In addition to editorial correspondence of general interest, we're including two special categories of letters this time. One has to do with clearing the air after a particularly foul-smelling anonymous communication circulated about The Audio Critic; the other covers exhaustively the "Mark Davis syndrome' originally brought up in our first issue as a kind of aside in Part I of the preamp survey. The Mark Davis correspondence is reprinted here in 8-point type for those who are sufficiently motivated to plow through it; we didn't feel we'd be justified giving it more space. The letters we publish in this column 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, Box 392, Bronxville, New York 10708.

In the last week of June and throughout most of July, the anonymous missive reprinted below was being sent out to literally hundreds of high-end audio dealers, manufacturers, reviewers, editors and other high-end audio practitioners from coast to coast. The anonymous commentator had invested 24 cents first-class postage per mailing, as each fat envelope also contained Xeroxed copies of our allegedly misleading ads and of trade-press clippings documenting that two of our part-time staff consultants made their daily living in the audio business, one in retailing, one in manufacturing. All of these envelopes were without a return address and postmarked either in Hartford, Connecticut, or Manchester, Connecticut (a suburb of Hartford).

The Audio Critic--A Ripoff?

Another "little magazine" has appeared.

This one bears watching.

Initially, our hopes were high. The price ($28.00 for a year's subscription) was high but then so was the promise. By combining objective measuring techniques with long term listening sessions, The Audio Critic promised information, not just opinion.

It now appears that the price was too high as The Audio Critic is, in our opinion, incapable of fulfilling the promise.

Contrary to claim, The Audio Critic has a number of substantial ties with the audio industry, both on the retail and manufacturers' level. These ties make objectivity impossible. In fact, these ties just might make The Audio Critic a "house paper." The attached information begins to point out the connections.

We feel there may be more tidbits yet to be un earthed.

But even if these suspected alliances prove to be untraceable, The Audio Critic has misrepresented itself to the audiophile-consumer. Where is the line between misrepresentation and fraud? Just as important-can anything written by The Audio Critic be believed? We think not.

The Audio Critic is, to us, severely compromised.

A publication with no credibility serves no other purpose than to make someone a lot of money.

Peter Aczel, the Editor and Publisher, has a lot of answering to do.

In the wake of the anonymous mailings, the following exchange of letters took place:

Dear Mr. Aczel:

No doubt you have seen the anonomous (sic) letter that is being circulated, accusing your publication of having 'substantial ties with the audio industry," and supplying what appears to be some documentation of that allegation.

We were of course sent the sheaf of papers in question, and while we do not intend to print the letter--it goes rather overboard, in my opinion--we cannot let the matter pass without at least some comment to the effect that the letter was circulated and commercial affiliations of two of your key staffers documented from other, independent sources.

May I please have a statement of your response to this, for publication?

Sincerely,

J. Gordon Holt Editor & Publisher, The Stereophile

Dear Mr. Holt:

Two things surprise me about your recent letter.

One is that you assume you're entitled to a reply, even though you're neither a subscriber of this publication nor a manufacturer whose equipment has been re viewed in it. I'll let that pass.

The other is that you, as a reputable person, should be impelled to initiate this dialogue by something as sleazy as an anonymous letter, which should have been disposed of in the same manner as bird droppings. Since the same information is available to you from other sources, as you say, I wish you had referred only to the latter. It would have been cleaner that way.

It seems, however, that in addition to his soiled character, the anonymous letter writer also has a reading problem. As the enclosed excerpt from the second issue of The Audio Critic shows, I had already discussed quite freely in print what he started to whisper behind my back months later. (Moral turpitude and lack of mental alert ness frequently go together.) To round out this editorial on the subject (which I must ask you to reproduce in its entirety along with this letter, otherwise I can't give you my permission to print either one), I would also like to make the following points:

My wife and I are sole stockholders of The Audio Critic, Inc., which banks the subscription fees. Neither one of us has any affiliation with audio manufacturers or retailers, nor does anyone else on our payroll. The two consultants (not key staffers, as you say) triumphantly tracked down by the anonymous creep are both old friends of mine who had lent their names to my infant enterprise without any remuneration. Now that The Audio Critic is a going concern, I am entering, and have entered, into many more consulting relation ships with technologists of all sorts employed by the audio industry-so many, in fact, that it's no longer practical to list them on the masthead, especially since some of them don't wish to be so listed.

I spend hours on the telephone or in person with these engineers and designers; I pick their brains; they pick mine; they influence my technical thinking; I in fluence theirs; they come to my laboratory, I go to theirs;

we discuss test methods; we have drinks or even dinner together; they show me their latest research and papers --and they haven't got a nickel's worth of influence on my income or my editorial policy. The list of their names would excite the anonymous letter writer into an orgasm of I-told-you-so.

I, on the other hand, feel that such alliances are better for my education and therefore better for my sub scribers than if I consorted exclusively with untutored audio freaks-which is obviously what the anonymous letter writer thinks an honest noncommercial audio journalist should do.

Does this answer your question? Sincerely, Peter Aczel Editor and Publisher (The editorial excerpt from our second issue, as enclosed with the above reply, is reprinted here.) We have come under criticism by about 0.1% of our readership (three persons, to be exact) for the professional involvement of some of our staff consultants in the audio industry. Doesn't that constitute a conflict of interest, we were gleefully challenged, in view of our simon-pure posture of independence? Ah, that's a good one. The basic reason for the amateurishness of the "underground" audiophile reviews is that they are staffed by amateurs. It would be very nice if one could come to valid conclusions about, say, the transient response of an amplifier by consulting music-loving dentists, accountants and shoemakers.

Unfortunately, such independent experts seldom know what they're talking about. That's why we have a professional record producer, a physicist/audio engineer, an audio-electronics technician and other qualified professionals on our staff. Sure, some of these people derive part of or all of their income from the audio business, but not one of them is a chief executive officer or majority stockholder, so that the worst that can happen is that the views they privately communicate to us deviate from the self-interest of their bosses. Tough. It just so happens that The Audio Critic has already dealt rather severely with products made or sold by said bosses.

The important thing is that the management of The Audio Critic is completely divorced from commercial audio. The Editor/ Publisher deliberately severed all connections with the industry before coming out with the first issue. Our subscribers are our sole business interest.

As a matter of fact, if you hear any malicious gossip about The Audio Critic's conflicts of interest, or especially about our taking bribes for favorable reviews (one of the ever-recurring fabrications about nearly all audio reviewers, perpetuated by a few pathetic little would-be authorities), we suggest you let us know, pro vided you're willing to identify the source. There's nothing a Hungarian loves better than a good lawsuit.

Dear Mr. Aczel:

We've had an anonymous packet regarding your publication's claim to impartiality. I imagine most publications in our field have. Would you care to comment on the present relationships of your staff to commercial interests?

Sincerely, Edward T. Dell, Jr.

Editor/Publisher, The Audio Amateur

Dear Mr. Dell:

In reply to your recent letter, I am enclosing a copy of a similar letter from J. Gordon Holt and of my response to it.

Since my answer to your letter, had it preceded Mr. Holt's, would have been virtually identical, there is really nothing further to say, except that this makes you the second reputable audio journalist without a clearly expressed distaste for anonymous letters.

Sincerely, Peter Aczel Editor and Publisher The Audio Critic

Dear Mr. Aczel:

We have received an anonymous note regarding your magazine, entitled "The Audio Critic-A Ripoff?" It's most probable that you have already seen a copy of this "material." It appears that considerable time, trouble and expense went into the preparation and distribution of this note, but we are unable to determine the intent of the sender; altruistic self-righteousness seems out of the question due to the effort involved.

Some of us take issue with some of your published opinions, Mr. Aczel, but we all agree that your conclusions are fairly presented and without discernable bias.

We question the composition of your reference system, having learned that there are amplifiers far more definitive than the Quatre, but we attribute that to subjective impressions rather than influence of industry.

You have our absolute support. We will continue recommending your fine publication to audiophiles within our area of influence. We are convinced that the more any client knows about audio, the better he'll like our way of doing business.

Sincerely, H. L. Eisenson Audio Dimensions San Diego, CA

And now for saner and more savory matters.

The Audio Critic:

... I applaud your attempts to correlate lab data with audible effects; this field seems to be limited to yourselves and Richard Heyser in terms of published material, especially since Bascom King no longer reviews equipment publicly.

My major complaint with your magazine concerns a fault common to both you and your competitors, and is a simple one. Simply, there is no state-of-the-art in audio today (with the possible exception of tuners); to attempt to recommend one piece of equipment as the best of its art is foolhardy. To truly be the '"'state-of-the-art" a component must outperform its competition in every parameter considered. That is, when two pieces of equipment are compared, there should be no doubt in the minds of the listeners as to which is superior. If the sound of both components is felt to have different positive qualities, then clearly neither represents the SOTA, since existing technology could better either component.

In fact, you allude to this characteristic several times, yet in the final analysis appear to ignore it. You could put yourselves yet another step above your competitors if you did not attempt to follow in their footsteps by finding a "best"; there is certainly more than one legitimate approach to the SOTA, and putting one above the rest smacks of the commercialism you are trying to avoid, as well as being unrealistic. . . .

Sincerely, Robert Bertrando Tucson, AZ.

We agree with you wholeheartedly: a mixed bag of superiorities and inferiorities doesn't constitute SOTA, even if on balance the equipment is judged to be prefer able to all others tested. We have tried to qualify our conclusions and recommendations to make that clear, unfortunately the word “best" tends to stick out and obscure qualifiers such as 'so far" or "depending." We'll try harder in the future.

-Ed.

The Audio Critic:

"Accuracy" is an impeccable word, and your bit on page 7 of the March/April issue, "In other words, accuracy isn't a matter of taste," is very neat and convincing. Up to a point.

As you say, "either they did or they didn't" match the color. Agreed also that it's difficult in audio, but your point remains valid, and taste has nothing to do with it. You look at your car and the repainted door and say, "They don't match." IT don't doubt that you see a mismatch.

I come along. I look from car to door, and I say,

"They did a fine job, didn't they!" I see a perfect match.

"Accurate," I say. "Not accurate," you say. And we are both right.

We both see what we say we see. But we happen to have different eyeballs. I have a friend who sees almost no red, but sees beyond violet into ultraviolet, to make up for it; most variations are not so spectacular. Which is to say that your color-matching analogy is valid, your "accuracy" motto is valid-until another person comes along. Eyeballs are not standardized.

Ears are not standardized either.

So that impeccable word ''accuracy" is not the solid support you may think it is. I hope you use it with care.

Sincerely, Burnett Cross, Hartsdale, NY

You have a point there but, in our opinion, you're bearing down a bit too hard on it. There exists a spread in physical perception from human to human, but be tween reasonably young and healthy specimens the spread is quite narrow, so that a norm for our species is a viable concept. If color matching, for example, were as subjective, indeed anarchic, as you suggest, wallpaper companies couldn't stay in business-nor could Eastman Kodak. In audio, we'll settle for a match to the original as good as the average car owner would let the body shop get away with.

-Ed.

The Audio Critic:

Would you please discuss the conflict (to my mind) between the ideal "straight wire with gain" concept, represented by the state of the art preamps and amps with a minimum of controls and frills, and the conflicting use of expanders, companders, equalizers, and noise reduction devices a la Phase Linear 1000 or Burwen, etc.

In other words, is it insane to own a Rappaport preamp, Quatre amp, DQ-10's, etc. and a ten-band/ channel equalizer-dbx expander-SAE "click and pop machine" -Phase Linear 1000-or any one of a dozen accessories? Do all the many add-on devices only serve to destroy the '"'art™ in "state of the art" represented by an exceptional amp or preamp? Thank you, David Gibbs Rancho Palos Verdes, CA The signal-processing accessories you mention are a necessary evil-and not quite as often necessary as a lot of audio people think. Such devices are designed to correct what somebody else has already messed up, either as a result of inadequate technology or through sheer neglect. With today's best program sources, played in a room that isn't an absolute pig acoustically, you're much better off with the shortest, simplest possible signal path from input to output. In fact, with all those "black boxes" in the signal path, you're quite right to suspect that the difference between, say, a Rappaport preamp and a Marantz would be obscured. We never touch the stuff ourselves, unless it's on loan for evaluation.

-Ed.

The Audio Critic:

There is one aspect of the preamp 'controversy' which confuses me and, perhaps, other non technical types. It is the question of stereo depth. If distortion gives the illusion of depth and true accuracy demands a sacrifice of the illusion, I would opt for accuracy. Isn't this question answered by using a record or tape where the reviewer would be aware of the precise location of the instruments during the live performance? A string quartet, even if the performers were local amateurs, might be ideal . . .

Yours truly, Edward J. Doyle

East Northport, NY

There's a subtle but audible difference between the haphazard artificial depth resulting from electronic anomalies and the focused natural depth rendered by accurate reproduction. There's no sacrifice involved in the latter; it sounds just right. We agree with you that complete familiarity with the depth and width of the sound source at the actual recording site is a great help in judging this phenomenon. Of course, the microphoning must be simple and straightforward; gimmicky multi miking can throw you off completely.

-Ed.

The Audio Critic:

Having purchased two Rappaport preamps on the basis of Vol. 1, No. 1, I must say your writers should have elaborated on the PRE-1's highs. They are objectively beautiful, but subjectively emotionally empty.

Most people agree the program-listener schism is anxiety provoking. Has anyone built a non-fatiguing SP-4 with sock?

James A. Debros St. Cloud, MN

No, no, no! What you need is a non-objectively transcendental preamp that is in perfect harmony with the vibrations of the universe, not to mention our precious bodily fluids.

-Ed.

The Audio Critic:

I have just read the first two issues of The Audio Critic and found them to be very interesting . . . . As an electrical engineer I am happy to see perceptive criticism coming from persons who seem at the same time to believe that there are laws of physics and electricity which must hold no matter what our wishes to the contrary. I look forward to future issues and your continued success.

There are two comments that I would like to make regarding the following quotes from The Audio Critic, Vol. 1, No. 2. On page 41 you state:

"The net result is linear-without feedback and hence without TIM." There is at the present time some confusion about the precise term that should be used for the distortion now commonly called "Transient Intermodulation Distortion." Some persons call it

"TIM," others call it Slew Induced Distortion or Transient Overload Distortion or Frequency Overload or any of several other terms. Whatever term is used, it is clear that this form of distortion can occur within a feedback loop or outside of any feedback loop.

Thus, it is not really correct to say that an amplifier, whether it consists of one or several stages, will certainly be free of transient distortion just because it has no feedback. A single stage amplifier with no feed back will fail the Otala test if its slewing capacity is exceeded. This is a small point but if we are ever to understand precisely what is going on in amplifiers we must think about them very carefully.

What is true about feedback amplifiers is that, if there is any tendency for internal slew or amplitude limiting to take place, the problem is greatly exacerbated by the feedback. However the problem is not caused by the feedback but is caused by the limiting tendency in the first place.

It is not correct to say "no feedback-thus no TIM." It is correct to say "little feedback-thus less problem with internal TIM." My second comment relates to the following parenthetical expression: ""(They'll never teach it this way at EE school, but then the tuition there is more than $28 a year.)"" I realize that this is just a nasty little comment set down to thrill your readers. (Critics are supposed to be mean to the core, | hear.) My issue with the comment is that it is not fair or accurate. In the attached reference list from an article by Jung, most of the references are from universities (Daugherty, Greiner, Leach) or certainly from highly trained and qualified electrical engineers.

We in the university are trying very hard to under stand the very same problems that plague us all. I have for 15 years taught about overload problems in feedback amplifiers to my students. I have also written about these problems in the technical literature. The public and amplifier designers have not been listening very well.

Certainly the university is not the source of all knowledge, but we have tried. Much credit for improved amplifiers must go to the individual inventor of course.

Unfortunately they have often been more reluctant to share their accomplishments and insights with the public.

What I am really suggesting to you is that we all need each other in order to both solve and understand the complexities of amplifiers of any kind, whether they are power amplifiers or pre-preamplifiers. In order to join together in these efforts we must make an attempt to communicate better. This will happen only if we try for harmony instead of acrimony.

Please consider these comments to be harmonious.

Let me close this time with a bit of acrimony of my own directed to the inventor-producer of some of the flood of equipment that exists.

It has been my observation that many writers, those that write for pulp magazines or the most esoteric journals, all have a primary purpose to spread their insight and understanding of problems to others. They are in fact teachers in the most generic terms. The editors of The Audio Critic are clearly doing the same.

This is the way to understanding since the process gets the best minds focused on the problems.

On the other hand, the individual inventor-producer all too often has just the opposite goal in mind.

He hides his techniques, he hides his theory, or worse, spouts garbage and nonsense; he misleads, he buries his circuits in epoxy, he gives erroneous specifications, he never writes about his circuits or theory of operation, he never supplies operational manuals that are correct, he never-never supplies repair manuals, and so forth.

This is not true of many manufacturers, of course, but all too often the most mystical manufacturers are the ones that make the most outrageous claims. It is very hard to tell the fakes from the frauds.

I would urge you to ferret out the fakes. Make them publish their theory, if any. Make them tell the truth about their specifications. Make them show the magic, if any, of their circuits. Have them explain how they defy the laws of physics.

I have hopes that you will accomplish more of this than the mystics among the golden-ear crowd have to date.

Very best regards, R.A. Greiner Professor of Electrical Engineering University of Wisconsin We stand corrected on our TIM / feedback simplism and fervently endorse the admonishment of undisciplined inventor-producers and muddleheaded audio mystics.

What distresses us, however, is that Professor Greiner, who is one of our heroes (he was coauthor of a superb 1966 paper that anticipated the entire Otala-Curl-Leach Jung scene), should so direly misinterpret our little quip about EE school and the $28. We were actually apologizing, in our own left-handed way, about the brevity and patness of our explanation of the analog multiplier ("gain cell"') principle. You academicians must really feel be leaguered if you take a mildly embarrassed journalist's foot shuffling to be a kick in the pants!

-Ed.

The Audio Critic:

Re your CES report on the Verion NF-1 Filter: The pros and cons of wide bandwidth, DC to MHz, versus narrow bandwidth, 20 Hz to 20 kHz, have been with us for more years than I care to remember.

The controversy simmers and boils. Now it has come to a boil again.

J. Peter Moncrieff in his belatedly issued Inter national Audio Review, July 1976, page 38, states that he considers a low-frequency cutoff of 0.5 Hz to be barely low enough for achieving accurate bass reproduction of music. On page 180 he considers an audio component with less than 0.5 MHz bandwidth to be audibly objectionable.

Mitchell Cotter believes that any bandwidth in a power amplifier (why not a preamp also?) in excess of 18 Hz to 35 kHz is detrimental to the reproduction of music and is asking you, as you say, to plunk down $300 to prove he is right.

This is a very serious issue to the dedicated audiophile interested in State of the Art equipment and should be explored thoroughly by The Audio Critic.

I am on the narrow bandwidth side even without the time compensation of the NF-1 filter. It has yet to be proven that phase distortion of the signal between the input and output of an amplifier is audible.

As you know, the reason for negative feedback is to reduce distortion. The wide bandwidth that is a concomitant of this feedback may be an asset or a detriment, depending which side of the fence you are on.

There is a new school of thought (of which you are a member) that claims the less feedback the better the sound. I am inclined to think that the reason is the reduction in bandwidth of the less-or-no-feedback amplifier. With the NF-1 filter you can eat your cake and have it, too.

Think of the poor amplifier designer. What will a 20 Hz or 20 kHz square wave look like at the output of an 18 Hz to 35 kHz amplifier. Horrible, that's what.

That it will sound better will not be believed by the re viewers who look at those square waves on their scopes.

Julius Futterman Futterman Electronics Lab New York, NY

No comment necessary-except to remind all you young whippersnappers to listen respectfully when Julius Futterman talks. He didn't learn what he knows from a Tech HiFi salesman three weeks ago; he was designing State-of-the-art tube amplifiers when you guys were watching Howdy Doody.

-Ed.

The Audio Critic:

When bi-(tri-, quad-, etc.) amping a system using electronic crossovers, the only speaker which needs a really high-slew-rate amplifier is the highest-frequency unit. In most cases, for high sound pressure level, low distortion treble, a low-power high-quality amplifier can be used, such as Mark Levinson's new class A job.

The rest of the power amps can be of lower slew rate, because the electronic filter networks will prevent any fast-rising waveforms from reaching their inputs.

This is particularly true for subwoofer and woofer amps.

High-slew-rate power amps with expensive, wide-band width output stages are simply a waste of money for the low-frequency end of a system.

Since TIM and related distortions cease to be a problem for amplifiers driven by low-pass filters, large amounts of negative feedback and compensation can be used to get extremely low output impedance. This will create the high damping factors necessary for good, tight bass and midrange. Op amps such as the 741 and many of the low-frequency high-power output devices should be naturals for such amplifiers, bringing prices down drastically.

This should cut down the $20,000 price tag considerably on your Super Dooper Mark Levinson System.

Alan Hoover, Indianapolis, IN

Cf. the letter from Chris Russell of the Bryston company, published alongside our power-amplifier survey in this issue.

-Ed.

The Audio Critic:

Since I had not received the first two issues and have had to see them from other sources I am quite annoyed. It is quite obvious that you have a grudge against me. However, your lack of ethics are in question as far as I'm concerned. (It are?-Ed.) You accuse me yet allow me no defense. You compound this by refusing to honor my subscription which I have paid for. The en closed photocopy is self-explanatory.

James Bongiorno, West Hollywood, CA

Everybody is ganging up on poor Jim Bongiorno, including the United States Postal Service. His subscription had been sent to the GAS company and went astray; we sent him replacement copies to his home.

The notion that we punitively withhold paid subscriptions from people we "have a grudge against" is down right pathological; of course, as Kissinger used to say to Nixon, 'just because you're paranoid doesn't mean they're not out to get you." As for allowing no defense, any manufacturer with anything factual to say in defense of his equipment or his company has this column open to him, and we never said anything to the contrary.

-Ed.

The Audio Critic:

It will be of some value to the other readers of your fine periodical if a couple of technical errors in the first two issues are corrected:

(1) In "Have Tone Arm Designers Forgotten Their High-School Geometry?" you criticized the Grace G-707 for its "incorrect" L sin 8 of 3.375". (Did you really measure it accurate to .001"?) I have measured 8 on my G-707 and have found it to be 23.25° (plus or minus 0.25°), which, in combination with the 9.33" L specified by Grace (L is adjustable by about 0.4", a fact you neglected to mention, which means therefore that L sin is adjustable) computes to, surprise, 3.68"; the exact number you described as being ideal. I think you owe Grace an apology.

(2) In "Fishing for Bass: A look at the Subwoofer Scene," you made the statement that a sealed system with Q = .707 has, among other things, "transient response . . . best possible for these conditions." Although it is not clear what is meant by "these conditions," I can point out that, while a Q = .707 sealed box loudspeaker does have a maximally-flat (i.e. Butter worth) frequency-amplitude response, it does not have a maximally-flat (i.e. Bessel) frequency-delay response, and hence does not have the best transient response obtainable for a closed-box system. This occurs with a Q = .58. The difference between the two transient responses is that the former tends to ring just a little at a frequency equal to the cutoff (and resonant) frequency.

In a subwoofer this is, of course, 20 to 30 Hz. Whether or not this is "audible" is an interesting but (as far as I know) unresolved question.

I would like, also, to contradict your somewhat pompous remark that there are probably no more than 15 engineers with a thorough grasp of the mathematics of speaker system analysis (naturally your boy is right in there around #7 or #9), since this writer has personal acquaintance with at least five. I suggest you get on with the business of reviewing audio equipment, and stop worrying about how incompetent everyone else is.

Leave that burning topic to The Absolute Sound.

Having dispensed with my technical complaints, I have a more musically oriented complaint:

I wish you would find someone more empirically well-qualified to write about records and recording than Max Wilcox, who has produced the most consistently awful sounding classical discs ever produced in this country, and therefore the world. (I refer, of course to his Philadelphia/Ormandy recordings for RCA.) Why not Kenneth Wilkinson?

Sincerely yours, Alan S. Watkins

Burroughs Corporation Pasadena, CA

(Six days later).

In my (previous) letter I pointed out your measurement error regarding L sin 8 of the Grace-707 tone arm, in "Have Tone Arm Designers Forgotten Their High School Geometry?" Since that time, I have measured (directly) L and 3 on the SME 3009 S2 Improved tone arm, and have found L (with Shure V-15 Type IT) equal 9.14" and 8 23.0° (plus or minus .25°). This corresponds to L sin 8 equal 3.57".

which is within 3% of your ideal figure of 3.68": Allowing for 1) a 1% measurement error, 2) the fact that L is in small part determined by the cartridge used, and 3) that few of us are equipped with a laser to align our cartridges with our headshells, I think this geometry can be called optimum. This, in combination with the exact geometry of the G-707 referred to in my previous letter, would seem to show that your statement "We don't know of a single commercial design in which 3 is optimum for the given L or vice versa," is based on your own faulty measurements.

Furthermore, I noticed that the latest issue of Audio contains a large ad placed by The Audio Critic, which makes use of the same misguided statement I have quoted above. This would suggest that the next candidate for roasting in "The Admonitor" is no one but The Audio Critic itself! This kind of irresponsible sloppiness on your part is causing your credibility to deteriorate rapidly in my eyes, and I am sure in the eyes of other technically aware readers. We are not all asleep out here! Sincerely yours, Alan S. Watkins Burroughs Corporation Pasadena, CA It isn't our policy to publish and refute letters presenting technical arguments that are just plain wrong, "without redeeming social value," but this one is so aggressive and righteous in tone that we've decided to make an example of it. So:

1. As long as you don't twist the cartridge in the headshell (which is what we now recommend elsewhere in this issue but isn't what you're talking about), the L sin 8 of any arm is constant and nonadjustable, even if L by itself is adjustable (i.e., if you can move the cartridge forwards or backwards, as in the Grace G-707).

Why? Because when you change L, you're simultaneously changing B. But you aren't changing-and can't change without twisting-the L sin (3. ( Eureka! You ought to apologize to your high-school geometry teacher.) We've re-measured the G-707, by the way, and its L sin 8 is nowhere near 3.68" without twisting.

2. For a fellow who's a little careless with his trig, you talk pretty good filter synthesis, and what you say about Butterworth vs. Bessel response is true-as far as you go. What you fail to mention is that a Bessel response profile has pretty grim amplitude characteristics, so that a Butterworth contour still represents the best trade-off between amplitude and time response-- which is what we meant all along.

3. Leave it to a mathematical type to derive 'no more than 15" from our "can probably be counted on two hands and maybe a foot." What's the matter, Alan? Don't they crack a few jokes around the water cooler at Burroughs?

4. Name a producer who has done better than Max Wilcox in that seventh-floor ballroom of the Scottish Rites Cathedral in Philadelphia. You think every re cording site is the Concertgebouw?

5. How far off your SME arm is, if we accept your measurements as accurate, will be indicated by the fact that the optimum offset angle for an arm length of 9.14" is 23.75° and that the optimum arm length for an offset angle of 23.0° is 9.42". Since even fully optimized geometry in a pivoted tone arm yields a fairly high distortion index at the three maxima, we don't consider that kind of deviation to be negligible.

6. Maybe you aren't all asleep out there, but at least one of you is showing the effects of insomnia.

-Ed.

The Audio Critic:

One of your subscribers sent to me for comment part of an article by Bruce Zayde, titled "A Rational Approach to Low-Frequency Speaker Design", that part of the article in which he describes transmission lines as "'non-optimal designs ..." Since we manufacture damped reflex, infinite baffle, and transmission line designs, and find that the latter are quite superior to the former on spark gap and pulse criteria, we are rather at odds with Mr. Zayde, when he says that "a better approach to producing low bass response . . . is to stick with the sealed or vented direct-radiator format"; and also because we don't know of any such designs which in fact do provide "low bass," except transmission lines.

Perhaps Mr. Zayde would care to comment directly to us, on what is "better" about his recommended "format." I understand that your magazine features measurements, rather than purely subjective reactions.

I am naturally interested in Mr. Zayde's measurements that corroborate his statements, not as a challenge to you, but based on my own curiosity; since most of the people around the world with whom I work would take my position, and base it on measurements.

Sincerely, Irving M. Fried

Fried Products Company

Philadelphia, PA

Mathematically, a transmission-line enclosure (or acoustical labyrinth) can be represented by a high-pass filter of the same order as certain types of vented (bass reflex) enclosures. Exactly the same expressions go into the mathematical model in each case, except that the transmission line has a large resistive component, corresponding to that long, heavily stuffed labyrinth. The upshot of this is that, by proper juggling of the interacting elements of either filter, vou can come out in the same place, i.e., achieve exactly the same frequency and time response. And that means exactly the same sound quality.

Mother Nature plays no favorites between networks of the same type and order. (We didn't have to consult Bruce Zayde to assert this simple truth.) The reason why the transmission-line approach is "'non-optimal' is that the resistive component dissipates a lot of amplifier power; in other words, all other things being equal, the resulting speaker will be considerably less efficient, even if not lower in fidelity.

The whole question was quite neatly disposed of by W.J.J. Hoge in his article on transmission-line speakers in the August 1977 issue of Audio, and there's an even more profound analysis in a 1975 thesis by G.S. Letts, a graduate student of Dr. Small at the University of Sidney.

When you get right down to it, the transmission-line bass enclosure has no serious academic backing; its sup port appears to be strictly by commercial and hobbyist interests.

When it comes to measurements, the question is--what exactly has been measured? Totally optimized executions of both approaches? We suspect that one reason why the transmission-line enclosure often comes out ahead in typical "as is' testing is that, by its very nature, it requires heavy bracing to be at all usable.

Makers of ordinary vented (as well as sealed) enclosures can get away with less conscientious bracing, but in that case the performance isn't optimal. Our value son why the transmission-line enclosure often comes judgment was based on what is possible in the light of available knowledge, not on what the industry is actually doing in response to audiophile mystique.

-Ed.

The rest of this column is devoted to the Mark Davis correspondence (in small type, as already explained). Enclosed with the letter written to The Audio Critic was a copy of a previous (but until now un published) letter to audio journalist Randy Tomlinson, which should be read first for full understanding of the Mark Davis syndrome, and whose text follows.

Dear Mr. Tomlinson:

Some work I had done with an under graduate named Jay Gurley on the audible differences in phono preamps was discussed on the WBUR program "Shoptalk," and apparently it was a tape of this show that you heard and commented on, indicating why you did not believe our results.

The fact that people such as yourself should disbelieve our conclusions comes as no surprise. Indeed, I did not set out to prove that overall frequency response was the only important parameter contributing to preamplifier sound quality. I firmly believed, as you apparently still do, that two preamps could have identical frequency responses, as measured from a test record with a particular cartridge, plus levels of noise and distortion below audibility, yet still sound different. The explanation most people seemed to give for this had something to do with the high-frequency performance of the preamp, where "high" means outside the audio band: 30 kHz, 50 kHz, or more. Unfortunately, none of the explanations got much more specific than that and, although test results were quoted showing measured preamp differences in these frequency ranges, the Fact remains that the ear cannot hear above 20 kHz, and that .no record cutter in existence can cut frequencies higher than 18 kHz or so anyway.

(What about cutting at half speed, as some of the more sophisticated mastering studios do? Accepting your figure, that would give a high-frequency limit of 36 kHz. -Ed.) Thus, what was missing was a clear-cut explanation of how the performance of a pre amp at 50 kHz, and high levels of 50 kHz at that, could affect something in the audio band. What I set out to do was isolate that cause/effect relationship, not prove that it didn't exist.

Having built the zippiest preamp A-B box I could imagine, I connected my painstakingly-designed Davis-Brinton Reference Standard preamp (souped-up FET input, blazingly fast slew rate, precision this, that and the other thing, ultrasharp ultra- and infrasonic filters, triple-filtered power supply, et al.) and compared it to the worst preamp I could think of at the time: the two-transistor circuit copied almost verbatim from the GE transistor manual (which actually has three transistors; the output is buffered through an emitter follower, but why quibble) and found I could not tell them apart.

And this was after I had been a party to a couple of "informal" A-B comparisons, where we listened to one preamp for 20 minutes, then disconnected it, connected an other, set levels to be vaguely the same, and found differences that were day and night. And which seemed to be repeatable.

After that initial A-B, Jay and I spent months trying to obtain the effects others seemed to be able to attain so effortlessly.

We tried substituting every link in the chain: cartridges, turntables, power amps, A-B boxes, speakers, headphones, yet the results kept coming out the same: whenever there was the most barely discernible difference, there was always a corresponding difference in frequency response. Whenever there was no measurable difference in frequency response, subjects could not differentiate audible quality. Nor did I limit the tests to naive subjects: some of Boston's goldenest ears were enticed down to the lab. (And, incidentally, I don't believe in the idea of a person with golden ears: all persons with normal hearing have perception thresholds which fall within certain known limits. A properly designed psychophysical experiment should yield identical results whether the subject has a $3000 stereo and goes to concerts or not. However, this gets into areas of things like bias, training and feedback, which I don't want to try to explore here.) Having failed to find anything relevant other than frequency response on our own, we approached some of the most vocal proponents of the other side, people like Al Foster of the BAS and Tom Holman of Advent. The listening tests they had done turned out to be informal to an extreme- which is not to knock them; they are not and do not claim to be psycho-acousticians- but when we tried to isolate differences between preamps they claimed sounded different, only frequency response again seemed to be the operative parameter. Neither we nor they could hear differences when frequency responses were equalized.

(Of course, this put Holman and Advent in a sticky position, since the success of the 300 receiver was and is predicated in large part on the preamp section. Since our tests, Advent has been extremely careful with the wording of their ads; they don't claim the preamp is the best, only as good as the best. The fact that it's also as good as the $160 Pioneer SX-450 (discount price) and a host of others is something they fail to point out. They have, however, deleted all reference to the square wave test Tom talked about in the original AES preprint paper.

(This is actually a problem with the entire industry: conflict of interest. If it turns out you can do a preamp as good as any with 2 transistors, and the public learns of this, folks like AGI, Levinson, and lots of others have a problem. So do retailers who depend on high-end sales. So do equipment review magazines, who need something to talk about each month ... Hell, not even Pioneer wants everyone to buy their

$200 receiver and ignore all their higher priced ones. This concept extends to other components as well-the industry needs things like TIM, phase response, moving coil cartridges, etc. to stay financially healthy.

Frankly, I disbelieve most of it-I think frequency response is 90% of the story. The other 10% is cartridge tracking and loud speaker radiation pattern.) (Like dynamic range is zero percent?-Ed.) Along the way, it did occur to us that long-term listening might reveal differences that A-B tests did not. Quite honestly, I think that there are some careful experiments that could be done in this area, but we did not pursue them because there is in fact al ready a small mountain of evidence to support their validity. In the parlance of psycho acoustics, an A-B test is normally referred to as a two-interval two-alternative forced choice experiment (2I12AFC). A 2AFC is simply an experiment where either stimulus A or stimulus B is presented on a given trial, and the subject indicates which he thought it was. In a 2I2AFC, both are presented, and the subject indicates whether the order or presentation was A-B or B-A.

There are other types, such as recognition experiments, in which one of a number of stimuli is presented and the subject indicates which. These experiments have been used to study a wide variety of acoustical phenomena, such as loudness perception, localization, speech intelligibility, etc., and in all areas that I am aware of, the 2I12AFC experiment results in the greatest reported subject acuity.

Our experience using the 2I2AFC paradigm in the preamp project strongly supported the validity of its use, in that whenever there was a measureable difference that fell above accepted thresholds of audibility, our subjects were able to pick it out on an A-B test. Among the differences that were audible: level differences of a very few tenths of a dB; frequency response differences over as little as an octave or so amounting again to a few tenths of a dB. At one point we attempted to use a home-brewed power amp comprised of a pair of Sanken 50 watt amplifier modules. Listening to a sine wave through them compared to a Crown 150 power amp revealed the presence of numerous high harmonics. Observation of the output waveform showed crossover distortion from the Sankens when they were connected to a loudspeaker, which was not present with the Crown. Spectrum analysis of the output con firmed the presence of the harmonics in the output of the Sankens, the absence of same in the Crown. Along the way, someone brought us an old HK tube preamp which apparently was defective. A-B testing indicated unusually high distortion, again confirmed with spectrum analysis. At one point, without informing the subjects, I hooked up an A-B which simply reversed the phase of one of the channels. The subjects, used to listening to ultrafine differences between pre amps, were almost knocked over by the magnitude of the difference. Yet it took them almost 5 minutes of guessing to figure out what I had done, despite the fact that they were well-heeled in both audio and electronics. To me, this strongly reinforces both the validity of the A-B test procedure, and the notion that even experienced listeners are notoriously inaccurate in inferring the

cause of differences they perceive.

(Incidentally, we made no attempt to control when subjects should switch, or how long they should listen before responding.

If a subject wished to hear a whole cut again before making a single response, we played it. Inter-switch intervals ranged from a couple of seconds to well over a minute.

Subjects were not told what to listen for.

Thus, it seems that there was an element of longer-term listening involved here-at least over a period of minutes.) There was one more check on the possible effects of long-term listening that 1 made.

Although it was an informal test, I found the results somewhat revealing. In a second attempt to make a flat-response/lousy-sounding preamp, I took a lowly 1458 dual 741 op-amp, put an RIAA feedback loop around each half, and called it a preamp. With a pair of 9-volt batteries the whole thing fit into a small minibox. I A-B'd it against the other preamps in the lab and, after tweaking frequency responses and input capacitances, the 741 preamp was indistinguishable, except at very high listening levels with exceptionally quiet records, where the difference in noise level was apparent audibly.

Even there, the difference was small: with a V-15 III cartridge connected, the preamps we looked at all had S/N ratios of about 82 dB, A-weighted, re 10 mV, | kHz. The 741 was 77 dB. After the A-B tests, I was still very suspicious of the results, so I took the 741 preamp home, where for months I had been listening to a good-sounding Davis Brinton unit. My first reaction after substituting the 741 was that things didn't sound right, but I couldn't put my finger on the cause, so I left it in the system, in place of the Dav-Brit. An hour later, I looked up pleased with how well the system was sounding until I realized that I was listening to the 741 preamp; then it didn't sound quite right again. This phenomenon was repeated several times over the next day or so.

As long as I didn't think about what preamp was on, the system sounded as good as ever, but as soon as I realized that the preamp consisted of nothing more than a 741, it stopped sounding good. Clearly my prejudice against the 741 was manifesting itself in very audible terms, whenever that prejudice was al lowed to function. When the prejudice wasn't active, neither were the sonic deficiencies.

Eventually, after a couple of days, I guess I became convinced of the basic goodness of the 741 preamp, and I stopped "hearing" deficiencies. I loaned this preamp to Peter Mitchell, of the "Shoptalk" show, and he subsequently informed me that he thought it was one of the better sounding preamps he had heard.

This I believe illustrates one of the deficiencies I perceive in the tests you briefly describe in your letter. Namely that the people doing the long-term listening tests know what preamps they are listening to.

(Or so it seems from your description.) I am afraid that this factor alone completely invalidates any results you might get. 1 maintain it is totally impossible to A-B knowingly a preamp made with $2 in Radio Shack parts and chicken wire to a $1200 preamp, without experiencing some measure of bias, and that this bias will manifest it self in apparent audible differences, even if the two circuits are in fact electrically identical. (You think any two resistors or capacitors of the same rated value and tolerance are electrically identical? Not so.

-Ed.) In running our tests, the subject was never informed in advance which preamps he was listening to. Nor was any feedback given after each trial; if we are talking about differences that really are hearable, you shouldn't have to tell the subject when he's right or wrong, he should tell you.

(A lengthy passage discussing the proper way to match cartridge/preamp frequency responses for A-B testing is deleted here, since the main issue has little to do with laboratory technique. -Ed.) There are a few other points in your letter that I would like to comment on.

You take note of 'certain superior characteristics of the JC-2 (imaging, definition, clarity)." While these terms may have some meaning to you, and may be valid descriptions of what you think you hear, I know of no serious psychophysical work that has been done to show these as being real, repeatable, measurable effects. The closest I can come is the concept of image fusion, from studies in localization and binaural lateralization, which is a function of how well the two channels are matched. The characteristics and thresholds of that matching have been extensively studied, and there is no preamp that I know of that comes within light years of having sufficient channel mismatch to cause detectable deterioration in binaural fusion, especially when auditioned over loudspeakers. (Never heard of cross talk, eh?-Ed.) The channel mismatches present in cartridges, records, loudspeakers, rooms, and ears far exceed those of preamps. Frankly, I suspect qualities like imaging, definition, and clarity are strictly functions of frequency response, level, and, alas, imagination. As you note in your letter, "the main problem is the psychological phenomenon that occurs in our heads. We seem totally unable to control the way the differences grow as exposure time increases." I agree with that statement in part, although I think there is a limit to how far those differences can grow. The problem is that a subject will tend to rein force a perceived difference whether or not it exists. Suppose for example that when switching from switch position A to switch position B one time, the high frequency con tent of the program material drops. The subject might begin to suspect that the preamp in the A position was brighter than the B preamp. Instead of asking himself, "Is there a difference between A and B?" he will now be asking himself, "Is A brighter than B?". The latter question shows bias, the former does not. Faced with a psychological compulsion to find a difference, the subject will tend to reinforce the perceived difference, real or imagined. If, how ever, the correspondence between switch positions and preamps is switched randomly after each trial, the subject can't always be sure that A will be the brighter one. Thus if the difference was only imagined, the subject's responses will be guesses, and he will be right only 50% of the time, providing enough trials are done. The point is that this psychological reinforcement makes it imperative that the subject not know what he is listening to, and that enough random trials are carried out to account for guessing, even in long term listening tests.

When we started out to study preamps, we realized it would be impossible to try every available preamp, power amp, turn table, cartridge, etc., in every combination, so we chose what we felt was a representative sample of equipment, making sure that everything was duplicated at least once, and that the preamps covered as wide a range as possible. Nevertheless, it comes as no surprise to have someone claim that the reason we couldn't hear differences was that we didn't use thus and such a cartridge, arm, power amp, speaker, whatever. Frankly, I don't find the argument terribly compelling.

In the first place, the experts out there can't seem to come to any consensus as to what the finest equipment is. You claim we should have used a Dyna 400 or an Ampzilla, but I seem to remember reading some other review that said the Phase 400 was the hot setup.

Then I think there was another one that said that the BGW, which we did try, was the only way to fly. In the second place, nobody ever gives a solid cause-effect reason why their particular super product is alleged to actually sound better than the other guy's.

(You say your preamp has .0008% distortion? Why should that sound better than their .01% preamp when the threshold of audibility is .2% under the best conditions, and when the cartridge is producing 4%?) Not even Mark Levinson can tell you why his preamp is supposed to sound better in explicit terms.

In the third place, as I've mentioned, whenever there was a certifiable difference, in level, frequency response, distortion or phase, the subjects heard it-and heard it as soon as it exceeded recognized thresholds of audibility. On all of the test setups.

(But earlier in this same letter you bracketed phase response with industry hypes like TIM and moving-coil cartridges.-Ed.) In short, Mr. Tomlinson, you seem to be asking me to believe in magic. You claim there are preamps out there with conventional specs-things like frequency response, distortion, noise, input impedance-identical to the "finest" preamps, which nevertheless are audibly inferior, but you do not tell me what these inferior preamps are doing to the signal to screw it up in an audible fashion.

You claim the audible differences start out as barely audible in A-B tests, but become more audible with casual long-term listening, without explaining why. You claim these differences are audible only on "state of the art" equipment, without providing an explicit definition of what makes a given piece of equipment state of the art. Having failed to specify the magical ether that seems to make it through the good preamps, and not the bad, you are obviously not prepared to say how it gets through "good" amps, tone arms, cartridges, and speakers to become imaging, definition, clarity, what have you, upon striking our ears.

About 12 months ago I set out to try to find that ether. I have been trying assiduously to track it down, but somehow it always seems to hide behind a tree when it sees me coming. I'm beginning to think it unfair that I should be the only one to have to carry the burden of proof. (Very interesting. Now tell the doctor when you first began to experience this feeling of isolation.-Ed.) Be fore too many more people are coerced into blowing twelve hundred bucks on mystique, I think it would be nice if someone showed in clear, irrefutable repeatable terms exactly what those twelve big ones buy in audio quality, and how that can't be just as easily bought for $17.50. In other words show, using controlled, blind listening tests, short-term or long, that a difference really exists, and then provide a clear cause-effect relationship that explains the origin and behavior of those differences. Then justify why those differences should cost as much as they seem to.

After some record producer has tossed equalizers, filters, limiters, compressors and reverb into the signal chain with gay abandon, I don't see why I should spend 5 times more for a preamp just to get an RIAA response that's 0.2 dB more accurate at 20 kHz. At the very least, if there are differences that can only be discerned on "near state-of-the-art" equipment, I think your readers should be so informed, so they won't waste large sums of money on a preamp if they happen to own AR-11 loudspeakers, or other equipment you deem not near state of the art.

I am willing to have my claims proved incorrect, Mr. Tomlinson, but I don't believe you have done so.

And now for the Mark Davis letter written directly to us, in response to our preamp survey.

The Audio Critic:

I was recently given the opportunity to read an article published in your premiere issue which discussed at modest length certain research concerning RIAA phonograph preamplifiers to which I was a party. The principal finding of this research was that competently designed phonograph preamplifiers, ranging in complexity from the costliest, most abstruse topology, to the simple and popular two-transistor configuration, are in fact audibly indistinguishable under any conditions of normal operation (i.e. reproduction of commercially available records by conventional magnetic pickups), providing the concept of competence in design subsumes the equivalence of overall playback frequency response to a tolerance of 0.3 dB or better.

(Moving-coil pickups are also magnetic, we subsume?-Ed.) It is possible that certain misconceptions may have been promulgated by my brief discussion of the research on the WBUR radio program, "Shoptalk," which you indicate as the source of your information. To the extent that you find it enlightening, I am enclosing a copy of a letter I have previously sent to Randy Tomlinson, of Accusound Engineering and StereOpus magazine, which treats in somewhat greater detail some of the questions at issue.

There are a few other questions your article raises that I would like to attempt to clarify. These generally involve electrical measurements or listening technique.

I concur with your conclusion that neither harmonic nor IM distortion is a significant factor in modern preamplifiers. Sometimes (too often) the opposite conclusion is reached by driving the preamplifier being evaluated with electrically generated test signals to levels hopelessly greater than will actually be encountered by the preamplifier under actual conditions of use (cf. Jung, 1977).

Obviously, I also take no exception to any reservations you may have about tests employing square waves, wideband noise, or other similar, unrealistically fast test signals. To the best of my knowledge, Tomlinson Holman no longer believes his square wave test to be relevant; nor does Alvin Foster endorse his white noise test (Foster and Swanbon, 1977).

It is gratifying that you find a pre amplifier with relatively low slew rate (1 volt per microsecond) can deliver excellent sound.

Our calculations indicate that such a figure is at least an order of magnitude larger than what is required by available records and cartridges. An interesting sidelight is that slew rate figures comparable to 1 V/uS are obtained from typical well-designed preamplifiers employing the two-transistor topology.

Your results on signal-to-noise ratio, that some units were 'measurably and audibly quieter than others," and that some were

"outstanding" (p. 8), runs somewhat counter to our own measurements, which showed fairly strong consistency among the units tested.

This may be due to differences in measurement technique; regrettably, you do not discuss the protocol you employed.

As you may know, the noise of a pre amplifier is lowest when the input is terminated with a short circuit. In trying to pre sent the most attractive-looking figures, manufacturers will often specify the signal to-noise ratio of their preamplifiers under this constraint. Occasionally, one will even optimize for it, to the detriment of the conditions of actual use; i.e. with the input terminated by a magnetic cartridge. Hallgren (1975) has shown that the principal source of noise in a preamplifier is in fact the cartridge, followed in importance by the 47K ohm input termination resistance.

Since it is the normal mode of operation of the preamplifier, terminating the input with a cartridge while measuring its noise seems to us to be the most realistic approach.

True, the result will depend to some degree on the cartridge chosen; nevertheless it will be more representative than the shorted in put specification. The latter is probably more apropos when a low impedance device, such as a pre-preamplifier, is connected to the preamplifier. Even then, it is of secondary interest, since the cartridge and pre-preamplifier input noise will inevitably dominate, to be masked in turn by record noise. The need to terminate the input with a cartridge applies as well to the measurement of the input noise of a pre amplifier, incidentally.

We used a Shure M9IE cartridge; while its sonic qualities may not be state of the art, its inductance and other electrical characteristics are typical of a large number of current state-of-the-art cartridges. With the cartridge mounted in a shielded box to suppress stray hum fields, and the stylus re moved to avoid acoustic pickup, we found almost all preamplifiers had a signal-to noise ratio of 82 dB, ASA 'A' weighted, re: 10 mV RMS, 1000 Hz. This is in good agreement with Hallgren's data.

We checked these measurements by listening to the noise of the preamplifiers, after first taking care to match levels with a test tone, with the respective inputs terminated in turn with the same cartridge. While there were clear variations in the spectrum of the noise among units examined, the overall level remained essentially constant.

Basically, the same laws of solid-state physics and thermodynamics hold for all hi fi companies. For about 10 cents you can buy, in production quantities, a transistor that will perform as well as possible with available cartridges. While there are some games you can play with the 47K input resistance, only small gains beyond 82 dB are possible, and even the least expensive receivers appear to hit that figure routinely. We have thus concluded that the concept of an ultra-quiet phono preamplifier is largely a myth. ( What about small-ly? That's our bag. -Ed.) One of the largest areas of disparity among preamplifiers was found to be input impedance, so I was a little sorry to see that you failed to measure it. The use of a low-inductance cartridge probably minimized the effect in your tests, but careful measurement of relative frequency response, including the effects of the cartridge, as described in the accompanying letter, would have been preferable to assuming the input impedance had no effect. Furthermore, not everyone uses the cartridge you employed;

many popular cartridges are sensitive to variations in input impedance, raising the possibility that a preamplifier that sounded fine in your tests would deliver inferior sound with a cartridge having greater inductance;

ergo the importance of reporting this specification.

(We don't know of a single high-inductance, i.e., input-capacitance-sensitive, cartridge that can compete sonically, regard less of loading, with moving-coils like the Denon or EMT, or low-inductance moving fields like the Grado. We don't recognize the *'rights" of low-fidelity cartridges; maybe that's why we hear preamps differently, too.-Ed.) Of greater concern to the accuracy of your report than the results of the electrical tests are the psychophysical issues implicit in your listening protocol.

Your measurements of preamplifier frequency response show unit-to-unit variations on the order of 0.75 dB, a tolerance which you claim is "good enough to eliminate the possibility of listening preferences on that score" (p. 7), further commenting that requirements of finer tolerance are to be considered “obsessions" (p. 8). As noted above, our investigation indicates the need to equalize overall playback frequency response within a tolerance of 0.3 dB. That is, the cumulative effects of input impedance, frequency response, and playback level must not induce a difference greater than 0.3 dB in level between the two preamplifiers being tested, at any frequency in the audio band.

Preferably, the difference in overall level, weighted on an octave basis, should be zero.

This constraint was not imposed capriciously; it resulted from concrete psycho acoustical data. Using the same AR speakers you think so little of, and the same A-B paradigm you denigrate without quantitative justification, we found subjects could reliably discriminate between sources solely on the basis of frequency response disparities of as little as 0.7 dB. By limiting differences to a maximum of 0.3 dB, frequency response/ level discriminability was found to be virtually eliminated, permitting accurate acoustical evaluation of other possible disparities. The tolerance of + 0.75 dB you encountered implies a worst-case deviation of 1.5 dB, which, when added to the tolerance of the level match you made, 1.0 dB, results in a maximum possible deviation of 2.5 dB, not counting the possible effects of input impedance variation.

Your plausibility arguments on pages 7,8, 11 and 12 notwithstanding, this difference is, by actual empirical observation, well be yond the point of becoming plainly audible.

Your failure to carefully measure and control playback frequency response/levels (casual tweaking with tone controls does not constitute careful measurement and compensation) makes it possible, nay likely, that your judgments primarily reflect preferences in frequency response. Of course, it is our contention that this must be so, as there is no other audible variation in preamplifiers;

but even if there is, it is likely that it was swamped out by the potentially large variations in frequency response and levels.

You dismiss the AR loudspeakers as "too forgiving" (p. 10), without citing any references to back up this claim, or stating how one goes about measuring the "'forgiving ness" of a loudspeaker, or what specifications contribute to this alleged quality.

(References? If authoritative criticism of speaker systems-by brand name!-were available in references, we wouldn't be in business. -Ed.) We acknowledge that there are loudspeakers which have flatter frequency response than the ones we used, but this is not a necessary prerequisite, since the same speaker is used for both A and B presentations. What is important is sufficiently wide response, without serious gaps, to ensure that all audible frequencies are reproduced at hearable levels, plus sufficiently low distortion to avoid masking possible distortion from the preamplifiers being auditioned. (But what kind of distortion? Steady-state or time dispersive?-Ed.) Both of these criteria are met by the AR speakers. We have measured the IM distortion of a number of commercial tweeters and found it in all cases to be under 0.5% at a level of 100 dB SPL, 6 feet on axis, the distortion dropping rapidly below this intensity. This is less distortion than the ear itself produces at such levels. (Yup. You meant steady-state. -Ed.) The most relevant justification of the speakers and associated equipment we used was that subjects were able to clearly discriminate between sources whenever there was a certifiable difference exceeding accepted thresholds of audibility, as for example in the case of frequency response/level deviations of only 0.7 dB (which you believed to be inaudible). This fact, plus the strong consistency in results with alterations in experimental equipment, including the substitution of Koss ESP-9 electrostatic headphones for the loudspeakers, lends fairly strong empirical credibility to the results we obtained.

(Hold everything! We never said that differences of 0.7 dB are inaudible, but rather that they aren't relevant to our conclusions. As for the electrostatic headphones, did you use them when you A-B-ed the two transistor GE preamp against the Mark Levinson? Right then and there? -Ed.) Lacking a proven alternative mechanism to account for differences in preamplifier sonic quality, your allegation that our equipment fails to preserve the effects of this un known mechanism would seem somewhat pre mature at best.

The rationalization you advance for eschewing the use of an A-B test is that it fails to show up differences that are really there, because "your ear tends to integrate the sound under these conditions" (p. 11). This position is simply untenable; there is no such integration effect to mask differences.

As explained in this letter and the one accompanying it, our A-B tests revealed extremely fine differences in level, frequency response, and distortion. Beyond that, numerous diverse disciplines in psychophysics make use of the instantaneous two-interval A-B discriminability paradigm, usually to explore the highest levels of human resolution.

I know of no authoritative, controlled study showing long-term listening as you describe to be a more accurate indicator than A-B testing; if you know of one, please cite it. In the meantime, I commend to your attention the Handbook of Mathematical Psychology (1963) as a viable primer on the subject of competently conducted listening tests and their analysis.

The attitude toward statistical testing is perhaps one of the most contradictory elements of your article. You claim, without providing justification, to have "very little faith in statistical surveys ('19 persons preferred A and 22 preferred B')" (p. 10), implying incorrectly that a preferential methodology was employed in our tests. Having made this claim, you then execute an about face, and perform just such a preferential survey.

Frankly, I completely agree with your lack of faith in such surveys, unless the basis for listener preferences can be objectively isolated, and unless it can further be shown that the preferences expressed represent a fairly universal consensus. Otherwise, it is indeed true that one may "come to a different conclusion with a different selection of listeners" (p. 10). You have made no attempt to show quantitatively that your preferences would remain the same with a different, independent group of listeners. Indeed, you have not even shown that your preferences are independent of the equipment chosen; to the contrary, the single substitution you did make, of the power amplifier, reversed the order of preference of the highest ranked preamplifiers.

(That's right. We were trying to help our subscribers choose a good preamp, not to write a textbook on psychoacoustics. -Ed.)

Almost without exception, the human auditory system responds in a stochastic fashion; present a stimulus to a subject at two different times and, in general, you will get two different responses. Such behavior makes the need for accurate statistical characterization manifest.

For the record, we never asked a subject which preamplifier was preferred, only whether there was a discernible difference.

Subjects were given no feedback or prior information that could cause bias; the sound of the preamplifiers was the only information available to the subject on which to base a decision.

The probabilistic nature of the results is reflected in the fact that most people could discriminate frequency response disparities of several dB most of the time; some people could discriminate frequency response disparities on the order of a major fraction of a dB some of the time; and no one could reliably discriminate disparities of 0.3 dB or less. Allegiance on the part of the subjects to a particular concert hall or type of music had no effect on the results. Toole (1977) has recently presented some data indicating that experienced listeners ("golden ears™) are not even required in preferential studies; naive listeners can be just as consistent.

In summary, your listening tests incorporated some very serious lapses of accepted psychoacoustical procedure: the failure to equalize for differences in level and frequency response; the failure to keep subjects from comparing opinions, despite the need for statistical independence; and the failure to conduct independent blind trials, without feedback, prior information, or other sources of bias, in sufficient number to en gender reliable statistics.

Even long-term listening tests, for which you express preference, can A done properly, providing the same rules of technique are observed. As it stands, I doubt that any competent psychophysicist would endorse your methods or accept your conclusions.

Lurking behind these issues is the conceptual question of whether it makes any sense to test preamplifiers when you profess not to know what makes them sound different. For one thing, the perceived superiority may not be worth the extra cost of the better units; it may be possible to incorporate the extra "goodness factor" which you claim the better preamplifiers possess into a cheap preamplifier for next to nothing. For ex ample, if you assume for a moment that our contention is correct, that frequency response is the only significant parameter, then it is foolish to recommend an expensive unit when an inexpensive one will either have the same response, or can be adjusted to have it by a suitable choice of input capacitance.

Without knowing the specific mechanism(s) involved, you cannot be sure that your conclusions are true in general for other installations. Again, as an example, if frequency response is all important, a revised choice of associated reproduction components would very likely have altered the preferences. As already noted, switching Just the power amplifiers changed the order of the two best-ranked preamplifiers. Without an alternative mechanism, clearly identified and characterized, you cannot prove that your preferences will hold for anybody else.

It is significant that the results and conclusions we reached have also been obtained by the "highly active and enlightened" (p- 9-here, we agree completely) Alvin Foster, of the Boston Audio Society, who formerly espoused the relevance and use of tests employing very high-frequency signals, such as wideband noise. Some of "these results will be published shortly (Foster and Swanbon, 1977).

The same conclusion was also reported to me by David Griesinger (1976), of Harvard University and the Boston chapter of the Audio Engineering Society. The phenomenon thus does seem repeatable.

Hopefully, this information will help clear up some of the misunderstanding that seems to exist. I know from personal experience that the results of informal listening tests often seem so obvious as to make a more precisely controlled environment appear unnecessary. More often than not, however, the impression is a false one. Be fore rejecting accepted psychophysical experimental protocol, you are urged to try it.

Sincerely, Mark F. Davis

Massachusetts Institute of Technology Cambridge, MA

References:

Davis, M.F. (1977). "A Simple Impedance Bridge," The B.A.S. Speaker ( Box 7, Boston, MA 02215), 5, 7, 3-8.

Foster, A., and Swanbon, W. (1977). "A Comparison of Low Priced Receivers," submitted to The B.A.S. Speaker. (Published July 1977. -Ed.)

Griesinger, D. Private communication.

Hallgren, B.I. (1975). "On the Noise Performance of a Magnetic Phonograph Pickup," J. Aud. Eng. Soc., 23, Sept. Jung, W. (1977). "Slewing Induced Distortion in Audio Amplifiers: Part 1," The Audio Amateur, 8, 1, 3-9.

Luce, R.D., Bush, R.R., and Galanter, E. (1963). Handbook of Mathematical Psychology VI (John Wiley and Sons, Inc, New York).

Toole, F.E. (1977). "Speakers--Part I," AudioScene Canada, March, 19-28.

Our overall reaction to these two letters is that, even though they fly in the face of all the slowly and painstakingly accumulated wisdom of an entire generation of audio perfectionists, they mustn't be dismissed out of hand. Obviously Mark Davis and his associates are highly knowledgeable, dedicated and sincere. What they claim they heard is undoubtedly what they actually heard-and quite possibly, had we been there, we too would have heard the same thing. The question to ask is why it sounded that way.

The most plausible explanation is that the system into which the preamps under test were inserted had insufficient resolving power. Our experience is that plugging different preamps into, say, a Beveridge Model 2SW system makes the differences laugh ably obvious, regardless of matched listening levels or RIAA equalization accuracy. (The signal source being a Denon DL-103S cartridge, fanatically aligned for optimum tracking laterally and vertically.) Through Dahlquist DQ-10's, used in our original survey, the differences are still easily audible though less flabbergasting. Through the Mark Davis setup-who knows? We have a feeling that total disregard for the masking effects of time dispersive distortion-originating from the cartridge and the speakers-and exclusive concentration on amplitude response and steady-state distortion channeled the Mark Davis tests in a self-fulfilling direction.

We say self-fulfilling because we also have the distinct impression that not only Mark Davis but the entire Boston Audio Society crowd (delightful and sophisticated practitioners as many of them are) still would like to discover an audio Nirvana for forty nine dollars and ninety-five cents. We discern in them a peevish hostility toward ultrahigh end equipment, perhaps because so many of them are academics and empathize with music-loving students who can't afford a Mark Levinson ML-1. And, of course, the best way to dismiss the expensive stuff is to find it scientifically indistinguishable from the run of the mill.

We must also remind our readers, to put things in proper perspective, that Mark Davis also considers TIM in power amplifiers to be a myth. Get the picture? It's not only the emperor who is naked, but also the empress, the crown prince, the archduchess . . .

-Ed.

---------

[adapted from TAC, Vol.1, No.4: July/August /Sept 1977]

---------

Also see:

Box 392: Letters to the Editor (V1-2, March/April 1977 Volume 1, Number 2)

Audiophile Loudspeaker Systems (Mostly Ribbon) from $1576 to $7000 for the Pair [Issue No. 11 Winter/Spring 1988]

Cartridge, Arm and Turntable vs. the Groove: Who's Winning?

Various audio and high-fidelity magazines