The LS6 combines old-school BBC-style monitor design with modern driver technology to deliver a very classy loudspeaker indeed.
Over the last few years, Graham Audio have developed a successful niche in reinterpreting classic BBC monitor designs, and I wrote a piece back in the June 2016 issue about their reincarnation of the LS3/5. That Chartwell LS3/5, as it was christened, has now been joined by a larger sibling, the Chartwell LS6, and that’s the subject of this review. The LS6, however, marks a compact monitoring departure for Graham Audio because, rather than being a reinterpretation of a BBC classic, it’s an an original design. Well, I say it’s an original, but there’s no ignoring the traditional look of the LS6, and the clear inspiration it takes from an earlier age of monitoring. There’s no ignoring, either, the fact that its designer, Derek Hughes, is the son of Spencer Hughes, who was a core member of the engineering team responsible for numerous BBC monitors (and later the founder of Spendor). So, the LS6 is an original monitor design that obviously benefits from four decades of driver development, but in many respects it has roots planted firmly back in the 1970s. Is that a bad thing? Read on.
The LS6 is a passive monitor, and despite the domination of the monitor market by active products, there’s still a market for passive units. I’ve no definite numbers to back this up, but my gut feeling is that passive nearfield monitoring is undergoing something of a renaissance at present. We’ve written about a few in the magazine recently, and they’ve all acquitted themselves vey well. Perhaps it helps that a consequence of so many studios still having a pair of Yamaha NS10s in residence is that they also still have a power amplifier available.
But back to the LS6: it’s generally a conventional two-way passive speaker with a 180mm coated-polypropylene-diaphragm bass/mid driver and a 19mm soft-dome tweeter. The tweeter is actually the same unit as the one employed in the Chartwell LS3/5. Both drivers are custom-designed units from Norwegian specialist company SEAS. A passive crossover with second-order low-pass and fourth-order high-pass slopes integrates the drivers at 4kHz. The crossover circuit also includes some elements dedicated to gentle equalisation of the overall system response. ‘Voicing’ a monitor though passive equalisation is a real skill that I suspect will, within a couple of decades, be mostly a thing of the past.
At 24cm wide, 36cm high and 26cm deep, the LS6 is a trouble-free size for nearfield monitoring installation. Its cabinet is finished all over in a tastefully retrained wood veneer and incorporates a large, generously exit-flared reflex port directly behind the tweeter on its rear panel. The only other feature on the rear panel is a pair of binding-post terminals. Around on the front panel, alongside the tweeter, is a switch surrounded by a small and suitably ‘70s-looking trim panel, which provides an optional +1dB or +2dB tweeter level adjustment. Magnetically secured fabric grilles are supplied and the user manual recommends that they are fitted for listening, so that’s what I did.
Before I write about listening to the LS6, there’s a little more to its cabinet than nicely veneered panels, and I think this deserves some description and explanation. Back in the 1970s, during the heyday of the BBC’s Kingswood Warren monitor-design department, the whole business of recorded or broadcast music was not quite as it is today. Recording audio was predominantly about accurately capturing a sound or a performance, and miles away from the creative post-process we have now, where ‘recording’ is only the start of a long and often tortuous road.
The term ‘hi-fi’, short of course for ‘high fidelity’, expresses perfectly where the recording motivation lay: in truthfulness. I hate to say this, but I fear that these days it is all too easy to lose ‘sight’ of the actual sound of the instruments and voices in the room (if indeed there ever were any) long before the end of the production process approaches. But back at the BBC in the 1970s, the principle of ‘high fidelity’ was important because recording and broadcast was predominantly of unamplified, or minimally amplified, acoustic instruments and natural voices in live acoustic spaces. So rather than listening in perhaps the way we do now — to compression characteristics, reverb tails or Auto-Tune artifacts — listening at the BBC was predominantly about being truthful to the source. A human voice announcing the news, or a piano at the Proms, needed to sound as if it was real and in the room. So the fundamental tonal fidelity of a monitor was emphasised above all else, and this was reflected in the BBC’s design philosophy. The tonal fidelity of a monitor is, however, not just about its frequency response. It is defined, among other things, by how the monitor stores and re-radiates acoustic energy, and often, the biggest influence in that respect is the the cabinet.
The significance of the cabinet vibrating along with the audio signal can be appreciated simply by comparing the radiating area of, say, a 180mm bass/mid driver, like that fitted to the LS6, with the potential radiating area of the cabinet enclosing it. The ratio is typically at least 30:1 in favour of the cabinet walls. If you factor in a modest musical dynamic range of, say, 60dB, which in ratio terms is 1000:1, it’s pretty obvious that the cabinet panels don’t have to move much to potentially make a significant contribution to the sound of a speaker.
The group of engineers working on monitor design at Kingswood Warren knew all this, and much research was carried out into the suitability of different cabinet materials and constructional techniques in order to identify those that had the most benign acoustic signatures. Perhaps the best-known example of cabinet research, a paper co-written by Dudley Harwood (who went on to found Harbeth), can be found at www.bbc.co.uk/rd/publications/rdreport_1977_03, and if you’ve always seen monitor cabinets as simply the boxes that the drivers are mounted in, it makes for eye-opening reading. Who would have thought, for example, that different species of wood used for corner fillets would be measurable in the acoustic signature of a cabinet, and audible too?
The result of all the BBC’s cabinet research was a constructional technique that came to be known as ‘thin-wall’, and it’s the technique Derek Hughes has employed on the LS6. The clue to how it works is in the name. Rather than attempt to create an inert cabinet by employing thick, rigid panels, the BBC thin-wall technique employs relatively thin (9 or 12 mm) panels that are stiffened with corner fillets and damped with strategically placed mastic elements. The arguments for thin-wall design are, first, that lower-mass panels are far easier to damp than heavy rigid panels and, second, that the vibration that still takes place will be at a relatively benign frequency. So, when you knuckle-tap a thin-wall monitor cabinet, like that of the LS6, you’ll hear just a dull thud, rather than the high-pitched ping of a really stiff braced cabinet. Once I’d finished listening to the LS6s, I removed a bass/mid driver from one and photographed the internals. Beneath the impressively engineered passive crossover, an area of black mastic damping is visible.
I’ve personally never been entirely convinced by the thin-wall technique, preferring a cabinet-engineering strategy that majors more on rigidity while still aiming for relatively low panel mass, and dealing with any resultant panel resonance though targeted mass damping. Having said that, however, if speaker design teaches one thing, it is that there are many ways to skin a cliché, and I can’t deny that many fine BBC-style monitors, the LS6 among them, have used the technique. So who am I to express doubts?
I rather gave the game away there by including the LS6 among my “fine BBC-style monitors”. However, before I elaborate, along with knuckle-tapping the LS6 cabinet, I also broke out a rather more high-tech analysis tool and made some FuzzMeasure acoustic measurements to illustrate a few aspects of the LS6’s performance. The LS6 is in many respects a conventional two-way passive speaker, so the acoustic characteristics it displays demonstrate quite well how similar systems fundamentally behave. To begin with, however, I’ve chosen to illustrate something that few manufacturers of monitors or hi-fi speakers really like to talk about: pair matching. If you ever get a trip around a speaker factory, ask about pair matching and watch the faces fall. If there are any aspects of speaker engineering that could be considered glamorous, troubleshooting of pair matching and long-term consistency is not one of them.
Pair matching, simply a measure of how similar the two monitors of a stereo pair are, can be pretty revealing of the true quality of a monitor and manufacturer because making two speakers the same, and even a whole production run the same, is a fiendishly difficult thing to do. The problem, fundamentally, is that the moving parts of drive units are necessarily constructed from lightweight materials that are asked work to beyond the limits of their mechanical rigidity, so they’ll bend and deform. Even worse, diaphragm materials tend to be temperature- and humidity-sensitive, and can display a range of mechanical properties depending on how warm they are and how damp the air is. But moving mass probably illustrates the issue most clearly. For example, say a driver diaphragm and voice coil together weigh 15 grams. Leaving aside the potential for the mass of either to vary, as part of the manufacturing process (through variation in diaphragm material thickness for example) they need to be be rigidly stuck together with a bead of adhesive. If, as a result, the total mass of the assembly varies by, say, 5 percent, that will lead (all other things being equal) to a change in driver sensitivity across the whole band of around 1dB.
Keeping control of driver moving mass is just one of numerous other similar challenges that monitor manufacturers face, which is why measuring pair matching is a useful pointer towards how serious they are: achieving good pair matching demands serious manufacturing and quality-control commitment. So, Diagram 1 shows the frequency response from 200Hz upwards of the two LS6s. Compared to many ‘pair’ matches I’ve seen, it’s pretty good. There are a few small regions where the two curves diverge, but even then, the degree of divergence is less than 1dB; ±0.5dB over 200Hz to 20kHz is a good result.
The second measurement I took examines Graham Audio’s advice that the grille should remain in place while listening. Diagram 2 shows frequency response curves from 200Hz upwards, both with and without the grille fitted. The general result of the grille is to attenuate the output by a dB or so from around 10kHz upwards, but there’s a second effect: discontinuities in the response around 3kHz and between 7kHz and 10kHz. These discontinuities are almost certainly the result of diffraction from the edge of the grille frame beneath the fabric. Discontinuities such as these, which result from the slight ‘smearing’ in the time-domain response caused by the diffractions, are of course not ideal, but probably not severe enough to be significant.
Moving on to the third and fourth measurements, these illustrate the LS6’s dispersion characteristics, and they are very typical of two-way speakers with spaced drivers. Diagram 3 shows the dispersion 15 degrees horizontally away from the perpendicular axis, and there are two features to note. Firstly, the dispersion of the bass/mid driver begins to narrow from just over 2kHz, and secondly, the tweeter dispersion does the same from around 10kHz. Both of these phenomena are fundamentally due to the wavelength of the radiated acoustic energy approaching the dimensions of the bass/mid cone and tweeter diaphragm’s dome. The narrowing mid-range dispersion demonstrates particularly how two-way speakers sometimes struggle to cover the full bandwidth: the bass/mid driver dispersion falls away just before the tweeter can come to the rescue. The LS6 is actually pretty good in this respect, thanks to its very well-behaved and relatively small bass/mid driver, but mid-range dispersion is one of main the arguments for three-way speakers.
Diagram 4 also illustrates dispersion, but this time 15 degrees vertically from the perpendicular, both upwards and downwards. These vertical dispersion measurements clearly reveal a fundamental issue with all non-coincident driver speaker systems: the phase relationship between the drivers changes with listening (and measuring) position in the crossover region where their outputs overlap. The result is a degree of destructive interference that leads to significant discontinuities in frequency response. I wrote ‘significant’, but it’s quite often the case with speakers that characteristics that appear so are actually not quite so bad. Big ‘suck-outs’ in frequency response curves caused by destructive interference are an example. There are two phenomena at play that help. First, our ears and brains are far cleverer than measurement microphones, and we don’t perceive sounds from instant to instant: we integrate sounds over periods of the order of 10ms (the exact ear integration time varies with frequency, level and a host of other parameters). Bearing in mind that the suck-outs in the frequency response will vary significantly with listening position, there’s great opportunity for the brain to hear a different version of the response. Secondly, we don’t listen in anechoic chambers, we listen in live rooms, so again, shortly after the direct sound that arrives at our ears first, there’s a host of early reflections that will each be imprinted with a different version of the speaker’s frequency response. The ear and brain integrate all these arrivals and we perceive an ‘average’.
Moving on, there are two more measurements to explain. Diagram 5 illustrates the output from the LS6’s reflex port. The measurement was taken with the measuring microphone positioned right at the mouth of the port. The blue curve illustrates the port tuning frequency at 55Hz, and shows that the port is very well behaved in terms of pipe resonances. The measurement illustrates clearly, however, that it’s not just bass that comes out of a reflex port. The green curve shows the response from the same microphone position but with the port blocked. I made this measurement to check that features in the open-port measurement are not simply the result of ‘crosstalk’ from the drivers.
There’s one more measurement to go before I cut to the chase. As part of my listening, I tried the LS6 with its ports blocked, but having decided to do so, I wanted to see what the resultant low-frequency response of the system would be. Accurately measuring the low-frequency response of a reflex-loaded speaker is not easy: the multiple sources (port and driver) and long wavelengths involved mean that the only option is to have a very large empty room. However, if the port is blocked so that there’s only one audio source (the driver), placing the measuring mic in very close provides an approximate measure of the LF vital statistics: cutoff frequency, approximate Q and roll-off slope. So, Diagram 6 shows that with its port blocked the LS6’s cutoff frequency (-3dB) is around 70Hz, with the -6dB point at about 55Hz. The roll-off slope is very gentle, at around 11dB/octave (it would decay further to 12dB/octave as frequency falls), which suggests that the fundamental LF resonance is over-damped with a Q of less than 0.7 — that’s certainly how it sounds. I’ve also added on Diagram 6 a green dotted line that illustrates the LS6s’ specified low-frequency response (which is specified with the port open, of course).
Which brings me neatly and finally to describing what I heard from the LS6 with a variety of CD, Spotify and Pro Tools Session listening. I’ll start with the LS6 bass, partly because that’s what I discussed in the last paragraph and partly because that’s the only element of the LS6 I was unsure about. The LS6 is, of course, a reflex-loaded monitor, with a port tuning frequency that’s relatively high (55Hz, if you fell asleep for that bit), and that’s what it sounds like. Although its low-frequency bandwidth is very well extended, especially considering the cabinet size, bass on the LS6 to my ears is not as sharply defined in terms of either time or pitch as I would have liked. It’s perhaps becoming a bit of a cliché to report this, but bass with a port-loaded character really does make it more difficult to be certain about the balance of, for example, bass guitars and kick drums.
The other sides of this coin (yes, it’s a weird three-sided coin) are, first, that the bandwidth extension achieved by reflex loading sheds more light on what’s going on at very low frequencies, and second, that with the LS6, there’s always the option of blocking the port. The LS6 with its port blocked displays low-frequency performance of a decidedly different character, which I far preferred. The down side of port blocking is that cone movement-limited power handling will be reduced, but that hasn’t stopped quite a few passive speaker manufacturers from supplying optional foam bungs to do the blocking job. Maybe Graham Audio could follow suit?
With its ports blocked, the LS6 for me was transformed from a monitor that I found not easy to get a grip on, to one that I had few doubts was delivering reliably and consistently what I needed to hear. Its mid-range, especially in the voice band, was, as perhaps would be expected considering its heritage, extremely natural-sounding, while at the same time revealing genuine depth of fine detail. There’s something uncannily realistic about the way the LS6 does voices. It’s a quality that’s reflected in the stereo imaging performance, too: voices hang there in space, as if the invisible man (or, indeed woman) has taken up residence. This fabulous mid-range performance to me speaks not only of a very high-performance bass/mid driver and skilful driver integration through the passive crossover, but also of a cabinet that’s not significantly playing along with the signal. The LS6 is a win for thin-wall cabinet design.
Moving on up, I had no doubts about the tweeter. I remember thinking while listening to the Chartwell LS3/5 that its custom-designed SEAS unit was a fine driver, and its use in the LS6 did nothing but reinforce my opinion. The tweeter matches the unforced detail of the bass/mid driver with a similarly revealing nature that does the necessary, but without imposing itself. I tried listening to the LS6 for a while with the grilles removed but found, as Graham Audio suggest, that the high-frequency balance became just a little overcooked. It’s perhaps a shame that the front panel doesn’t offer any attenuation options, because then the option of listening without the grille would have come more into play.
Perhaps the big question to ask in summing up the review is this: does a passive monitor that might have been designed in the 1970s (although the LS6’s contemporary drivers are a few steps ahead of 1970s fare) really have any relevance today? It’s a question that I was particularly motivated to muse upon while writing this final paragraph, because the next monitor I’ll be reviewing, already in position either side of my DAW, is one of the most up-to-date and high-tech speakers available. The answer I came up with was, yes, of course the LS6 has relevance, because fundamentally it is a skilfully engineered and classy-sounding monitor. If you have a suitable power amplifier or are willing to invest in one, there’s absolutely no reason why the LS6 won’t provide exactly what’s needed of a nearfield monitor.
The LS6 is obviously up against passive monitors from ATC and Amphion — the SCM20 and One18 respectively. Monitors, similarly inspired by BBC design techniques, from Harbeth and Spendor would also be worth considering.