The debate on Bob Carver's transfer-function duplication (t-mod) claims, techniques and results, originally started by a 1983 preprint of the article on the preceding pages, has recently been rekindled by his M-1.0t power amplifier and is eliciting some characteristically undisciplined and inane commentary in the "alternative" audio press.

It is assumed that readers of this article have read and digested our original, notorious Carver t-mod article, either in the preprint form that has been circulating since January 1983 or as reproduced in this issue. Without a reasonable grasp of what transfer-function duplication is and what it is not, the recently renewed controversy about Bob Carver's favorite sport is devoid of meaningful content.

Aftermath of the first Carver t-mod.

Soon after the January 1983 Consumer Electronics Show in Las Vegas, where Bob Carver first announced the "Levinsonized" t-mod of his top-of the-line amplifier and started to circulate our article, the outraged yelps of the high-end cultists began to be heard. The hue and cry was of a different sort, however, than we had expected; we had obviously overestimated the intelligence of certain sectors of the audiophile press. We had expected attacks on (1) the very concept and feasibility of transfer-function duplication, (2) the validity and completeness of Bob Carver's methodology, (3) the theory and practice of our blind A/B tests, and possibly (4) the lack of corroboration by independent ob servers outside the project. We had airtight arguments against all such attacks, but they never took place.

Instead, the press reaction consisted mainly of moral indignation. Their message was that it is not ethical to mimic the sound of somebody else's amplifier (why can't Bob Carver come up with his own sound if he is such a genius?) and that it is similarly unethical to circulate an article that has not appeared yet in the publication it was written for. We were flabbergasted to see the audio pundits miss the main point, namely that there is no inevitable connection between the ultrahigh-end sound and the ultra high price, and we were tempted to rephrase Samuel John son's famous aphorism to read "moral righteousness is the last refuge of an incompetent critic." (If Adolf Hitler stated on a rainy Tuesday morning that 2 + 2 = 4, then on that subject the morally warped, inhuman monster was 100% right and cannot be faulted by an intelligent observer. As for the ethics of copying a famous amplifier's sound, is it equally reprehensible to copy a top home-run hitter's stance at the plate or a beautiful movie star's shade of red hair or Henry Kissinger's accent? Maybe we should have assembled an interfaith committee consisting of a priest, a minister and a rabbi to give us pastoral guidance before we braved the moral complexities of the t-mod project.) A couple of the aforementioned lightweight reviewers did end up auditioning the production version of the Carver M-1.5t and reported that it sounded okay but not nearly as good as the Mark Levinson ML-2. Was the latter available to them for side-by-side comparison? Nope. Their exquisite hearing and phenomenal aural memory made that small in convenience unnecessary. More about one of these gents further below.

Our own follow-up of the M-1.5t story was not particularly thorough, since The Audio Critic had sunk into limbo, but we did have the opportunity to examine two or three samples of the production version in 1983 and early "84 that satisfied us as being exact duplicates of the hand wired t-mod Bob had left with us in our laboratory. Considerably later, in mid-1984, we came upon a sample that did not sound right, but we were unable at the time to investigate what was wrong with it or whether it was in any way typical of Carver's then current production. It is possible for the design parameters of an amplifier to "migrate" after months or years of manufacture if they are not rigidly con trolled against a calibrated prototype (or in this case the ML-2 standard), but we have not the slightest evidence that such was ever the case at Carver and, frankly, we doubt it.

More about that subject, too, further below.

Segueing now to 1985-87, we come to the currently unfolding chapter of the Carver t-mod saga. As many of our readers know from other sources, Bob has done it again and

this time with an added touch of virtuosity, using a very high-end tube amplifier as his reference and copying its transfer function into a Carver solid-state amplifier.

A Conrad-Johnson in Carver's clothing.

In 1985, Bob Carver made the same challenge to Stereophile magazine as he had made to us three years earlier ("give me an amplifier, any amplifier at any price," etc.). He did not mention to them that he had already done the whole thing once before for The Audio Critic and, astonishingly, the Publisher, Editor and staff of Stereophile had never heard of the first Carver t-mod project or of our article, de spite the preprint with a circulation well into six figures and the ensuing press commentary. Those golden ears were apparently not kept to the ground. No wonder, then, that they took up the challenge as eagerly as we had, and luckily they introduced an interesting new wrinkle by designating the Conrad-Johnson Premier Five mono tube amplifier ($3000 each, $6000 a pair) as Bob's target.

An 8-page article by J. Gordon Holt in the October 1985 issue (Vol. 8, No. 6) of Stereophile relates in detail what happened next and provides excellent insight into the high-end audiophile subculture's primal anxieties on the subject. First of all, the policymakers of the magazine decided to withhold the identity of the target amplifier; they would not even divulge that it was a vacuum-tube unit with an output transformer or that it cost more than ten times as much as the second-from-the-top Model M-1.0 that Bob selected from the Carver line to "Conrad-Johnsonize" into the M-1.0t. We soon found out from our own sources that the amplifier they used was the C-J Premier 5 (we try hard to keep our golden ears to the ground), but the hemming and hawing and rationalizations in the article to explain away the concealment of that enormity are quite depressing. It all comes down to their tacit but clearly evident belief that the whole truth is bad for business. To compensate for such equivocation, the article addresses with ecclesiastical gusto and without the slightest intellectual embarrassment the nonsense moral issue of mimicking other people's sound.

In all fairness, we must go on record here that, in our opinion at least, Gordon Holt has one of the keenest ears in the business, understands the technical aspects of audio quite thoroughly and is 100% honest. Those parts of the article that are politically untrammeled, such as the account of Bob Carver's t-mod procedures and the reporting of the listening tests that followed, can be unhesitatingly taken at face value. Our own, earlier article was a somewhat more rigorous tutorial on the physics and logic of transfer function duplication, but Gordon's story is basically the same: Bob toiled and tested, there were a few temporary set backs, but in the end a -70 dB null was obtained between the two bridged amplifiers (in our case it had been -74 dB) and after long-suffering back-and-forth comparisons the two were found to be "sonically identical." We could say, "What else is new, Gordon, baby?" but in fact there were some significant differences between the two t-mod projects and they are worth noting.

To begin with, tube-to-solid-state transfer-function duplication is undoubtedly a greater feat than solid state to solid state, as we already speculated in our first article, and we feel that Stereophile deprived Bob Carver of full credit on that count by concealing such a key element of their project. We have always maintained that a truly competent designer who understands tubes and transistors equally well can in a given situation achieve equal black-box results with them, although with very different techniques inside the black boxes. Here was a perfect test of that tenet, but apparently it was impolite to bring it up.

On the other hand, the Conrad-Johnson unit has its own characteristic shortcomings that limited the scope of the exercise; there is no mention in the Stereophile article of the input/output null test that figured prominently in the first Carver t-mod project because the results with the C-J would have been quite poor. The output of this amplifier, with its vacuum-tube circuitry and output transformer, does not resemble the input closely enough to produce an impressive input/output null; the amplifier is in effect a mild signal processor rather than "a straight wire with gain" and preferred by certain audiophiles for that very reason. (See the power amplifier recommendations in this issue for our own views on the subject.) Gordon Holt reports that in the penultimate phase of the transfer-function duplication the two amplifiers sounded absolutely identical, except that the Carver unit had better bass definition; Bob had to muddy up the bass of the modified M-1.0 a little bit to make it sound totally indistinguishable from the Conrad-Johnson! Not surprising when you take into account the difference between direct coupling and an output transformer, but too damn bad-because the production version of the Carver M-1.0t incorporates the exact C-J Premier 5 transfer function, warts and all. Bob would not have it any other way; he wanted to prove a point, not give the world a better Conrad-Johnson.

The most amusing difference between the two Carver t-mod projects, however, is that the success of the first gave us a great deal of intellectual gratification, since it was a celebration of the laws of physics and the dictates of common sense, whereas the Stereophile people were obviously and miserably unhappy about the success of the second, perceiving it as a deadly blow to the high-end mystique. Every paragraph of Gordon Holt's article exudes a feeling of we all-wish-it-weren't-true, but he is honest enough and hears well enough to admit that it is indeed all true. With all that resistance to simple truth (not so much by Gordon, as we read it, but by the front office and their cohorts), something eventually had to give, and that brings us to the current state of the Carver amplifier controversy, which is the main reason for this follow-up article.

It's bad for high-end audio, therefore it isn't true.

A long letter by our old friend Harvey Rosenberg (the Tube God of New York Audio Laboratories) in the February 1986 issue (Vol. 9, No. 1) of Stereophile sets the tone of subsequent commentary on the subject. Harvey is a with-it guy and immediately points out that The Audio Critic was the first to do a t-mod project with Bob Carver, but then he goes on to state that the production Carver M-1.5t does not sound like the Mark Levinson ML-2, without making it clear (perhaps deliberately) whether or not this perception is based on an A/B listening comparison. He questions Bob's "reformation" as reported by us, brings up once again the ethics of copying somebody else's sound and circulating an unpublished article, challenges Bob to "give us your [own] state-of-the-art stuff if you've got it," and expresses strong doubts about the feasibility of duplicating any experimental Carver amplifier in production. Most significantly, his deepest fears are betrayed when argues that an inexpensive Bob Carver clone of one his tube amplifiers, if it could be done, would actually be good for business. Talk about Don Giovanni inviting the statue to dinner...

A brief answer to a reader's letter in the January 1987 issue (Vol. 10, No. 1) strongly hints at the emerging new Stereophile party line: the production Carver M-1.0t does not appear to sound like the laboratory prototype; tests are under way. The denouement comes in the April/May 1987 issue (Vol. 10, No. 3), where some 14 pages are devoted to the production M-1.0t, including reports by Gordon Holt, John Atkinson and Publisher Larry Archibald, plus a lengthy reply to these by Bob Carver. The gist of the re viewers' message is that the production M-1.0t is sonically distinguishable from the original target amplifier, with some difficulty according to Gordon, easily according to John and necessarily according to Larry. They now admit that the target amplifier was a top-of-the-line vacuum-tube unit (without revealing that it was the C-J Premier 5), which gives Bob the opportunity in his reply to nail them on fundamental procedure.

What happened was that the reviewers compared only two amplifiers, the production M-1.0t and the original C-J, when the scientifically correct procedure would have been to compare three amplifiers, the third being the original hand wired Carver t-mod. The latter had been found one and a half years earlier to sound indistinguishable from the C-J, with a -70 dB null on the bridging test, whereas the production Carver appeared to sound at least slightly different and nulled only an average of -28 dB against the C-J. Even a novice equipment tester should have concluded at that point that there were two distinct possibilities: (1) the production Carver was not a perfect duplicate of the hand-wired proto type or (2) the C-J amplifier was not a perfect duplicate of its older self because of tube aging over a period of one and a half years, a widely known phenomenon. Nothing would have been simpler than to test both possibilities, since Bob had sent the prototype back to the reviewers, but they never touched it, preferring the self-fulfillment of their prophecy of production variations. Bob quietly seethes over this in his letter but remains a gentleman; Larry Archibald has the last word in a postscript that admits the omission and then makes light of it, not very convincingly.

Two issues later (August 1987, i.e., Vol. 10, No. 5), Stereophile tries to retire the subject with six letters from readers and a final editorial summation by John Atkinson.

One letter is a foul exudation of petty ill will, ignorant disparagement, and provable untruths (who let that one past the wastepaper basket?); the other five are reasonable but not always to the point. John Atkinson does a choplogic postmortem on the key issue of the two-way versus the three-way comparison, takes a gratuitous potshot at Bob's trustworthiness, and then with perceptible irritability declares the matter closed, with Carver advertising thenceforth barred from the pages of the magazine "in response to Carver pressure with respect to the editorial content of Stereophile." From reluctantly acknowledged wizard to banished undesirable in 22 months-such was the meteoric rise and fall of Seattle Bob in New Mexico.

All this, in our modest opinion, added up to pretty lightweight audio journalism, quite lacking in credibility no matter what the ultimate truth turns out to be, although we are certain that the project was originally undertaken by the magazine in good faith. Somewhere along the line some body got cold feet, the raison-d-étre of the high-end business appeared to be threatened and a cop-out mentality crept in.

Meanwhile, in the March/April 1987 issue (Vol. 12, No. 46) of The Absolute Sound, the staff member listed as Senior Editor, one John Nork, also reviewed the production M-1.0t. By sheer coincidence, he happens to be one of those prodigies of aural memory referred to above who in 1983 found the M-1.5t inferior to the Mark Levinson ML-2 without needing to listen to both at the same time. By even more remarkable coincidence, he is a former Carver dealer from Indianapolis with a widely reputed grudge against the Carver Corporation. In his M-1.0t review, he magnanimously admits the theoretical possibility of duplicating a highly dissimilar amplifier's sound in the laboratory but then goes into the familiar song and dance about the hopelessness of doing the same in production. Although he, too, seems to have found out that the target amplifier was the Conrad-Johnson Premier Five, he makes no attempt at a comparison, contenting himself with some approved clichés about imaging and soundstaging, after which the M-1.0t emerges as a surprisingly decent amplifier for the money ("substantially better" than the M-1.5t-again strictly from memory) but not really world-class.

We could walk away from this sort of subjective expertizing without comment, since it is exactly on the level of suburban restaurant reviewing, but there are some bitchy remarks in the review about The Audio Critic, obviously believed by the author to be mute and incapable of retaliation after all these years, which prompt us to disabuse him of that belief. He refers to our 1983 article on the first Carver t-mod as "the last great journalistic burp from Peter Aczel" and, having established himself with that polished phrase as a literary stylist and arbiter elegantiarum, he goes on to characterize the long subhead of that article and another sentence from it as "some of the most narcissistic prose ever to adorn an audio periodical." It appears that John Nork is incapable of recognizing the posture of tongue in cheek (at least when it is your own tongue in your own check) and is innocent of rhetorical devices such as hyperbole and When you buy an amplifier for many thousands of dollars, what you pay for is not the sound. You pay for status, pride of ownership, cosmetic appeal, possibly mechanical rugged ness and long-term durability, in some cases easy servicing and tender loving care by the manufacturer, but not what you actually hear. That is achievable for a lot less money.

Such a conclusion, inevitable as it is in our opinion, should not be hysterically interpreted as undermining every article of faith of the high-end community. Judging from the level indignation in some circles, even mounting to un disguised hatred here and there, Bob Carver is perceived as sending a message of "Who needs you?" to designers and manufacturers of very high-priced amplifiers and "Sucker!" to their customers and admirers. That, of course, is the rankest nonsense. Before there was a Carver M-1.5t, there had to be a Mark Levinson ML-2; before there was a Carver M-1.0t, there had to be a Conrad-Johnson Premier Five; the order could not have been reversed, and there also had to be some brave and well-heeled purchasers of the high-end product, otherwise the issue could not even have arisen. Bob is the first to admit all that. Equipment of this caliber always originates as a concept in sound, valid or not, without any restrictions on the cost of implementation; the problem starts when the makers and owners of the equipment stake off some sort of exclusive claim to that sound by virtue of the price tag. Bob's real message is that there is no such exclusivity because nobody owns, and money cannot buy, a transfer function-and the transfer function is the sound.

So where's the review? One more thing. Since the Conrad-Johnson Premier Five is not our cup of tea, having been engineered as a subtle signal processor to suit certain tastes rather than as a totally neutral conduit, we are not particularly interested in reviewing the Carver M-1.0t. To us it is primarily the proof of a principle, although to lovers of the "tube sound" it appears to be an almost unbelievable windfall. Let them listen to it, argue about it and have fun with it.

-------

[adapted from TAC, Issue No. 10]

---------

Also see:

Box 392: Letters to the Editor

Various audio and high-fidelity magazines