphony No.

22 in E-flat Major ("The
Philosopher"); Symphony
No. 45 in F-sharp Minor

("Farewell"). Orchestre

de Chambre de Lausanne,
Jesus Lopez-Cobos, con­ductor. CO-75660 (1992).

This traversal of the
Haydn symphonies with
modern instruments con­tinues to go well. Lopez­Cobos conducts strong,
clear, idiomatic perfor­mances of these well­known works; the orches­tra is very good but not
great. The recorded sound
is smooth as silk, possibly
the main appeal of this CD
to audiophiles.

•

Felix Mendelssohn: Vio-

lin Concerto in E Minor,
Op. 64. Henri Vieux-

temps: Violin Concerto
No. 5 in A Minor, Op. 37.

Chee-Yun, violin; London

Philharmonic Orchestra,
Jesus Lopez-Cobos, con­ductor. CO-78913 (1994).

This very recently is-

sued recording of the 24­year old Korean violinist
shows considerable artistic
maturation since I first
heard her play as a teenag­er at a private dinner party,
and even since her earlier
Denon recordings (namely

Vocalise: Violin Showpiec-

es, CO-75118 [1992] and

three Sonatas by Fauré,
Debussy, and Saint-Saëns,
CO-75625 [1993]). Her
technique is all that could
be asked for; her tone is
gorgeous; and she is clear-

ly beginning to assert a
strong interpretive person­ality. I'm impressed. Lo­pez-Cobos and the London
Philharmonic support her
with considerable panache.
What's more, this is an­other of the new Master­sonic 20-bit recordings,
and the sound is fabulous.
I can honestly say I've
never heard a better­sounding concerto record-

ing. There is a sense of
completeness and dynamic
ease conveyed by the re­corded sound that I find
quite unusual.

W. A. Mozart: 6 'Haydn'
Quartets (G Major, K.
387; D Minor, K. 421
[417b]; E-flat Major, K.
428 [421b]; B-flat Major,
K. 458 ["Hunt"]; A Ma-

jor, K. 464; C Major, K.

465). Kuijken Quartet: Si­giswald Kuijken & Fran-

gois Fernandez, violins;

Marleen Thiers, viola; Wie-

land Kuijken, cello. CO­75850/51/52 (1990-92).

What a disappoint­ment! These unquestion­ably excellent and scholar­ly musicians play very
chastely and precisely on
period instruments, which
are tuned to A4 = 430 Hz.
Nothing wrong with that,
but the playing is vibrato­less and expressionless
(just the opposite of the
Carmina Quartet), and the
tone is unpleasantly nasal
and penetrating, repro­duced to perfection by the
English engineers. This is
some of my favorite cham­ber music—indeed, some
of everybody's favorite
chamber music—but I find
this version virtually unlis­tenable, complete and au­thoritative though it may
be.

W. A. Mozart: Concerto in
A Major for Clarinet and
Orchestra, K. 622. Fer-
ruccio Busoni: Concerti­no for Clarinet and Small
Orchestra (1919). Aaron
Copland: Concerto for
Clarinet and String Or­chestra, with Harp and
Piano (1948). Paul Meyer,
clarinet; English Chamber
Orchestra, David Zinman,
conductor. CO-75289 (1992).

The wonderful Mozart
concerto receives a fluent,
smooth, professional per­formance without much
expressive nuance. The
Busoni piece is uninterest­ing, but the Copland has a
certain Stravinsky-ish flair
to it; both are performed
competently. The sound

isn't Denon's absolute
best, but even their second
best is awfully good.

Camille Saint-Saëns: Sym-

phony No. 3 in C Minor,

Op. 78 ("Organ"); Le
Rouet d'Omphale, Op. 31;
Phaeton, Op. 39; Danse
Macabre, Op. 40. Michael
Matthes, organ; Orchestre
National de Lyon, Emma­nuel Krivine, conductor.

CO-75024 (1991).

Krivine is highly musi­cal and idiomatic in these
works, but his "Organ"
symphony is a bit tame for
my taste; it should be
weightier and more rip­roaring. The orchestra is
good; the organist is excel­lent; the Denon recording
is once again of demo
quality.

Deutsche Grammophon

The new "4D Audio
Recording" system actual­ly seems to have helped
the generally quite dread­ful DGG sound, whether
or not as a result of all the
much-hyped massaging of
the analog and digital elec­tronics. The fact is that the

4D Schubert disc reviewed
below sounds incompara­bly sweeter and sonically
more believable than the

non-4D Bruckner disc of

the same orchestra and
conductor, recorded by the
same DGG team in the
same church. So the leop­ard can change his spots,
Jeremiah! (Well, maybe
only a German leopard...)

•

Anton Bruckner: Sympho­ny No. 7 in E Major. Staats­kapelle Dresden, Giuseppe
Sinopoli, conductor. 435
786-2 (1991).

A rather uninspired,
heavy-handed, heavy-foot­ed performance, although
the Dresden orchestra is a
great one, as we all know.
The violins scream into
the microphone(s) in the
traditional DGG manner
whenever the music gets
loud. I'm distinctly under­whelmed.

•

Franz Schubert: Sympho-

ny No. 8 in B Minor, D759
("Unfinished"); Symphony
No. 9 in C Major, D944
("The Great"). Staatska-

pelle Dresden, Giuseppe

Sinopoli, conductor. 437
689-2 (1992).

Now this is more like
it. These are much more
loving, idiomatic, nuanced
performances, with many
musical felicities, although
they won't displace any
number of truly great re­cordings of the past. Still,

they're good enough even

for first exposure to these

masterpieces. The 4D sound
is smooth, spacious, au­thoritative, highly listen-

able. About time, DGG.

Dorian

Craig Dory, president
and chief engineer of this
extremely audio-conscious
and audio-savvy company,
took me to task for my
perhaps too casual Editor's
parenthesis (Issue No. 21,
p. 66) suggesting that he
experiments with more
and more ambience out of
boredom or restlessness. I
think I had the right to the
opinion that he sometimes
goes overboard ambience-

wise but I was out of line
when I got personal re­garding his motivations.
He claims that all of his
Dallas Symphony Orches­tra recordings are micro­phoned exactly the same
way. OK, but what about
the McDermott Hall's tun­able reverberation time?

•

Ernest Chausson: Sym­phony in B-flat Major, Op.

20. Jacques Ibert: Escales;

Divertissement. Dallas Sym­phony Orchestra, Eduardo
Mata, conductor. DOR-

90181 (1993).

Chausson's only sym­phony, a fine work in the
Franckian idiom, receives
a very solid performance
here without providing
major thrills (is Mata a
kind of Señor Ormandy?).
The Ibert pieces could be
played with a little more
sparkle, but the orchestra
plays very well and the
musical points are clearly
made. This time the ambi­ence is just right, and Do­rian's 20-bit technology
yields gorgeous sound.

•

Manuel de Falla: La Vida
Breve. Marta Senn, mezzo­soprano; Fernando de la
Mora, tenor; Cecilia An­gell, mezzo-soprano; Simon
Bolivar Symphony Orches­tra of Venezuela with cho­ruses, Eduardo Mata, con­ductor. DOR-90192 (1993).

Mata with a different
orchestra, recorded by
Craig Dory in a different
hall—that's an interesting
situation. Unfortunately, the
performance is not so in­teresting. The singers and
the orchestra are good but
not special, and so is the
performance, which never
really catches fire, al­though Falla's first opera
is fiery enough. The
L6pez-Cobos performance
on Telarc (see below) is
definitely more exciting
and more authoritative.
The 20-bit Dorian record­ing, however, is quite
wonderful—slightly more
refined than the also su­perb Telarc but still very
dynamic, a little more
closely recorded than the
Dallas sessions, in a hall

of the Universidad Central
de Venezuela in Caracas.
No excessive reverbera-

tion—how about that?

•

Joseph Jongen: Sympho­nic Concertante for Organ
and Orchestra, Op. 81.
Camille Saint-Saëns; Sym­phony No. 3 in C Minor,
Op. 78 ("Organ"). Jean
Guillou, organ; Dallas
Symphony Orchestra, Edu­ardo Mata, conductor.
DOR-90200 (1994).

Blockbuster music. The
incredible Fisk organ. Mc­Dermott Hall. The one and
only Jean Guillou. A 20­bit Craig Dory recording.
This CD promises a lot
and delivers most of it. For
a 1926 composition in the
French idiom, the Jongen
work is almost as conser­vative as Marshal Pétain,
but Mata and Guillou play
it for all it is worth. The
Saint-Saëns is performed
with broader strokes and
much more excitement
than in the Krivine/Matthes
version (see under Denon
above). The sound is just a

wee bit more reverberant
and hazy than I like, and
the low notes of the organ
are almost too powerful,
but even so this is quite an
audiophile showpiece, very
clean and dynamic.

•

Sergey Prokofiev: Alexan­der Nevsky Cantata. Dmi-
tri Shostakovich: Sympho­ny No. 9 in E-flat Major,
Op. 70. Dallas Symphony
Chorus; Mariana Pauno-

va, contralto; Dallas Sym­phony Orchestra, Eduardo
Mata, conductor. DOR­90169 (1992).

Mata is in his element
here; these are absolutely
first-rate performances, per­haps without that special
touch of Russian "soul,"
but well-rehearsed, inci­sive, and idiomatic. The
"Battle on the Ice" scene
in Nevsky is awesome. The
tuneful, catchy 9th is Shos­takovich's "classical sym­phony" and is played here
with the appropriate light­ness. As for the sound, it's
pre-20-bit Dorian but one
of their absolute best, ever.
I still insist that not all of
the Craig Dory recordings
in McDermott Hall have
this kind of definition and

just-right liveness. The

bass drum alone is worth
the price of admission
(that's one of my favorite
cliches, as you may have
noticed).

•

Maurice Ravel: Piano Trio

in A minor. Cécile Chami-
nade: Piano Trio No. 1 in
G Minor, Op. 11. Camille
Saint-Saëns: Piano Trio

54 THE AUDIO CRITIC

pdf 44

No. 1 in F Major, Op. 18.

The Rembrandt Trio: Vale­rie Tryon, piano; Gerard

Kantarjian, violin; Coen-

raad Bloemendal, cello.

DOR-90187(1993).

The Ravel trio is a mi-

nor masterpiece; the Saint-

Saëns is utterly charming;
the Chaminade I could
pass up without regrets.
The Rembrandt Trio plays
impeccably; they're fine
musicians. The 20-bit re­cording, in the Troy hall as
usual (but without Craig
Dory) is perfection itself;
the instruments are vivid,
and the hall sound is not
exaggerated. Demo quality.

•

Robert Schumann: Piano
Quartet in E-flat Major,
Op. 47. Johannes Brahms:
Piano Quartet in G Minor,
Op. 25. The Ames Piano
Quartet: Mahlon Darling­ton, violin; Laurence Burk­halter, viola; George Work,
cello; William David, pia­no. DOR-90194 (1993).

Another 20-bit record­ing of chamber music in
the Troy hall, again with­out Craig Dory and again
of the utmost refinement
and palpable realism with

just-right ambience. On
top of it, this is lovely mu-

sic (both works, even if
I'm more of a Schumann
fan than a Brahmsianer),

and the Ames quartet
plays very stylishly, with a
fine combination of nu-

ance and brio.

•

"GRAND CONCERT! Vo­cal and Instrumental Mu­sic Heard in 19th Century
America." D.C. Hall's New

Concert and Quadrille

Band. DIS-80108 (1991).

This is a release in the

Dorian Discovery series,

which includes licensed

miscellanea not necessari-

ly recorded by the Dorian
team (in this case by Sony
Classical). The D. C. Hall
band is a weird antiquarian

group specializing in re-

creating mid-19th-century

American genteel/cornball

(parlor, gazebo?) musical
performances in the most

authentic manner possible.

It's quite amusing. There's

a quintet of flute/piccolo,
clarinet, violin, viola, and
bass violin, plus a tenor—
and what a tenor! His
name is Kevin M'Dermott;

he has a distinctly Irish
tenor sound; and in sheer
vocal quality I would rank

him just a smidgen above

John McCormack. What?

Am I out of my mind? Not

at all. I know a great voice

when I hear one. Of

course, "Come into the

Garden, Maude!" is not
the same test of vocal ar-

tistry as "// mio tesoro,"
but I think if he sang Don
Ottavio at the Met (assum­ing he could learn the
role), the critics would be
falling all over themselves.
He has a much more beau­tiful voice than any of the
aging Three Tenors, for
example. At this point he
seems to be the guiding
spirit of the D. C. Hall
group and maybe that's all
he wants to do. And may­be I'm overreacting, so
find out for yourself—if
you can stand the almost
comically banal music on
this CD. One of a kind and
very well recorded.

Erato

Without being consid-

ered an "audiophile" label,

Erato consistently delivers
as good sound as any­one—and, of course, they
have the Time Warner
clout when it comes to ob­taining the best artists.

•

J. S. Bach: St. Matthew

Passion, BWV 244. The
Amsterdam Baroque Or­chestra; choruses and solo-

ists; Ton Koopman, conduc-

tor. 2292-45814-2 (1992).

An original-instrument,
period-practice Matthäus-
Passion, a bit on the cool
side but effective nonethe-

less, with a nice sustained

flow. The instrumental
playing is better than the
singing; the sound is ex­tremely transparent.

•

J. S. Bach: The Art of
Fugue, BWV 1080. Marie-

Claire Alain, organ. 4509-

91946-2 (1992).

The absolute pinnacle
of contrapuntal composi­tion (left in open score

form by Bach but obvious­ly intended for some kind
of keyboard) is realized
with great clarity here on a
baroquely voiced Alsatian
church organ built in 1975.
Alain uses very colorful
registration but wisely
avoids the full organ. Her
phrasing is chaste in com­parison with that of a Jean
Guillou, but Bach is prob­ably better served that
way. An authoritative, yet
not pedantic, exegesis of
this abstruse masterpiece,
beautifully recorded and
surprisingly listenable.

•

Johannes Brahms: Sym­phonies Nos. 1-4; Tragic

Overture; Academic Festi­val Overture; Variations
on a Theme by Haydn.

Chicago Symphony Or­chestra, Daniel Baren­boim, conductor. 4509-

95191/2/3/4-2 (1993-94).

Surprisingly sensitive,

nuanced, and uneccentric

conducting by Barenboim;
world-class playing by the
orchestra; absolutely love­ly sound captured by the
paradoxically named Larry
Rock in the supposedly
"impossible" Orchestra Hall
—overall this turns out to
be one of the more desir­able Brahms symphony
sets of the digital era. It's
easy to nitpick even the
greatest conductor's phras­ing of every four bars of
this music, but no one can
call these performances
routine or boring. One stu­pid boo-boo: in the third
movement of Symphony
No. 3, at 2:30, there is an
intolerably bad splice that
ruins the flow of the music.

•

Johann Strauss: Waltzes

and Polkas. Chicago Sym-

phony Orchestra, Daniel

Barenboim, conductor.
2292-45998-2 (1992).

Unsurpassed melodist,
master craftsman, master
orchestrator—J, Strauss is
underestimated as a "light"
composer. But you have to
play him with a light touch,
otherwise the debonair
Viennese magic is gone.
Barenboim doesn't have
the light touch; he leans
into the music and drives it
hard, on a tight rein. Even
Erich Kunzel and the Cin­cinnati Pops (Telarc) are

better, because they're

looser, and that's not half

the virtuoso orchestra the

Chicago is. The sound it-

self is as good as in the
Brahms set.

Harmonia Mundi

This is another label

that strives for authenticity

and serious musical values
in their recording projects,
rather than the exploitation

of trendy glamour. Handel

appears to be their favorite
composer.

•

George Frideric Handel:

Messiah. Five soloists; Les
Arts Florissants chorus
and orchestra, William

Christie, conductor. HMC

901498.99 (1993).

Maybe if I were more
of a period-practice purist,
I would find this chamber­sized, elegantly played and
sung performance (of the
original Dublin score only)
less wimpy and ineffec-

tive. This not the Handel

who has thrilled millions
over the centuries. The
magnificence is missing.

The earlier Harmonia

Mundi Messiah, conducted
by McGegan, was emo-

tionally more satisfying—

and of course much more
complete. The recorded
sound of this new version
is, on the other hand, won-

derfully transparent and
timbrally accurate.

•

George Frideric Handel:

Radamisto. Ralf Popken,
countertenor; Juliana Gon­dek, soprano; Lisa Suffer,
soprano; et al.; Freiburg­er Barockorchester, Nich­olas McGegan, conductor.

HMU 907111.13 (1993).

Handel's half-forgotten
operas are an incredible
treasure trove of marvel­ous music, and Radamisto
is one of the greatest, even
though this is the first
commercially available re­cording of it—can you be­lieve it? This beautifully
sung and superbly con­ducted performance (and it
isn't even McGegan own
orchestra!) would deserve
a feature review by a Han­del specialist; let this dilet­tantish reviewer merely
state that, if you like the
early-18th-century kind of
florid singing (and it ain't
Puccini, paisan), then this
is a must. The recorded
sound is as limpid, palpa­ble, and properly scaled as
it always seems to be, re­gardless of the engineer,
when Robina Young is the
producer.

George Frideric Handel:

Arias. Lorraine Hunt, so­prano/mezzo-soprano; Phil­harmonia Baroque Orches-

tra, Nicholas McGegan,

conductor. HMU 907149

(1989-91).

Speaking of the florid
Handel style of singing, if
you want a sampler by one
of the great practitioners,
here it is. There is no bet­ter Handel singer than Lor­raine Hunt; there is no
more authoritative Handel
conductor than Nicholas
McGegan; there is no
more convenient single­CD introduction to this
wonderful music than this
one. Highly recommended.

•

Gustav Mahler: Sympho­ny No. 1 in D Major (with

"Blumine"). Florida Phil-

harmonic Orchestra, James

Judd, conductor. HMU

907118(1993).

The Florida Philhar-

monic is the big surprise
here; they play within a
hairsbreadth of the world's
great orchestras. A little
more weight and plush in
the strings, and they'd be
right up there. After his
outstanding Hoist Planets
(Denon), James Judd was
no surprise to me; he does
an equally intelligent, sen­sitive, meticulous job with
Mahler, making this one of

the better Firsts of the digi-

tal era. Peter McGrath's
recording is state-of-the-

art—again no surprise.

The audiophile community

picked up on it right away.

Koss Classics

This audiophile label is
a subsidiary of the Koss
headphone company, an

outfit that presumably

knows good sound.

Maurice Ravel: Valses no­bles et sentimentales. Oth­er short works by Villa­Lobos, Nazareth, Scriabin,
et al. José Feghali, piano.
KC-1018 (1991).

The piano sound here
is of the life-size, in-your­face, in-your-room variety
and absolutely perfect of
its kind—fantastic dynam­ics, low distortion (another
Larry Rock classical job).
The young Brazilian-born
pianist Feghali has a big
technique (don't they all
these days?) and a respect­able understanding of" the
idiom, but his competition
in this music is just over­whelming.

London

Of the Cleveland Or­chestra performances dis­cussed below, all but one
were recorded by John
Pellowe in Severance Hall
over a period of about a
year and a half. Only the
Till is a John Dunkerley
taping. The audio quality
in all instances is typical
of the English Decca mul­timiked approach, achiev­ing a big, clear, layered
sound with excellent dy­namics and strong bottom
foundation. The stunning
intimacy and impact of the
best Telarc and Delos re­cordings, however (to name
only two examples), isn't
quite equaled. It's more of
a generic and somewhat
homogenized sound—but
still transparent, musical,
and highly listenable. A bit
of edginess obtrudes at
rare moments, especially
on loud brass, perhaps be­cause of too close miking.

•

Anton Bruckner: Sympho­ny No. 5 in B-flat Major.

The Cleveland Orchestra,
Christoph von Dohnányi,
conductor. 433 318-2 (1991).

Anton Bruckner: Sympho-

ny No. 6 in A Major.
Bach/Webern: Ricercare.
The Cleveland Orchestra,
Christoph von Dohnányi,
conductor. 436153-2 (1991-

93).

Bruckner's great, soar­ing themes and colorful or­chestration almost play
themselves, but coherence
isn't his long suit. A con­ductor who can make it all
hang together is way
ahead of the game. Doh-

ISSUE NO. 22 • WINTER 1994-95

55

pdf 45

nányi has that ability, and
the superbly clean, articu­late, disciplined playing of
the Clevelanders helps a

great deal. No two Bruck­nerites will ever agree on
the "best" interpretation of
these problematic works,
but I think these perfor­mances are right up there
with the top contenders. At
the very least, none are
better played technically.

•

Franz Liszt: A Faust Sym­phony. Royal Concertge-

bouw Orchestra, Riccardo
Chailly, conductor. 436
359-2 (1991).

This is much greater

music than most of the

Liszt that you hear much
more often, and Chailly
gives it all the nuanced at­tention and expressive
phrasing a Romantic mas­terpiece deserves. This
kind of music is perfect
grist for the great orches­tra's mill, and they give it
everything they've got. In­terestingly, after my blan­ket comments above on
the sound of the Cleveland
recordings, John Dunker­ley's excellent results here
in a much better hall ap-

pear to have just about the

same virtues and occasion­al faults. Decca blood is

thicker than Amsterdam

water.

•

Gustav Mahler: Sympho­ny No. 4 in G Major.
Dawn Upshaw, soprano;

The Cleveland Orchestra,
Christoph von Dohnányi,

conductor. 440 315-2 (1992).

There isn't nearly as
much portentousness and
breast-beating in the cheer­ful Fourth as in the other
Mahler symphonies; that's
why some people like it
best/least of all. Dohnanyi
responds to the relative
simplicity and classical
proprotions of the work;
those who expect a lot of
heaving and convulsions
will be disappointed. The
orchestral playing is abso­lutely superb, and Dawn
Upshaw is a highly intelli­gent and musical interpret-

er of the last movement,
even if vocally less than
amazing.

•

Richard Strauss: Ein Hel­denleben, Op. 40; Till Eu-

lenspiegels lustige Streiche,
Op. 28. The Cleveland Or-

chestra, Christoph von
Dohnányi, conductor. 436
444-2 (1991-92).

A very solid, impres-

sive, flawlessly played per­formance of Heldenleben
is diminished by the fact
that the Reiner/Chicago
version of 1954 is even
better and almost as well

recorded. ("The best is the
enemy of the good," said
Voltaire.) Till used to be
something of a spécialité
de la maison of George
Szell in his Cleveland
days, and it seems the or­chestra is still good at it.

Mapleshade

This label, devoted

mainly to jazz and blues,
uses custom electronics
and minimalist recording
techniques to obtain a
sound of astonishing trans­parency and presence,

probably the best of its
kind known to me. Their
master tapes are live-to-2­track analog, recorded on a
special machine with very
extended frequency re­sponse—and as far as I'm
concerned that's their
tweaky privilege if the re­sults are this good. I have
picked three of their recent
releases here almost at ran­dom; there's more, and

just about all of it of super

demo quality.

•

"Masters from Different
Worlds." Clifford Jordan,

tenor and soprano saxo­phone; Ran Blake, piano;
Julian Priester, trombone;

etal. MS 01732 (1989).

The idea here is to con-

trast the basically main-

stream Chicago postbop
Jordan (who died recently)

with the crazy avant-garde
Blake. The result is some
very interesting sounds
and damn fine jazz.

•

"Portraits in Ivory and
Brass." Jack Walrath,
trumpet; Larry Willis, pia­no; with Steve Novosel,
bass. MS 02032 (1992).

This is pretty sophisti­cated modern jazz played
by two classically trained
musicians. It may be too
far out for some but no
more than, say, Sonny
Rollins. To me it sounds
just right.

•

"Highways of Gold." Har­vey Thomas Young, vocals
and acoustic guitar; with
Junior Brown, guit-steel
and pedal steel. MS 02252
(1992-93).

This is the kind of
singing and guitar playing
you might hear in a Texas
bar—if you're lucky. Sing­er-songwriter Young has
an appealing style, but the
locally more famous "gui­tar legend" Junior Brown
is very laid-back on this
disc, probably because
he'd steal the show if al­lowed to cut loose.

•

MusicMasters Classics

A label that uses Max
Wilcox to produce their

major releases, such as the
one below, is already on
the right track in my book.

•

J. S. Bach: The Well-

Tempered Clavier, BWV
846-869, Book I. Vladimir
Feltsman, piano. 01612­67105-2 (1992).

Feltsman is a brilliant
technician and also some­what willful in his phras­ing of Bach. The result
ranges from the sublime to
the irritating. The piano
sound is simply gorgeous.

Reference Recordings

This label confesses the

audio gospel according to

Saint Johnson (Keith O.).

•

"Trittico." Dallas Wind
Symphony, Frederick Fen­nell, conductor. RR-52CD
(1992).

"Pomp & Pipes!" Paul
Riedo, organ; Dallas Wind
Symphony, Frederick Fen­nell, conductor. RR-58CD
(1993).

Both of these programs
of short, mostly 20th cen­tury, mostly showy pieces,
effectively conducted by
the ancient Fennell, were
recorded in McDermott
Hall and are intended to be
played back with HDCD
decoding, which I don't
have yet as of this writing.
Undecoded the recordings
have sensational dynamic
range, awesome bass, very
nice ambience and dimen­sionality, but I've heard
greater transparency and
finer detail on other RR

CDs. Whether HDCD is
an advancement in accura­cy or just another proces­sor remains to be seen.

Telarc

What other audiophile­oriented label has as many
good artists as Telarc? None.

•

J. S. Bach: Brandenburg

Concertos. Boston Baroque,
Martin Pearlman, direc­tor. Nos. 1, 2, & 3: CD­80368 (1994). Nos. 4, 5, &
6: CD-80354 (1993).

As authentic, propul­sive, and convincing as
any period-instrument per­formance in the catalog—
and better recorded. What
else is there to say?

•

Ludwig van Beethoven:

String Quartet in E-flat
Major, Op. 74 ("The
Harp"); String Quartet in
F Minor, Op. 95 ("Seri­oso"). Cleveland Quartet:

William Preucil and Peter
Salaff, violins; James Dun­ham, viola; Paul Katz, cel­lo. CD-80351 (1991-92).

The utmost finesse, as
against the powerhouse
school of Beethoven play-

ing, characterizes these per­formances, which are at
least as successful artisti­cally as the previous two
CDs in the series. The re­cording is pure silk and of
superb transparency. (The
producer was Judy Sher­man, Max Wilcox's ex.)

•

Manuel de Falla: La Vida

Breve. Alicia Nafé, mezzo­soprano; Antonio Ordonez,
tenor; May Festival Chor­us, Robert Porco, director;
Cincinnati Symphony Or-

chestra, Jesus López-

Cobos, conductor. CD-

80317(1992).

Beautifully sung, idio­matically conducted, and
superbly recorded perfor­mance—definitely the pick
of the digital era in this
short, uneven, but ex­tremely vital early opera
of the composer. Killer fla­menco passages.

•

Alexander Glazunov: The

Seasons, Op. 67 and 67a;
Scènes de Ballet, Op. 52.
Minnesota Orchestra, Edo
de Waart, conductor. CD­80347(1993).

The scrumptious Rus­sian ballet music is just
one attraction here; an­other is the verve of the
Minnesota players; and
still another the audio

quality—unusually clean,
smooth, and natural even
for Telarc. Is it the new
and different (for them)
venue or the 20-bit A/D
processing? Only Jack
Renner knows for sure.

•

W. A. Mozart: Cost fan

tutte. Felicity Lott, Marie
McLaughlin, Nuccia Focile,
sopranos; Jerry Hadley,
tenor; Alessandro Corbel­li, baritone; Giles Cache­maille, bass-baritone; Edin­burgh Festival Chorus;
Scottish Chamber Orches­tra, Sir Charles Macker­ras, conductor. CD-80360­A/B/C (1993).

A very important re-

lease, deserving of an ex­tended feature review a la
David Ranada. In future
years this may possibly be
ranked as the digital era's
equivalent of the 78-rpm
era's Busch/Glyndebourne

1935 recording. Mackerras

conducts with a combina­tion of easy flexibility and
disciplined control in per­fect dynamic balance, and
the voices are all fresh and
lovely. Of course, the star
is still Mozart. (Stravinsky
played this music over and
over again on his phono­graph while composing
The Rake's Progress.) The
20-bit recording is abso­lutely beautiful in texture
but presents a relatively

small soundstage, perhaps
deliberately, to suggest a
small, intimate theater.

•

Sergei Rachmaninoff: Sym-

phony No. 3 in A Minor,

Op. 44; Symphonic Danc­es, Op. 45. Baltimore Sym-

phony Orchestra, David

Tinman, conductor. CD­80331 (1994).

To me, this CD proves
two things. One, that
Rachmaninoff composed
his best music in his 60s.
Two, that Jack Renner can
occasionally surpass his
normal high standard in re­corded sound. The string

tone, dynamics, and round­ed solidity of this record-

ing are extraordinary. And,

yes, the Baltimore has be­come a very fine orchestra.

Teldec

See also the Wagner

Ring review on page 52.

•

Béla Bartók: Concerto for

Orchestra; The Miracu-

lous Mandarin (Suite for
Orchestra); Deux Images.
Philharmonia Orchestra,
Hugh Wolff, conductor.
9031-76350-2 (1993).

The great acoustics of
Watford Town Hall, su­perb playing by a world­class orchestra, sensational
recording by Teldec's Ger­man team—these are the
strengths of this release.
As for Wolffs conducting,

his inauthentically slow
tempo in the second move­ment of the Concerto tells
the story: too careful, not
incisive enough, not at all
like Reiner or Solti. (Call
me a Hungarian chauvinist.)

•

Gaetano Donizetti: Lucia
di Lammermoor. Edita

Gruberova, soprano; Neil

Shicoff, tenor; et al; The

Ambrosian Singers; Lon-

don Symphony Orchestra,
Richard Bonynge, conduc­tor. 9031-72306-2 (1991).

I agree with Toscanini

about Lucia: "Che bell'
opera!" Never mind the
coloratura shenanigans; lis­ten to the melodies, the
ensemble writing, the lyri­cism alternating with dra­ma—it's good stuff. Here
is the same pair in the
leading roles that I liked so

much in the Teldec Travi-
ata: Gruberova has touch­es of greatness (vocally if

not dramatically) and Neil

Shicoff has a still fresh,

unabused, beautiful tenor

voice. Bonynge's conduct-

ing could be more dynam-

ic but he keeps producing

lovely sounds, and that's

what this opera is about.

The recording is perhaps a

bit constricted in the busy

moments but always clean.

56 THE AUDIO CRITIC

pdf 46