Not exactly, he altered some aspects (probably FR and probably output R) so these amps could not be told apart in a level matched blind test.
Not exactly. At least not exactly as far as Bob's methodology went. And at least not as far as a
level-matched blind testing goes. Of course, in the 'final analysis' what you suggest is certainly the case.
I'm referencing Bob's original (first) 'Carver Challenge' (not the subsequent
Stereophile fiasco) conducted in 1982 at Peter Aczel's lab. In this first setup, Bob's stated goal was to match the transfer function of his (then) new M1.5 amplifier to Aczel's ML-2 class A mono unit. The idea was to 'null' the two amplifiers so their output signals would cancel each other, thus proving that two amplifiers of dissimilar topology could effectively have the same input to output transfer function. I'm not sure that FR was ever measured as a working or deciding criteria. The test report (
Audio Critic 10) indicates that an oscilloscope and voltmeter were used. However it was, the mod took two days of work.
According to Aczel:
In the process, the open loop of the (Carver) was retailored, the negative feedback was readjusted several times, various transistors were rebiased, the input impedance characteristics were changed, lower-resistance wiring was substituted in crucial places, but no active devices were replaced and no changes in basic topology was made. German polypropylene capacitors were used whenever new values had to be inserted.
Aczel reported that except for a 'miniscule' ground loop hum, the amps nulled at -74dB (Bob's actual production 't-mod' amp was claimed to be nulling in the mid 50s against the ML-2). One thing: during listening sessions program material and playback level had to be carefully selected because of the Levinson's limited power output (35 watts v 350 watts), which could easily lead to clipping on loud passages.
Secondly, it is not clear that levels were 'matched' rigorously, if at all (see below for an interesting and telling sequel). Aczel did not own an ABX comparator until some years later (1988 or so). Nevertheless, the Carver listening evaluation was done 'blind' in that during test panelists did not know which amp was connected.
Panelists were first given time to listen 'sighted' to each amp in order to determine what they thought were sonic attributes of each. In this, panelists were convinced they could identify unique 'sonic signatures' exclusive to each amp. Then they left the room and formal testing began. Bob hooked up one of the amps, and left the room (that is, he was not present during the actual listening tests). Listener had no idea which amp was connected, but was allowed to listen as desired. Once a panelist made his determination as to what amp was playing, he left the room. Bob then switched (or didn't switch out) the amp--based upon the predetermined toss of a coin. 15 times panelists would come in and listen, then decide.
Tallied scores indicated randomness. No one correctly identified anything. The conclusion was that previously certain 'differences' between the two amplifiers had been 'nulled out' (again, see below). Bob was said to have made both amps 'sound the same' as a result of his tweaking. But did he really do that?
What is telling is that after Peter got his ABX device and proved to himself that level matching was the key, he refined his thinking. In fact, he told me that the 'secret' to the Carver Challenge was probably that the Carver and Levinson sounded same all along, even before the tweaking. But because he never matched levels, he had been fooling himself, imagining that differences were there in the first place, when they likely were not. Because of the intensity of the two day 'workout', and the suspected 'belief' that the amps were now sounding the same when at first they 'sounded different', and because participants couldn't reliably tell anything when the brands were hidden, they all believed that Carver had actually worked some magic, when in fact he was just doodling.
Peter told me that he came to realize that all of his pre-ABX 'tweako' amp reviews were worthless (he didn't use that word, but it was clear to me what he meant). The only thing I could tell him was that if he was fooled, the rest of us who 'followed him' were bigger and derivative fools, so he shouldn't take it too hard. Live and learn. That's what it's all about, anyway.
I think what at first bothered him was that after he changed his thinking, the tweako crowd piled on him pretty hard, calling him a sellout, a Carver shill, and a lot worse. It was all pretty ugly. Later, however, after the
Stereophile Conrad Johnson/Carver affair, I think he felt vindicated to a degree, but was mostly simply over it, and could not have cared less.