Features

Alvin Gold responds
It's not often I take up arms as a result of what I read in the hi-fi press, but there are moments. When I am told that all competently designed amplifiers used within their input/output limitations sound the same, I see red. And then there's this business about single speaker demonstrations (SSDs). . .

The reason I see red here is simple—I actually spend a lot of time listening to music using hi-fi equipment, and my ears tell me that SSDs work. BS seems content to use paper logic to prove that they don't, apparently without reference to any practical experimentation. Thought experiments conducted in ivory towers have their place, but as we are dealing with such complex, interwoven variables, it would need a computer model and a lot of data to quantify the problem adequately. Drawing conclusions comes even later in the process.

I'm not above thought experiments myself, it's just that I know their limitations. Although all objects, including walls and ceilings within a listening room, inevitably have an effect on the reproduced sound, there are different classes of interaction in volved. Some are common to all systems, those caused by the floor for example, and others tend to be there in one form or another as a result of furnishings and so on.

To a first approximation, these environmental effects are relatively linear: if they involve storage and re-radiation of the sound, they do so at very low frequencies (2–3Hz for a structure like a wall for example), or the extra path lengths involved introduce delays that tell the ear/brain that what they are hearing comes from the room, not the original sound. Alternatively, we're talking about the effects of a table or other item of furniture where resonant effects are probably very well damped.

In acoustic terms, boxes and loudspeakers cannot be directly compared. A box may have little effect for the reasons given, or it may indeed have an effect which says to the listener that "there is a box here that is modifying the sound." Big deal. With transducers, however, what we're dealing with are not linear or simple effects; but highly reactive, frequency-selective ones; that can result in a surprisingly large output from an undriven loudspeaker when measured with a voltmeter (yes, I've done it).

Presumably the loudspeaker, thus wound up, releases this energy, mostly of it near its fundamental resonance and smeared over a period of time. It's the selectivity of this mechanism to frequency and temporal aberrations that marks it out as important. Such distortions in the soundfield are undoubtedly quite difficult for the ear to sort out.

I have noticed that undriven panel loudspeakers seem to have much less of a negative effect on the sound of a system than most box speakers. This runs contrary to BS's assertion that the larger the obstacle, the greater the effect; and it reinforces the idea that it's a resonance problem, probably associated with the fundamental driver resonance when mounted in a box, that's at the root of the observations that led to SSDs. Diffraction is hardly a candidate for "severe boxiness" as Bill Sommerwerck suggests; much more likely is that the boxiness comes from the usual source of boxiness in loudspeakers—the combination of a drive-unit mounted in a box!

Some of the other points in BS's article are either spurious or just plain wrong. The one about stereo listening is a good example of the latter. We're stuck with two loudspeakers for stereo whatever happens. More to the point, loudspeakers that are in a circuit do not have the same effect on the sound as undriven one: a driven loudspeaker is effectively shorted by the amplifier, and is a much stiffer and well-damped structure. It will respond in quite different ways in a radiated soundfield than an undriven speaker. I know of one dealer who inserts a wire shorting link into all loudspeakers not in use for precisely this reason.

To say the IT ignored the stereo problem is quite fallacious. I personally heard him make a semi-ironic remark at the Chicago CES about 3–4 years ago to the effect that multi-mike recordings don't sound as good as purist ones because of the number of transducers in the recording room! He even made a similar comment about mono vs stereo. And if you want to talk about the very special qualities of some old mono recordings . . .

Far be it from me to defend IT. I haven't done so for a long while, and the last time I was up in Glasgow we spent 24 hours at each others throats about Linn's awful Index loudspeakers—yes, I've listened to them rather more thoroughly since I wrote about them in Vol.8 No.3—and other topics. But his ideas about SSDs (as I understand them) bear repeating.

The point of SSDs is simple: they are designed as a dealer aid. Now it may be argued (as James Michael Hughes has done in the UK) that they are counterproductive for 101 different reasons, or that the effects are small in absolute terms. On the latter I'll concede the point, though small differences are what hi-fi is all about. Moreover, the kind of distortion eliminated by SSDs (time-smearing distortion) is an important one.

As far as I know, no rigorous analysis has been done, but it's easy to show empirically that additional transducers in a room adversely affects the sound of a good system—be they loudspeakers, telephones, or cassette recorders. The effect tends to be particularly insidious, because it is manifested as a loss of pitch and timing integrity quite distinct from the classes of effects that arise from furnishing and room boundaries. Tonal and other more commonly anticipated effects are not usually severe, so the degradation is sometimes missed first time round by the less experienced listener.

We're talking now about quite subtle differences, but high fidelity is all about quite subtle differences. No one, not even IT, as far as I know, has suggested that all transducers other than the driven ones should be excluded from listening rooms. The very sound of an orchestra is partly determined by similar interactions between highly resonant instruments.

The point is, however, that many listeners already have listening rooms that approximate SSD environments. It's only comparator-equipped dealers that don't. So why not try and achieve the same as the home situation when a dealer demonstrates his products to the public? This way, the sound is as representative as it can reasonably be made at the point of sale, and the equipment is given a chance to do its damndest.

My personal position on this is that when reviewing it's essential to hear what the product being tested is doing, and that reviewing under SSD conditions is clearly the best and safest way to achieve this end. At best, extra transducers will have an unpredictable effect, so why not eliminate what is after all an unnecessary variable? For this reason, I have auditioned loudspeakers under SSD conditions for a number of years—but I wasn't foolish enough to subject myself to the not inconsiderable inconvenience without first establishing that the differences were not merely audible, but also musically important. You can pontificate until the cows come home, but with a good, well-optimized hi-fi system, the facts quite literally speak for themselves.

By the way, Bill Sommerwerck, grandma is quite adept at sucking eggs. Naturally, serious listening is not done with a coffee table interposed between myself and the loudspeakers. I have a wife, you see, and although she tolerated a Krell and a pair of Maggies, eventually the coffee table has had to live in the room between times. You might like to know that the room has been switched around. The loudspeakers sit just forward of the bay, and the (now round) coffee table is to one side of the hot seat.—Alvin Gold

J. Gordon Holt adds some thoughts
I agree with AG that BS shouldn't pre judge SSD until he tries it, but I also agree with BS that other speaker boxes in the room probably will have an effect, and probably a detrimental one, on the sound. But so, of course, will any other object in the room, including people. (Why not a people-less demo, for the ultimate purity?)

It appears to me that both views are partly right and partly wrong. BS, for example, overlooks the fact that an enclosed volume with a hole in it is a Helmholtz radiator, which will readily absorb energy from the surrounding air at its resonant frequency. AG reports only that the sound is better when other speakers are out of the room, without ascertaining whether this is due to the effects of their undriven cone surfaces or of reflection and refraction from the outer surfaces of their enclosures. Both, it seems to me, overlook the fact that the smaller the offending object, the less effect it is going to have on anything, and that there will be a point somewhere down the line where its effect will be too small to be detectable by even the keenest ears.

IT's stand on this appears to be based on a common misconception: namely, that vibration of an undriven speaker cone creates spurious sound waves in the room, just as though that speaker were being driven at low level by another signal source. It doesn't, because it isn't.

The amount of vibration induced in an undriven speaker cone from impinging sound waves is in fact significant, and can be easily felt with the fingertips. The amplitude of those vibrations is similar to that that would be caused by an electrical signal in the tens of millivolts range feeding that speaker, and would be clearly audible if no other speakers were playing. But the amplitude of these induced vibrations is much lower than those of the main, driven loudspeakers, because the latter are creating all the air-pressure variations in the room, while the former is being acted on by only a very small percentage of those variations.

Contrary to one's gut feeling about this, these vibrations of the undriven cone are not being re-radiated, because they represent a loss of energy from the soundfield, not a contribution to it.

That undriven cone has mass, and it takes energy to set any mass in motion, so practically none of the sound-wave pressure reaching that cone will ever leave it. Virtually all of the sound-pressure energy will have been used to move the mass of the cone. The only re-radiation that would occur from that "undriven" cone would be due to the release of stored-energy resonances from it after the cessation of air-pressure-induced motion. And in any loudspeaker we would give house room to, that stored energy should be more than an additional 50dB lower in amplitude below that of the ! cone vibrations which stored it. It is exceedingly unlikely that they would be audi ble, even to AG.

It is much more likely that the audible effects of other speakers in the room are due to reflections and refractions from the enclosures, plus their behavior as Helmholtz radiators. It is also obvious that, the smaller the volume and the smaller the hole, the less effect either is going to have; below a certain magnitude, the effects will be imperceptible.

For this reason, the idea that a telephone or a cassette recorder could affect anything at all strikes me as bizarre. Certainly, their presence is going to cause a change in the soundfield in the room, but I cannot believe that anyone could hear that change unless the object in question is suspended from wires, a foot in front of his head. But we are speaking here of common sense.

Or are we?

Incidentally, a frame-core room wall has effects far above the 2–3Hz cited by AG. Just pound on one and listen to the sustained thud it produces. That "thud" will usually be between 25 and 40Hz, and the resonance which causes it will absorb, not add to, a speaker system's LF output.

There is, it seems to me, enough justification on scientific grounds to assume that other loudspeakers in a room will affect the sound of the main ones, and to conclude that SSDs are probably a good idea. I would hesitate to assert, though, that MSDs (multi-speaker demos) have enough detrimental effect to swamp the enormous differences between competing loudspeakers, and that in-store auditions are still of value even with other loudspeakers present. But any system chosen from a MSD is likely to sound better under the more ideal SSD conditions at home.—J. Gordon Holt