The Spendor 1/2 Loudspeaker
TWENTY-FIVE YEARS IN THE MAKING
Robert E. Greene
TAS issue 90 Fall 1993
Twenty-five years ago, the Spendor BC-1 redefined the world's ideas of how neutral and transparent a box speaker could be. The BC-1 marked the start of a new era. Before its introduction, dipole speakers were the only possibility for the finest possible sound; afterwards, box speakers were in full contention. Almost two decades later, (the late) Spencer and Dorothy Hughes, the founders of Spendor, brought out a new embodiment of the same design principles. Known as the SP-1, it incorporated the improvements in driver materials and technology of the intervening years. The SP-1, too, was an extraordinary speaker in its time.
Now Derek Hughes, Spencer and Dorothy's son, has designed a new speaker, the SP1/2, again with the same fundamental design principles but with new, further improved drivers. The SP1 /2 is far more, however, than a pleasing illustration of British family tradition. The SP1/2 is the fulfillment of a grand vision of speaker design; behind its unpretentious appearance lies an almost Promethean ambition, to steal not the secret of fire, but the secret of sonic truth. The perfect speaker has yet to appear, but the SP1/2 comes far closer to certain aspects of perfection than almost any other.
The 25-year journey to the SP1 /2 began with a concrete idea, derived from BBC research, of how to solve the many problems of box speaker designs. The idea consists of a relatively small, carefully damped box (2'x1'x1'), a single driver covering almost the entire frequency range of the fundamentals of musical instruments, a carefully designed crossover, first to the larger of two tweeters, then higher up to the smaller, and free-space mounting on a dedicated stand about one foot high. The height is chosen with regard for the first floor reflection. The goal was to achieve a seamless transition between drivers, and a virtually flawless tonal smoothness and accuracy, together with almost entirely nonresonant behavior. Following this plan, the SP1/2 retains the two cubic foot ported box; it uses an 8" bass/mid driver designed and manufactured by Spendor and two Scanspeak soft-dome tweeters, a 1 ½ " and a ¾ ".
To praise one speaker design is necessarily to denigrate others, at least indirectly, and this necessity is sometimes unfortunate. But I have no qualms, in thinking about the BC-SP designs and their coherence, in pointing out that the vast majority of multi-driver box speakers are in fact not coherent in actual listening. Coherence is destroyed, for one thing, by intrusive vertical lobing effects and interference patterns between the drivers. Some interference is bound to occur, but some patterns are far more audibly troublesome than others. (First-order crossovers in unpaired driver arrays are hopeless.)
A potentially great advantage to the use of a single bass/mid driver over a large frequency range is that real coherence is possible. It is, however, not guaranteed: Many designs of this type end up with an audible "hole" at the top of the bass/mid drivers' range that isolates the tweeter sound. With the SP1/2, even more than its predecessors, near total coherence is obtained. One can barely "hear out" the different drivers, even with specialized test signals. For all practical purposes, integration on musical material is complete.
More surprising even than the coherence is that the SP1/2 approaches the seemingly impossible: The SP1/2s are almost exactly neutral in tonal balance in real listening rooms.
The quality of performance in actual listening rooms is what counts, of course. Nowadays, anyone with some pretty good drivers and a software package can produce some sort of anechoic, theoretical flat-on-axis setup. But how the thing will actually sound, after the effects of baffle loading, diffraction, and, most of all, the listening room are incorporated--that's another story.
Many speakers look good in the anechoic test chamber or on the computer analyzer. Few of them sound good at home. And if sounding good means producing an audible facsimile of the input for a real listener in a real listening room, then most speakers do a rather poor job indeed. There are some that do all right; the SP1/2 is among the best. The level of sophistication and refinement of design, especially crossover design, that is shown here is extraordinary.
Overall, Derek Hughes and his associates have addressed and, it seems, effectively solved some, perhaps even most, of the problems of a box in a room, at least to the extent that these problems can be solved. Would it be too extreme to say that there are major problems that often aren't really noticed by most designers? Or even that audiophiles can come to interpret the non-solution of the problems as sonic correctness? Typical two-way speakers in typical placement have a big hole in the 100-300 Hz region from the floor cancellation.1 They have another hole in the 2-4 kHz octave from failure of the crossover to be smooth in audible terms, regardless of what the software package says. Audiophiles who listen to the SP1/2s will probably be inclined to think that they are too warm and are over-projected in the upper midrange. They aren't. They just don't do the wrong things many people are used to.
The question arises: How does one know? After all, reviewers are always saying that this, that, or the other speaker is almost neutral, even though few of these supposedly neutral speakers sound anything like one another. The answer comes down to listening-listening to familiar material and comparing the sound heard to what should be heard, listening to well recorded vocal material to verify midrange naturalness, and so on. The SP1/2s offered exceptional truth to source. In addition to their apparent excellence with commercial recordings, their performance was outstanding with recordings of my own playing and of various recordings of my musical colleagues, with whose sound I am intimately familiar. To my ears, the SP1/2s came surprisingly close to making instruments and voices simply and literally correct in tonal terms.
Of course, one might hope to verify such an impression through measurement, but no one has ever been quite sure what to measure in order to evaluate performance in real listening rooms. (Most of what has been passed off as "scientific" analysis of this question has been either too crude or too biased to be of any real use or validity.) But there are relevant things to easure, things that are closely connected to what one hears. The graph below shows the frequency response of the first ten milliseconds of sound arriving at the listening position from 300 Hz to 10kHz. This frequency range determines almost completely the intrinsic midrange tonal character of a speaker. Below 300 Hz, room effects begin to operate and they assume importance to the intrinisc behavior of the speakers. Above 10kHz, one is in a range where only "air" and "texture are involved, not really tonality and timbre as such. The first 10 ms /300 Hz-10 kHz , for the SP1/2 is amazingly good , with only a slight projection at 1-2 kHz marring the overall flatness at all, seriously.
The in-room RTA over the same frequency range is equally impressive, shown in the next graph.
Beyond measurements of this sort, a new and interesting feature of the situation has arisen in the form of DSP digital correction systems.
Anyone who has not been asleep will have realized that digital signal processing is bringing a revolution to speaker design and to audio in general. The possibility of DSP correction is revealing just how far from perfect most speakers really are. Not all analogue speakers will survive the winnowing that the DSP revolution will bring; few will still seem acceptably low in coloration and accurate in stereo imaging in their unprocessed form once their DSP-corrected version has been heard. But, as so many of the expensive giants are left behind in humiliation, I believe that the small and fairly inexpensive SP1/2 will be like Cinderella emerging from the kitchen to find that the glass slipper fits.
I don't want to set DSP in general up as a shibboleth. Listening judgment remains supreme. And a DSP system always involves choices that must prove themselves in listening terms. But it is striking what the SP1/2, so neutral in listening, underwent in the DSP correction process by the SigTech AEC 1000 (see my review in Issue 89 and reprinted on this site ). Other speakers changed dramatically. But with the SP1 /2, there was almost nothing to correct.
Let me clarify. Some micro-colorations in the midrange vanished with DSP, the top end came up a bit, and the deep bass was made somewhat smoother, extended, and tightened. But in the midrange, where most of music is, there was remarkably little tonal change with DSP in versus DSP out.2 Cinderella, indeed!
An idea of why the change with the SigTech was perceived as small on musical material can be obtained by looking at the SigTech "before" graph that follows. This graph shows the response that SigTech corrects, a response made by a complex time-windowing process. ( This in-room measurement is actually from a different room: the first 10ms and RTA graphs are from my new listening room . But the main features are similar.) The SigTech graph looks somewhat irregular, though note the expanded vertical scale. But if you recall that narrow-band dips are not very serious in audible consequences, the smoothness of the over-all behavior is impressive and indicative of little to correct in audible terms tonally.
The SP1/2 is a conventional, or at least conventional looking box speaker, and at this point in standard (non-TAS) audiophile reviewing one might expect some pieties about it. After all, as a box speaker, the SP1/2 cannot be expected to perform as well as a panel speaker, and so on. Some such pieties reflect partial truths. But I want to look at the situation a little more carefully than that, in general terms as well as in the specific instance of the SP1/2.
Conventional audiophile wisdom has dictated over the years that dynamic drivers are "slower" and "store energy" because they are much heavier than the driven membranes of planar speakers. The energy storage was, as often as not, real enough, but the conventional explanation was largely wrong: The energy storage was not in the mass of the driver as such, but rather in its break-up modes. The drivers have improved in their behavior, and energy storage has become much less of an issue. Indeed, in the nearly ubiquitous "waterfall" plots of decay behavior, good dynamic speakers are competitive indeed with many, if not most planars in midrange and higher frequencies. The audible significance of waterfall plots is a matter of controversy, and many of the people publishing them and interpretations of them are whistling--or perhaps I should say slamming doors--in the dark. But waterfalls do reveal information about energy storage; that's what they are for.
Regardless of such considerations, the SP1/2 is a clean, clear speaker from the midrange on up. Compared to some other audiophile speakers, the SP1/2 has a somewhat subdued top end in the far field. Though its on-axis anechoic response is flat and extended, except for a bit of droop at the transition to the top tweeter, its audible character is possessed of a non-aggressive top. Of course, audiophiles have a long history of tormenting themselves with excessive top end response in the name of accuracy, even though the sound thus obtained is demonstrably inaccurate, both in terms of musical values and technical truth. One need only think of the moving coil cartridges of yore and metal dome tweeters to get the point.
Exactly how much top end energy there should be in the reverberant field is a matter of controversy, as I have often discussed. You can hear extra top end power response, as anyone who has experimented with the switchable rear-firing tweeters on Snell speakers can attest. On the other hand, experience with those rear-firing Snell tweeters that bring up the top end reverberant field has shown that the sound is always more natural without them. There is simply no circumstance in which live music contains the sound of "flat power response" in the high frequencies. With the SP1/2, you have no rear firers to fill in for comparison, but you won't miss that extra zip if you are really looking for the sound of live music. I have gone over the logic here so often that I am not going to do it in detail again (see, e.g., my Spendor SP-1 review in Issue 37, and my article "Records and Reality," Issue 38 and on this site). But be assured that the high frequency balance of the SP1/2 is both musically natural and technically accurate (within a decibel or so, anyway) according to one entirely logical technical standard. (No one should sacrifice musical truth to artificial technical standards imposed arbitrarily by the engineering practice, which in the case of flat power response is not very carefully analyzed. The argument for flat power response is both superficial and wrong.)
The top end of the SP-1 is smooth, clean, and natural regardless of questions of prominence. The soft domes used (from Scanspeak) are to my ears far superior to the hard, metallic, colored, dirty, and unpleasant metal domes that stalk the speaker world like bloodthirsty beasts. (At this point, I guess you're sure you're not reading Stereo Review.)
It is mostly in the bass that the SP1/2s show their limitations. A two-cubic-foot box and an eight-inch driver cannot be expected to pump out enormous quantities of deep bass. And the SP1/2s won't. The only time I overdrove them with acoustic music at rational volumes was with the grossly exaggerated bass drum on a Telarc. (Hasn't anyone told the people at Telarc that bass drums are not like that in orchestral music, heard from the audience?)
Bass dynamic capacity and extension to the lowest lows are limited. That's why Spendor put out the S-100, I suppose (see Issues 64, 68), although my impression is that the SP1/2 is (even) less colored than the S-100, so the choice does involve some compromise. In my room, the bass of the SP1/2s covers the normal orchestral range, down to 40 Hz, smoothly and without much roll off. But they don't plumb the deepest depths, and orchestral music played through them lacks some "scale" in consequence. This deficiency is not great, but it exists nonetheless.
A second point in this regard is that, as I have often stated, box bass interacts with listening rooms differently from dipole bass. The particular way a bix radiates sund into a room acounts, in my view, for most of what one calls "boxiness" in box speaker's bass. Indeed, when taken outdoors, the SP1/2's bass is almost entirely non-boxy. The speaker's bass performance is clearly well engineered. In the listening room, careful placement can make the bass quite smooth and even. But inevitably a box speaker will betray a bit of the fact that it is effectively an omni radiator in the bass but a half-spaced (or less) radiator above. The SP1/2s, though non-boxy as boxes go, do suffer from this characteristic problem, as they must.
The SP1/2s achieve their remarkable room-independent neutrality via our old friend, control of radiation pattern. I explained the general role of controlled radiation in perhaps excessive detail in my review of the Snell Type B in Issue 75/76 of TAS , so I won't recapitulate the theory here. Whereas the Snell's use the D'Appolito symmetric driver arrangement to control the upper midrange dispersion, the SP1/2 accomplishes the same goal by a different mechanism: running a relatively large (eight inch) driver up to a rather high crossover frequency (3 kHz). Thus the driver is fairly beamy near the top of its range.
This procedure is risky, and has the potential to result in the development of an audible "energy hole." But when it is done right, as it is in the SP1/2, it works wonders in minimizing wall reflections in a critical frequency range.
Because of their deliberate beaminess, the SP1/2s have to be pointed at the listener if their full potential neutrality is to be appreciated. But the same controlled radiation that aides in the neutrality of the SP1 /2s also contributes to their remarkably convincing stereo performance. Altogether, the SP1/2 is de facto what I was calling a "second-revolution" speaker in my Snell review--even if the basic design vision of the SP1/2 goes back many years.
To appreciate how nearly perfect the stereo performance of the SP1/2s can be, and to get a goal to shoot for in your own listening room, you should listen to the SP1/2s outdoors. Their ability to "float" central images is uncanny, perhaps in part to their superb pair matching. And, when the material has widely spread images (through phasiness), they are also surprisingly capable of projecting images beyond the speaker boundaries. Their performance is surprising in that it seems to be a common perception among audiophiles that speakers that need to be "toed-in" tend not to do wide soundstaging well. Moreover, depth of image is superb. So is the insight provided into the recording venue.
Indoors, wall reflections will tend to interfere with stereo performance, though the degradation will tend to be less severe with controllable radiation pattern speakers. But careful placement of the SP1/2s and well-placed damping material will enable you to recover an impressive approximation of nearly flawless outdoor, reflection-free stereo. To my mind, the result is a striking affirmation of what is possible with direct radiation, two-channel stereo. Incidentally, I should add that the SP1/2s should be placed widely apart, in something along the lines of the theoretical Blumlein position (900 apart, from the listener's viewpoint). With their superb center-fill characteristics, such an arrangement will generate no hole-in-the-middle effects--except those on the recording.
This brings me to one last problem, if it can be called a problem as such. It arises with the all too numerous recordings with exaggerated left-to-right separation and with, say, a whole first violin section jammed into the left channel only. The left channel with the SP1/2 will naturally localize itself as a point, the SP1/2 being effectively a point source. But this situation, in which you are effectively listening to one channel at a time, will also make one hear the sound as if coming from a box. Under these circumstances, every speaker reveals its size and shape.3 Panels certainly sound like panels, e.g., the Sound Lab Pristines I commented on in Issue 87.
Let me turn away from all this technical stuff and conclude by just talking about music. I had a wonderful time with the SP1/2s. I was still listening to entire operas on them long after I had listened enough to write this review, just to glory in the voices so magnificently reproduced. Similarly, string chamber music was hypnotic in its beauty of tone. Popular vocal recordings, for all their over-close miking, nonetheless acquired a compelling, if close-up, naturalness. Organ music had the richness of timbre and texture it has live, even if the 32-foot stops were a bit curtailed. And so on.
I don't want to suggest that the SP1 /2s are perfect. As noted, orchestral music lost some sense of scale and had some hint of dynamic limitation, although what I estimate to be realistic audience location levels were certainly possible. There was some bottom bass missing, and there were some box effects. But the overall truth to live musical experience was startling. If you listen not for whether various audiophile vocabulary items are being provided, but rather for basic resemblance to music, then the glory of the SP1/2s will shine forth. They simply come remarkably close to what music sounds like.
We all tempted to think in words and to listen via verbal categories, to be analytical, rather than intuitive listeners. But try instead to summon up the literal sound of music in your mind. Then listen to the SP1/2s. I believe that you will understand why I consider them a masterwork of speaker design.
1 Roy Allison's research has pointed out this situation, and suggested solutions via special placement of speakers. But these solutions seem to be implemented rather seldomly, and the whole problem tends to crop up again and again.
2 With the SigTech, one can switch the correction in and out instantly with matched volume levels. With the SP1/2s, discontinuities were minimal from such switching, even in the middle of notes! (See my SigTech, review in Issue 89 and on this site.
3 Almost everyone: Speakers like the Quad ESL-63 or the Stax F-81 that approximate ideal behavior by electronic delay lines are almost free of the problem.
Click HERE to go to the article as published on The REG ON AUDIO website.