Confessions of a Cautious Objectivist: Zanzibar & Linearity

by Allen Edelstein | May 28, 2010

irst of all, it's great to be back. Back from where, you ask? For ten years in the late '70s and early '80s, I was fortunate to write for the granddaddy of perfectionist-audio publications, Stereophile. Now, after a couple of decades of just writing letters to audio mags, it's great to have the opportunity to write for a real audio publication again.

I am an objectivist at heart, someone who believes that if there is a signal, acoustic or electrical, you should ultimately be able to measure it, comparing what's going into a system (or a piece of gear in the chain) to what comes out. You should have a good replica of the input at the output. However, we are unable to measure what "good" means in a comprehensive sense, so I ultimately believe in combining objective and subjective approaches, which is the best we can do here and now.

To explain myself further, I'd like to recount two audio experiences from the '70s that made big impressions on me. The first was from the early part of the decade and relates to the perils of pure subjectivism. The second occurred at the 1979 CES and was my first experience with relevant objectivism.

purchased my first high-end speakers, IMF Studio Mk 2s, in 1971 from Music & Sound, one of the early high-end emporiums in Jenkintown, Pennsylvania, a suburb of Philadelphia. I got them home and tore off the grilles to look at the drivers. Staring me in the face was a paper-cone midrange. I expected a plastic cone and instantly called Mel Schilling, the owner of the shop. He told me to call Bud Fried (Irving M. Fried = IMF), fortunately from the area. Bud told me to bring the speakers to his laboratory on the following Saturday morning.

It turned out that the cone was not plastic, but a laminate of paper and PVA, the same stuff that gave the bextrene cones of the day a shiny appearance. But for some reason, one month the worker responsible for coating the cones decided it was ugly and didn't coat them. Bud got out a jar of PVA and a paintbrush and repaired the cones. This is how I met Bud Fried, a truly fascinating audio character whose career spanned the period from right after WWII to the early 21st century.

I stayed in touch with Bud, and we were good friends until he died. He taught me a lot, and I got all the updates for my IMF speakers for free. One morning while at work, I had some free time and phoned Bud. The subject became "guess who's coming to dinner?" It turned out that Bud was hosting a dinner for Percy Wilson, an early technical editor for Gramophone and a doyen of British audio, and his wife. Also attending were David Hafler and Gordon Holt and their wives. Would I like to join the group? I instantly gulped and said, "Yes!"

I recall cursing my way down City Line Avenue at rush hour, convinced I would be late and miss everything. But I made it and had a wonderful time. Percy was special. I still picture him playing a British prime minister in a '40s British film. David Hafler was very quiet and all the more impressive for it, knowing all he had accomplished in audio. Gordon was quite the character. Most important, he gave me his phone number, which was the beginning of another special friendship.

Percy and his wife were staying at a downtown hotel. Bud asked me to drive them there, and he made Percy promise to tell me his tale, the Zanzibar Fallacy, which Bud loved.

It goes like this: There was a captain of a freighter who had been sailing the seas for a very long time. He was a fine captain, liked by his crews. He had a tradition. Outside the ship's wheelhouse was a small cannon. Each day at noon the captain shot off the cannon.

One day, the captain got his crew together to announce his retirement. He was going to build a house outside the capital of Zanzibar, and he invited all his crew to visit him after it was finished. One morning a crew member showed up. He and the captain reminisced as the captain showed him around his home. On the second floor was a replica of the wheelhouse with a cannon outside, just like on the ship. It was nearly noon. The captain checked his watch and at noon shot the cannon.

The crewman had to leave, but as he left he asked the captain how he knew it was noon, so the cannon was shot at the correct time. The captain replied that it was simple. In the city was a man who studied time and had a grand collection of clocks. He just made sure his watch was coordinated with those clocks, and he advised the crewman to visit the man in order to see his collection.

The next morning the crewman was in town and visited the man. It was just before noon, and at noon all the alarms of the clock collection went off. The crewman complimented the man on his impressive collection. But he was curious. How did he know the time was correct, that it was really noon when the alarms went off? The man replied that it was easy. Outside the city lived a retired sea captain, and every day at noon he shot off a cannon. The man simply coordinated all his clocks with the captain's cannon shot.

We have to fast-forward about half a decade to the 1979 CES for the second story. It was evening in the Precedent Audio exhibit, and I was with Murray Zeligman, the head of the company, and David Berning, who is still designing significant audio electronics today. They were using a pair of Precedent Audio Mod 3 dual-transmission-line (both woofer and midrange) modular three-way speakers, the first Berning preamp, and an Audionics BA 150 amp designed by Berning, a mostly tube, triode-output amp rated at 150Wpc (actually, it was closer to 225Wpc; David conservatively rated it with 100 volts AC). The BA 150 featured variable negative feedback ranging from 16bB to 0bB in about four steps.

Murray and David wanted to show me the effect of varying the feedback, going from highest to lowest. At each step, as the negative feedback was reduced, the sound became clearer, more focused and more dynamic -- more like live sound. I knew that the reduction in feedback meant that harmonic distortion and intermodulation distortion were increasing. And things were even worse than I knew at that time. I didn't know then that the output impedance of the amp was also increasing, meaning the lumpiness of the frequency response of the amp/speaker combination was also increasing.

I was bothered. I certainly never expected that adjustments producing worse measured performance would result in improved sonics. Were objective measurements totally misleading?

Irritated, I asked David and Murray what was going on. I asked if they ever found an objective measurement that correlated with what they heard, expecting to be dissatisfied with whatever they told me. But they surprised me. Both replied yes and gave the same answer: linearity. I didn't know what linearity was then, but it was simple. If one volt in gives five volts out, then the gear that comes closest to producing ten volts out with two volts in is the most linear and should sound best. As the amp's negative feedback was reduced, the linearity improved, even though the conventional distortions were worse. And let me add that for an amplifier, this is particularly true with a reactive load like a loudspeaker, even more than with a test resistor.

o these are my two tales from the 1970s, when high-end audio seemed full of potential, some of which has been realized since. The first is a warning to be wary of circular reasoning, particularly in subjective judgments. One also need be wary when doing objective tests. There are many examples of scientists searching for a particular result and unwittingly designing an experiment to produce it. The second tale was my first experience with worse measured performance correlating with better linearity and better sound. I find this to be very relevant for evaluating speakers. I believe the single most vital factor for a speaker that produces the illusion of live music to be dynamics -- not just loudness, which is not really dynamics, but accurate changes in level whether small or large. This is simply linearity in the speaker's output.

Of course, I'll have firsthand opportunities to test my theories with music -- this is about that, after all -- any cannon shots being purely coincidental.