Living Voice Auditorium R25A Loudspeaker

Living Voice has long been a significant presence in the British speaker industry. The “25” in the model number here is in reference to the speaker being a 25^th anniversary model. In fact, the history of the design goes back even a little further, in that the original Living Voice Auditorium was apparently strongly influenced by an earlier design from DALI, the Model 104. Meanwhile, the original Living Voice model has been steadily and carefully revised in a process that resembles the Old Masters of violin-making refining their designs year after year, even generation after generation. Living Voice is not unique in this respect. It seems a British tradition, with the original LS3/5a and the original Spendor BC1 having led to decades of refinements and variants, over a 50-year period for the BC1 and close to that for the LS3/5a. Still, 25 years is in itself impressive in a world where technical products tend to have a short life.

The design goals of Living Voice in the R25A and its related models (there are other similar designs at higher prices in the Living Voice line) have pursued an additional goal compared to BBC designs, which were seeking complete neutrality. Living Voice sought not just accuracy but high sensitivity. The Auditorium R25A has a listed sensitivity of 94dB. Compare this to say the LS3/5a, with its 83dB sensitivity. Clearly, we are in a different world here, intended to be suitable for lower-powered tube (or Class A solid-state) amplifiers and, indeed, Living Voice’s designer Kevin Scott voiced the speakers with tube amplification. Living Voice also makes horn speakers with the low-power tube user in mind: The Vox Olympian has a sensitivity rating of 105dB!

But none of this is to suggest that the Auditorium R25A is not of interest to people who are happy with higher-power solid-state amplifiers. It is, in fact, a speaker of general interest, even though it may be particularly attractive to the low-power tube people.

Still, the sensitivity issue is an important aspect of the design, and it makes sense to look at that general context first before getting to the more specific sonic description.

Size, Sensitivity, and Bass

In the early days of the modern high-fidelity era, the late 1940s and early 1950s, amplifiers of high power by today’s standards were not available. An amplifier with 30 watts (per channel) was considered a powerhouse. But people wanted bass with both volume and extension. The answer was to make speakers big. The logic of this is based on physics: Bass wavelengths are long. A 40Hz tone is over 28 feet in wavelength. This means that a small driver will not couple to the air at that frequency at all well and will have to move violently to produce any substantial sound at all. Large drivers fare better. So, people ended up making really large woofers: the ElectroVoice Patrican 800 had a 30-inch driver! That was then, but acoustic principles do not change, and if one wants bass with high sensitivity, size remains the way to go. The Cerwin-Vega CLS215s—my go-to speakers for reproducing the full bass power of the orchestra—are 96dB sensitive, and they get this combination of bass power and extension and high sensitivity by being almost the size of a small floorstanding refrigerator and using two 15-inch woofers. At the other extreme, the diminutive Spendor S3/5, exquisite though it is at moderate volumes in midrange reproduction, is a fish out of water faced with orchestral power at full throttle. Its bass limit is 70Hz, the sensitivity is 84dB, and the maximum power input is 70 watts: beautiful sound within its limits, but orchestral bass power and extension are not on the program and most definitely not so with a low-powered amplifier. A small speaker can pump out bass with enough power driving it and the structure to handle the power: Subwoofers can be small and still produce low bass in a big way. But they need a lot of amplifier power. Many have amplifiers as powerful as 1500-watt built into them. And the big power is needed, and the bass driver is driven very hard.

The Living Voice R25As are a well-chosen compromise in the inevitable trade-offs among size, sensitivity, and bass power and extension They are sensitive enough to be driven by quite small amounts of power, small enough to fit gracefully into ordinary domestic environments, and they have enough bass extension and volume capability to be quite satisfying for most types of music.

The designer of Living Voice, Kevin Scott, is firmly of the opinion that amplification of any real power (by contemporary standards) involves serious sonic errors. I must say I do not find this to be so. But, of course, there is no harm in principle to having a speaker that is so high in sensitivity that high power levels are not even close to being needed. The power that is available but not used is not bothering anything. And for some people, the relatively high 94dB sensitivity will be desirable. To check out the lower-power situation, I used the Schiit Aegir (solid-state, but relatively low power at 10 watts in pure Class A, 20 watts in what Schiit calls Continuity mode). It did a fine job with the R25As. I did not have a low-power tube amplifier to try, but in general it is definitely the case that 20 watts is plenty, and I would guess that 10 would be enough for most purposes. Certainly, the Aegir was working superbly well at the levels involved, with Class A sound in the best sense—beautiful amplification indeed, at a very rational price ($799 if you do not need more power than is on offer). The power was sufficient in the case of the R25As as long as no substantial equalization was used—more on that point later. I also used the 100-watt Benchmark AHB2, which, of course, was just loafing along in its neutral, distortion-free way at the power levels likely with a 94dB-sensitivity speaker with or without EQ.

What is in the R25A?

The R25A has two 170mm mid/bass drivers in an MTM (midrange/tweeter/midrange) driver configuration with a 26mm soft-dome tweeter. The drivers are made by Scanspeak. The mid/bass driver has a coated-paper membrane and is built specifically for Living Voice, the original form of the driver having been otherwise discontinued, it seems. The cabinet is a slim floorstander, unusual in that it is supplied with a podium to lift it to a better listening height. (Bravo for this—all too many smaller floorstanders require the listener to sit almost on the floor to be on a good axis.) This is a good idea and also a necessary one, because, in fact, the response of the R25A varies a good deal with the vertical position of the listening position relative to the speaker, unless the listener is really far away. The speaker has quite a wide dispersion pattern overall, but it is not particularly stable in frequency response with variations of the vertical listening position (listening height). On the plus side, the MTM arrangement makes the speaker interact with the floor, it seems, rather less than usual. The R25A is ported, with the port in the rear. Positioning the speakers fairly far into the room seemed to work best in my setup. (Close to the wall behind brought up the bottom but in a somewhat colored way.) The speakers sound like themselves in general terms in a variety of positions, but the best possible performance depends on careful adjustment, apparently especially so because of the speaker’s complex radiation pattern. Patience is rewarded. This is not a speaker design of the appliance approach, popular with what used to be called “mid-fi” companies, where you load up the dishes any old way and push a button and the dishes get washed. (One of the really peculiar ideas involved here is that one ought to compare speakers by switching between them but putting all the different pairs in the same place successively, never minding that the best place might be different for different speakers.)

Starting to Listen and the Imaging Behavior of the R25As

When I first set up the speakers, I ran Stereophile’s Test CD1 (as is my usual practice) to verify that the channels were wired correctly and that the two channels were “in phase.” This is easy to detect on music material, but it is quicker still to use signals designed for the purpose. The channels were correct (track 2) and so was the polarity (track 3). But the latter was startling to hear. The difference is clear on any pair of speakers that are closely matched in frequency response. But for the R25As, the difference was extreme. Focus was perfect in polarity, and in reverse polarity, the barking dog was just nowhere at all—total lack of spatial focus. This was more so than with almost any other speakers I have tried, almost at the level of the Gradient Helsinki 1.5s, which are the absolute champs at this in my experience. No doubt there are those who think that performance on such test signals is not relevant to music, but this particular kind of test, and the similarly impressive result of the pink noise correlated versus uncorrelated, is directly related to stereo imaging in theory, and so it proved in practice, with music material. The lateral stereo performance of the R25As was super. In addition, the depth of image was also quite startling good.

It is natural to ask whence this arises. What makes the R25As so good at stereo? The currently popular belief is that superior stereo arises from uniformity of radiation pattern. This is practically a mantra among semi-technical writers on audio, that wide and uniform radiation gives rise to stable and correct stereo. Stable, maybe, but my experience is that the best possible focus and depth is actually attached to certain kinds of deviations from theoretical point-source uniformity. It is hardly surprising that this might be so. The ear/brain senses location of images acoustically not from the signals at the two ears with a fixed head position, but from how the signals vary with small, virtually subliminal head movements around that position. It is thus more than likely that certain small variations of the soundfield with such variations of head position would give enhanced results, and others not. In any case, my experience has been that speakers that run a single driver (or in the R25As a pair of drivers) up to a relatively high crossover point before the tweeter comes in tend to have extraordinary image focus. And the R25As have the crossover to the tweeter at 2.85kHz. The off-centered tweeter placement, which tends to minimize overall diffraction effects, might also play some role. Finally, the MTM driver arrangement, which makes the floor interaction different from a two-way with two drivers plays a role I think in the generation of depth and in freeing the front of the soundstage from the plane of the speakers.

In any case, leaving such tentative theorizing aside, the imaging behavior of the R25As is top drawer. And this is a constant enhancement of the musical experience. Even if you think you do not care, that only the sound itself not the spatial aspect counts for you, I think you will find the way in which the R25As inhabit the room without seeming to interact with it spatially enhances the sense of listening to music as if in the original venue. I, at least, found this a continuing source of musical satisfaction and to an unusual extent.

The Tonal Character

The R25A’s basic tonal character is somewhat distinctive, and the perceived frequency response is a bit up and down, although overall it is smoothly changing on the whole. From the bottom up: Deep bass is missing, though extension is not bad considering the size and the high sensitivity. There is a kind of lumpiness around 100Hz, and below that the bass is a bit lightweight, though it goes nominally down to below 40Hz. Above that, there is a dip (not unusual, from floor effect, not so severe as in some speakers). The midrange in the 500Hz-to-1kHz region is somewhat over-prominent. Then, there is a perceived broad dip around the crossover point (in the 3kHz region), arising from the fact that anywhere but exactly on the right axis vertically there actually is a dip. The top seems smooth and extended, even slightly rising exactly on the tweeter axis but somewhat rolled not far off that axis. The most observable aspect of these features is the overall lightness and midrange emphasis. And this really is observed, not so much as a coloration in any narrow resonance sense but as a shift in balance. The male speaking voice is lightened up, the opposite of chestiness. And the female speaking voice is similarly affected. The cello sounds more in the viola direction than it should, more alto than tenor. On musical instruments this is not typically unpleasant, but it is not really right, and the speaking voice is often quite peculiar, especially compared to the BBC school speakers which are optimized for the speaking voice in particular. To paraphrase Lewis Carroll, “Collar that midrange” is the thing to do with some EQ. Otherwise, things are just too strange, to my ears anyway. It almost sounds as though the baffle-step compensation had been omitted or at least was incomplete.

What is peculiar to my mind here is that this midrange emphasis, which is quite conspicuous to my ears, is hardly commented upon in other reviews. The RM25A has been quite widely reviewed, but almost no one seems to have even noticed the distinctive tonal character of the speaker. This is part of a trend: In the apparent zeal to talk about other things than frequency response, contemporary audio reviews often simply ignore frequency response effects altogether. And then there is the fact that the combination of the “floor dip” and the roll-off off-axis in most speakers leads quite often to the midrange being somewhat projected, as you can see in in-room measurements when they are published. In any case, the effect in the R25As does not look spectacular in measurement terms, but it is there, and it has a musical effect, whether one likes this effect or not. To my ears, the R25As sounded a lot better if one pulled down the 800–1200Hz region somewhat, smoothly but definitely down. With that done, something like neutral balance appeared outside of the (not disagreeable) sense of dip around 3kHz arising from the near-but-not-on-axis dip if one moves vertically. The corrected response was actually extremely attractive to my mind and made a surprising number of recordings sound “good” in the sense of sounding rather like concert music. Euphonic in the best sense!

Back to the Sensitivity Question

This brings up the issue of exactly what sensitivity measurements mean. Traditionally, sensitivity is measured at 1kHz. The idea is that people tend to set the volume according to the level heard at 1kHz, and this is likely more or less true, although part of this is just the convenience of the round number 1000. But in acoustic music, orchestras in particular, the big power demands are actually further down in frequency, as you can verify for yourself by using a peak-power-level meter and playing some music. The big power demands happen when the orchestra goes “whomp” in the bass. The effect of this is that if a speaker is elevated at 1kHz and down a bit lower down, the sensitivity rating at 1kHz will give a misleading impression of how many watts of amplifier power you need overall. (This was discussed by RH in his editorial in Issue 335.) One watt may be loud enough in the mids, but when the bass whomps come along, you may be in trouble with the level that was all right for 1kHz.

This is a very real issue. An amplifier that puts out 1 watt will play 1kHz at 94dB before clipping. That is quite loud. But when the big bass comes along, clipping will happen at the drop of a hat. How extreme is this in practice? I tried a while back a McIntosh amplifier that had peak-level meters on my Harbeth M40s, sensitivity 84dB. Playing orchestral music at moderate levels—mid-80dB forte level overall on an SPL meter (C-weighted), the big bass whomps drew 170 watts frequently. People who think their micro-powered amplifiers are not going to clip are likely mistaken, unless they listen exclusively to clavichords and soprano recorders. Of course, one of the nominal selling points of SETs is that they tend to clip gracefully—and they surely need to.

Looked at the other way around, with 1kHz being the determiner of the setting for the volume, a speaker that projects the mids will need less bass power because the volume setting will be lower. But if you EQ the bass up to match the mids, then power demands go up. In effect, the speaker operates as having lower sensitivity.

The effect of all this is that, whether consciously or subconsciously, the high-sensitivity/low-power world tends to pursue midrange forwardness more often than not. As one might expect, SETs are usually rolled at the frequency extremes. People talk about their midrange superiority. But the midrange of SETs is not necessarily superior—it is just more abundant compared to the frequency extremes. These effects are large in listening terms. And similar bias is involved sometimes in speakers.

I am not trying to suggest that anyone sets out to make a speaker with pushed-up mids in order to make its listed sensitivity higher. Rather, there are factors that make for higher sensitivity that also promote midrange forwardness. Among these are lightweight drivers, which are more likely to be lively at the top of their operating range, and the absence or near absence of internal damping. This contributes to higher perceived sensitivity, since damping out sound in effect converts sonic energy to heat, so there is less sound energy emitted to the outside world. Furthermore, the undamped sound seems livelier even if it is not all that much louder: Resonances make things sound louder. This is, of course, no place for a long technical discussion, but the R25As do have rather little internal damping, according to Living Voice.

Let me give an analogy. If you sit down at a grand piano and depress the right-hand pedal (the damper pedal or sustaining pedal), which lifts the dampers off all the strings. The instrument seems almost alive—the very act of applying the pedal and lifting the dampers makes a small sound reminding one of the resonances waiting to be unleashed by pushing down the keeps and having the hammers hit the strings. Now these resonances are what make the music, creates the sound of the piano playing. But they are all waiting there to be released.

At first sight, one might think that for a speaker to be lifelike, it should be similar; it too should be a bunch of resonances waiting to be unleashed. But this is not really the right idea. If a grand piano with the dampers raised is a racehorse waiting for the starting gate to be opened, aquiver with excitement of the power to be unleashed, a speaker ought to be more like a brick lying on a mud flat. It is not about to move unless shoved and when the shoving stops, it stops moving. Speakers need to reproduce the recorded result of the resonances of musical instruments, but they are not supposed to have any resonances of their own.

Now, people worry a lot about how to handle the movement of the cabinet walls in the lower frequencies, either by making the cabinets super-rigid or by making them slightly flexible but damping them. What I am describing is further up in frequency than the bass. A midrange driver playing at say 2kHz is producing a backwave of a wavelength of about seven inches. It is not pressurizing the box as a unit the way bass does, but it is still bouncing around inside in various ways. What is one to do with that sound? Ideally, one wants simply to get rid of it because it is stored energy that one does not want to hear. So, usually, speakers are stuffed with something or have sound-absorbing damping material on the walls of the box. This material gets rid of sonic energy—it is supposed to, and it does.

You can skip doing this. If you skip it, the emitted sound will change. The resulting sound is livelier and less “dead.” But dead is really what you want to hear—or, at least, I think so. In general terms, controversies along these lines have been going on in audio forever. How dead should things be? How damped down should a listening room be? What about the backwave of dipole speakers, absorbed or reflected? It depends on whom you ask, especially about listening rooms.

To return to the R25As, they belong somewhat in the “undead” category. And to my ears, they are (in the words of Rex Stout’s mystery classic) “not quite dead enough.” Even if you EQ down the midrange prominence to get the measured response flat, there is a kind of glow to the sound at a subliminal level that I was not fond of. I kept wishing I could “put a sock in it,” literally—stuff something into the box, not the port but into the box itself.

Don’t misunderstand: This is not a huge thing. Musical instruments, after all, do “glow” in this sense—they are resonators. But to my ears they tend to acquire a little more glow than is really there. This is true especially of the piano. I freely admit that lots of people seem to like this. Reviews of the R25As have either ignored the question or have liked the “lively” sound. And some people have praised them for percussive sound especially.

In any case, this is a smaller matter than the frequency-response effects, as such. But if you are touchy about this, listen carefully to some piano sound to get the idea. This makes much less difference on sustained material, however.

Good News

It is important to realize that the R25As do a number of important things really well, whatever issues I might have had with absolute neutrality or lack thereof. As already noted, the R25As do a fine, indeed a spectacular job of stereo imaging. And they do this in a way that is of real musical significance. Orchestral music, for instance, is properly separated spatially, which adds to the listener’s ability to hear individual lines. (Ernst Ansermet’s comment in the early days of stereo about how the main musical job of stereo was separation of parts came to mind.) Comes to that, the R25As do a fine job of keeping complex textures straight and unconfused. It is hard to say what is involved here technically, but the lack of auditory confusion is striking and musically rewarding.

In the Telarc Cleveland Quartet recording of Beethoven’s Opus 59, No. 2 and No. 3, the four instruments are always separated in space and identity, but the whole has a unity at the same time. It really sounds much like a string quartet does sound in a gratifying way. (This is a good recording for those looking for the repertoire.) And when one moves up to more complex multi-lined material, the gestalt remains excellent. The Swan of Tuonela on Reference Recording’s Reveries CD (Minnesota Orchestra, E. Oue cond.) sounds both beautiful and convincing, not only when the solo English horn is playing with sparse accompaniment but also when the whole orchestra speaks in response. And the big moment is just big, not in any way muddied up. This feeling of effortlessness remains when even bigger orchestral moments arise in other repertoire. Again, the clarity without any harshness and without any sense of strain (as long as volume is kept rational—this is after all a small speaker) is musically gratifying. Even the heaviest scores of R. Strauss and Mahler stick together while remaining unconfused (the recording permitting). The sense that so often turns up of things falling apart, the center not holding when the going gets intense, is blessedly missing to an extent that is remarkable, especially in a speaker of this size. (It was a bit of a surprise to me to find this in a speaker with so little internal damping. I would have expected that the after-sound that one hears on the piano recordings would obscure slightly complex orchestra textures. but for whatever reason, the clarity is there.)

At very close range, one hears the somewhat complicated acoustic relationship among the two drivers and the tweeter geometrically between. But at a reasonable listening distance, the speaker becomes very coherent and of a piece. And the sound is very “un-speaker-like.” These descriptions may seem like standard audio-review boilerplate, but in this instance, they are actually true. The long work in adjusting crossovers and choosing crossover parts which has been involved with the design seems to have paid off—something has paid off, in any case. (It is always hard to reverse-engineer, to know what part of the design process caused what, but the audible effects are there no matter what the reason.)

One starts to understand how other reviewers have decided to ignore issues of complete neutrality and concentrate on what the speaker does well—because the things it does well it really does do well, and they are musically significant. It is perhaps silly to call Beethoven and Sibelius “fun,” but in some sense the R25As really are “fun” to listen to or, to use a better word, a joy.

Long Thoughts

As I was writing this, I recalled the passage on page 102 of Issue 335 where our editor RH discusses how even speakers at the grandest price and size range (in this case the Göbel Divin Noblesse) have a certain distinctive character of their own, likely to be different from other large and expensive speakers. This is perhaps not surprising. Indeed, it even seems inevitable in the context that just looking at on-axis response alone, never mind the effects of off-axis and room sound and all the rest, the threshold for audible difference is about ±0.1dB for broadband shifts in response. All speaker designers are in the position of having to choose among a multitude of possibilities, even if one supposes that they have complete control, which, of course, they do not.

But the odd thing is that people, even faced with the fact that there are going to be inevitable differences, become devoted to one or another possibility. We all want to believe that the speaker we like is somehow “right” in a way that others are not. Even among people who agree largely, impassioned disagreements arise. Have a look at any forum where different versions of essentially the same speaker design are discussed.

In this spirit, I am in fact quite convinced that the R25As are too “midrangy.” But I am prepared to admit that others might feel differently. Perhaps this is a moment when one should recall the lines from the Youngbloods’ “Get Together”: “Come on people now, smile on your brother.”

The R25As are a careful work, refined over a long period of time and aimed at a specific aesthetic in sound, as is any serious speaker design. In this case, the aesthetic seems to include the maximization of midrange life in a speaker that is, even so, nearly full range, moderate in size, and high enough in sensitivity to be driven by lower-power amplifiers.

If the latter point, in particular, is of compelling interest to you, then you definitely need to audition the R25As. If you are not determined to stick with lower-powered amplifiers and do not mind a little EQ to collar the midrange, the R25As are still of interest for their imaging, coherence, and bass extension in a relatively small speaker.

Specs & Pricing

Driver complement: 1x Scanspeak 26mm (1″) soft dome tweeter, 2x Scanspeak 170mm (6.5″) coated-paper mid/bass drivers
Frequency response: 35Hz–22kHz, ±3dB
Sensitivity: 94dB
Nominal impedance: 6 ohms, nonreactive
Dimensions: 215mm x 1170mm (including podium) x 270mm
Weight: 42 lbs. (each)
Price: $6995/pr.

LIVING VOICE
Long Eaton
Derbyshire, UK
44 (0) 115 9733222
livingvoice.co.uk

BORDER PATROL (U.S. Distributor)
(301) 705-7460
borderpatrol.net