At around £20,000 a pair, Meyer's new X10 monitors lie well beyond the budget of most SOS readers — but the technology inside them represents a breakthrough in loudspeaker construction, and should filter down to more affordable products over the next few years. Hugh Robjohns demystifies the design...
California‑based Meyer Sound Laboratories have acquired an enviable reputation over more than two decades as a manufacturer of exceptionally well‑engineered, high‑quality, robust PA loudspeakers and stage monitors. They have also developed a range of advanced tools — their SIM system, for example — to help optimise and align the amplitude and phase responses of their sound systems for the environment in which they are installed, whether permanently or as a touring rig.
Meyer Sound and their chief designer, John Meyer, have been responsible for a number of innovative products over the years, such as the active HD1, a two‑way nearfield monitor. This was the commercial result of five years of research aimed largely at achieving an accurate impulse response from a conventional speaker (see the box on page 140 for more on impulse response).
In many ways, the latest addition to the Meyer product line, the new X10 control‑room monitor, is an evolution of many of the acoustic and engineering principles developed in the HD1. However, it goes far beyond anything previously developed for studio monitoring systems. The X10 specifically addresses, in radical and revolutionary ways, most of the technical limitations of existing large farfield monitors. Essentially, Meyer have produced a well‑controlled, very low‑distortion compression driver/horn system for the top end, and combined this with a bespoke long‑throw bass driver incorporating a state‑of‑the‑art servo‑feedback system which reduces distortion dramatically. If terms such as compression driver and servo feedback are unfamiliar to you, don't worry — they'll be explained later.
The X10's asking price is currently around £20,000 per pair after installation and room tuning, so you can be forgiven for thinking that few SOS readers are likely to invest in a set for their home studio. However, the innovative technology employed in the X10s represents a substantial step forward in loudspeaker design and, as it is inherently scalable, it will inevitably filter down to more modest products in the near future. As we can confidently expect that these products will soon include project‑studio speakers based on the same design principles as the X10, those principles surely merit closer examination. So what makes the new Meyers so amazing?
John Meyer's primary goal was to produce a very high‑resolution studio monitoring system that would be capable of reproducing the same sound in the control room as that produced in the studio, both in terms of linearity and sound level. Theoretically, such a system would make comparisons simple, and recording techniques could then be judged fairly.
Historically, the process of improving linearity in loudspeakers to minimise distortion has been slow and incremental, particularly where moving‑coil transducers have been concerned. The mechanics of these systems are so complex and unpredictable that few, if any, engineers really have a good understanding of their working dynamics. Almost inevitably, most moving‑coil drivers impart relatively high levels of distortion, and this has come to be accepted as the norm. It is not uncommon for the bass driver of a full‑range monitor loudspeaker to produce over 20 percent harmonic distortion at high sound‑pressure levels. The problem is that the distortion artefacts manifest themselves as unwanted energy, mainly in the mid‑range. That means reduced overall clarity, resolution and detail, and often leads to the impression of a rising frequency response through the mid‑range. In compensating for this kind of monitoring problem, people often create mixes which don't translate very well between different monitor systems.
However, thanks to the advanced and innovative proprietary techniques Meyer Sound have developed, the X10 speaker produces less than three percent of distortion at elevated sound pressure levels of around 130dB, and only a few tenths of a percent when operating below about 90dB SPL. This is a whole order of magnitude better than most conventional designs.
Many big studio monitors employ dual 15 or 18‑inch bass drivers, in order to generate sufficient acoustic power at low frequencies. However, the wave fronts from a pair of drivers will interfere above about 250Hz which forces a crossover to a mid driver below 250Hz and then, typically, on to a high driver above, say, 2kHz — the classic three‑way system. Following their experience with the HD, though, Meyer designed a two‑way system, because creating an accurate impulse response from three drivers proved virtually impossible.
Although a two‑way hi‑fi speaker is relatively simple to produce, a high‑quality system capable of 130dB SPLs is rather more of a challenge. A single bass driver requires an enormously long throw in order to move sufficient air, but then needs a colossal magnet to maintain a linear magnetic field across such a long gap. On the high‑frequency side, dome tweeters are simply not capable of generating the peak SPLs required without tearing themselves apart or introducing huge levels of distortion. Only a compression driver (more on this below) fits the bill — but that requires a horn or waveguide which, traditionally, suffers from higher distortion
I'll return to the tweeter problem shortly, after looking at Meyer's solution for the bass driver. Their first goal was to make the motor itself was as quiet as possible. To this end, an air‑cooled bass driver was developed in contrast to the majority of high‑power systems which use ferrofluid — moving fluid was found to generate too much noise!
To move the required amount of air, the 15‑inch driver required a massive four‑inch diameter voice coil with a one‑inch excursion. Creating a linear and consistent magnetic field over that kind of area is not easy — we are talking about four times the magnetism of a conventional 15‑inch driver! Consequently, no fewer than 28 neodymium magnet slugs are mounted in two concentric steel rings at each end of the motor assembly. The result is an enormous magnetic field (it measures a vast 16,000 Gauss, or 1.5 million Maxwell).
However, the practical result is that this unique long‑throw, low‑distortion 15‑inch bass driver can generate the same acoustic power as an equivalent dual 18‑inch system, but with the advantages that it generates far lower distortion, and can fit in a cabinet a quarter of the volume.
Feedback servos are a common element in everyday technology. The servo control essentially monitors the output of any given system and uses that information to generate a correcting signal at the system's input. In an amplifier, for example, the output is analysed and 'negative feedback' sent back into the input to help minimise distortion. Applying this idea to a loudspeaker, you might expect the output signal to be measured in some way (say by means of a mic placed in front of the speaker) and the information so gathered somehow to be used to generate corrective signals at the speaker's input. Sure enough, this is basically how the X10's bass driver cuts down distortion.
Naturally, there is more to it than this in practice — integrating a feedback servo system with a reflex‑loaded loudspeaker is far from straightforward, mainly due to non‑linearities caused by the resonance of the tuned cabinet as well as those in the driver itself. This is a very complex problem to solve, and one which has defeated all previous attempts. Indeed, many audio veterans will recall some of the disastrous attempts in the past — the one I remember best was Philips' 'motional feedback' which, although quite effective with a slow swept tone on the test‑bench, sounded like it was underwater when replaying music!
Meyer had experimented with a servo design during the development of the HD1, but without success. However, very sophisticated mathematical techniques called Mu‑Control (see box on page 138) have been developed for the design of complex servo‑feedback systems and they are now capable of dealing with the complexities of a loudspeaker system. Meyer Sound call their implementation of this servo system 'Pressure Sensing Active Control' (PSAC) and it was developed in conjunction with the University of California in Berkeley.
Applying and refining the PSAC technology took around a year. Before the servo system could even be designed, a 'model' of how the complete loudspeaker/cabinet worked had to be developed, as the servo can only generate its correcting signal if it is 'aware' of the way the loudspeaker system works. This is, clearly, the hard part of the problem, but once all the various controlling inputs (air pressure, amplifier current and voltage) and desired outcomes (minimal distortion, flat frequency and phase response) have been accurately specified and modelled, a series of mathematical equations describing the servo are derived. From these equations, the electronic circuitry (or even DSP algorithms) can be constructed to implement the servo in hardware.
The main controlling input to the PSAC system is feedback of the air pressure wave emanating from the driver. This pressure wave is measured with a microphone mounted on a bar directly in front of the bass driver. This has to withstand SPLs in excess of 150dB on peaks and requires a flat response down to 1Hz to maintain stability in the servo controller. Other feedback information is derived from current and voltage sensing at the output of the bass amplifier which ensures the drive unit remains within safe working limits.
The use of a microphone as the main controller input means that the servo system is, to some degree, inherently sensitive to external sound, including room reflections. When I was shown the Meyer X10s, I was amazed to notice that the bass cone moves visibly in response to the low‑frequency pressure change caused by opening a door to the listening room. In practice, this means that it is imperative for the system to be carefully aligned and tuned to its environment during the installation.
One of the problems which defeated earlier attempts at feedback‑control systems was the apparent delay — it takes a finite time for the speaker diaphragm to move in response to an input signal. Intuitively, using a microphone a few inches in front of the cone to sense the pressure wave it creates would seem to make this situation even worse. However, at such low frequencies we are not really talking about delay but, rather, a phase shift. The Mu‑control design technique allows the servo to be designed to accommodate this.
When it came to designing the high‑frequency driver, Meyer decided that a compression unit feeding into a waveguide was the only way to achieve the required SPLs for a large studio monitor. Compression drivers are specialised moving‑coil loudspeakers with a metal diaphragm. They are designed to operate with a tiny diaphragm excursion in a high‑pressure environment created in the throat of a horn or waveguide. The expansion of the horn not only controls the dispersion of the signal, but also matches the impedance between the high‑pressure area in front of the diaphragm and the atmospheric pressure outside. Compression drivers are inherently very efficient and can achieve SPLs far greater than any other loudspeaker technologies (see diagram above).
Meyer developed their own compression driver for the X10 — a custom‑built four‑inch device with an aluminium alloy diaphragm. One of the innovative aspects of its design is that the diaphragm is supported by a hybrid suspension arrangement which has been designed to remain quiet when it moves, unlike the folded metal suspension designs of most equivalent compression drivers.
Conventional pressure‑driver and horn systems are often associated with relatively high levels of distortion (the precise profile of the waveguide necessary to produce a linear, low‑distortion output is both critical and complex) but Meyer Sound have overcome this through the combination of their in‑house compression driver coupled with a painstakingly developed waveguide, which took over a year to perfect.
Conventional horn designs tend to become very directional at high‑frequencies making integration with a low‑frequency driver virtually impossible. Part of Meyer's solution involves a very carefully developed profile in the throat of the horn, where there is an expansion area in front of the compression driver. A restriction at the mouth of the throat then leads into the main body of the horn. Finally, the horn flares out to the front with what is referred to as a 'Constant‑Q' profile. This complex form is very difficult and expensive to manufacture, but Meyer claim that it is essential to achieve the low distortion and uniform off‑axis response they desired.
At modest SPLs, say around 90dB, the Meyer combination of bespoke horn and compression driver produces well below 0.2 percent distortion — similar to that of a good soft‑dome tweeter, which is an impressive achievement. At sound‑pressure levels around the 130dB mark its distortion is only slightly higher, and is considerably lower than a dome driver could manage at the same level. The system also has the advantage that overloading results in a progressive rise in distortion, whereas a dome system breaks up very suddenly. Needless to say, it is also able to operate continuously at levels which would destroy a soft dome.
During their research into how best to obtain accurate time‑alignment, Meyer discovered that two factors need to be managed simultaneously. The acoustical alignment defines where the sound appears to be coming from, and the impulse response determines how well the wave fronts from the two drivers reconstruct in the monitoring environment. These two separate factors can be tackled independently, but they both have to be right if the speaker is to perform accurately.
Achieving the correct acoustical alignment means arranging for the drivers to be in the same vertical plane. Since a compression driver is physically mounted behind the baffle at the rear of the horn, it is inevitably quite close to the plane of the bass driver. The alignment is perfected through the use of delay circuits within the crossover and can be easily checked by measuring the frequency response at different distances from the loudspeaker in an anechoic environment. Ideally, all frequencies should drop in level with increasing distance by the same amount — a uniform broad‑band decrease of 6dB with each doubling of distance indicates that sound is emanating from a common focal point and so has perfect acoustical alignment.
If the alignment is not accurate the frequency response will only be correctly balanced at one specific distance from the monitor, causing unacceptable and unpredictable results in different listening environments. The proof of the pudding is that the HD1 (Meyer's first attempt at precise acoustical alignment) is apparently accurately balanced out to around five meters, after which the linear change in amplitude across all frequencies breaks down. However, this is acceptable in a nearfield monitor. The X10 builds upon the fine results of the HD1 and has a near‑perfect acoustical alignment.
The physical construction of full‑range electrostatic speakers means they are inherently accurately time‑aligned and also produce a very good impulse response — far better than that of conventional moving‑coil loudspeakers. However, the attention paid to time‑alignment in the X10, together with careful honing of the system electronics, has bestowed the X10 with an impulse response which is notably better than that of any electrostatic. Indeed, the acoustic output from the speaker is virtually indistinguishable from a test impulse applied to its input.
There is great debate about the importance of accurate time‑alignment, but Meyer's view is that if it is measurable it should be correct in a properly engineered product, whether its effect is audible or not. My personal opinion is that correct time‑alignment is critical for the accurate reproduction of complex transient‑rich sounds — most notably percussion and brass instruments — which the X10 (and HD1) reproduce extremely well.
That the X10 represents a substantial step forward in loudspeaker technology is beyond doubt. The PSAC servo might appear to be the most revolutionary aspect of this new loudspeaker, but the design of the horn profiles, the bespoke compression driver, innovative bass unit and the method of precise time‑alignment are all equally advanced.
However, technology for its own sake is pointless, so the real issue here is: does the X10 sound significantly better than more conventional control‑room monitors? I have auditioned many state‑of‑the‑art monitoring systems, of all kinds, in some of the best acoustic spaces in the world, so I believe I have a reasonable grasp of what is currently available and considered to be 'the best'. I first auditioned the X10s in a less‑than‑ideal listening room at the Meyer Sound offices in Berkeley, California, but even there, their phenomenal transparency and accuracy was evident. Subsequent auditioning of a 5.1 installation of X10s in a properly designed control room reinforced these first impressions.
Distortion is simply not an issue with these loudspeakers, at any practical listening level — it is as if the control‑room window has been removed and you are listening to the acoustic sound in the studio itself. The X10's astonishing clarity also reveals flaws in recording equipment, techniques and signal processing astonishingly well. By way of a test, I listened to some material processed through a well‑known predictive data compression system. This kind of system damages unpredictable parts of signals — mainly transients — and whilst these artefacts usually go unnoticed or are hard to detect on most monitoring systems, they were certainly clearly audible on the X10s.
Stereo imaging and depth is pinpoint sharp, precise and stable over a remarkably wide listening area, and the character of the sound remains absolutely consistent where ever you are in the room. Whether this results from the almost perfect impulse response, the very low distortion, the well controlled off‑axis dispersion, or a combination of these I could not tell you. All I know is that it works, and works extremely well. Instruments with naturally wide dynamic ranges and percussive transients are reproduced with breathtaking clarity and naturalness — no other monitoring system I know of reproduces brass instruments with as much fidelity.
The X10 is a most impressive monitor which, through the introduction of significant new technology, redefines the art and science of loudspeaker design. Meyer Sound have repositioned the goal posts for the high end market. Let's hope it's not too long before the techniques employed here filter down to more modest studio reference monitors.
In the 1970s, computational 'optimisation processes' began to be employed to help design complex servo‑control systems with multiple inputs, outputs and feedback loops. First, information would be gathered describing the important aspects of the system to be controlled, the objective of the controller, the available inputs and required outputs, and so on. Then all of this information — a 'model' of the desired system — would be processed by a special computer algorithm to determine a mathematical solution which described the resulting servo controller.
Frequently, however, these algorithms found solutions by taking advantage of undefined aspects of the original model, and were therefore often unacceptable. In contrast, the manual computation of experienced engineers tackling the problem in a more conventional way generally resulted in workable solutions, albeit ones which required considerable effort and often still worked sub‑optimally.
Nevertheless, the mathematics of these optimisation techniques continued to develop and by the 1980s had evolved into 'Mu‑Control' — an advanced technique able to accommodate 'uncertainty' in the model. In a nutshell, Mu‑Control determines solutions for very complex servo‑control systems with multiple interacting inputs and outputs from relatively simple descriptive models of the system, along with information about the conditions where the model is known to be accurate and, crucially, where it is not.
Mu‑Control is very good in situations where many variables are being manipulated to achieve many simultaneous objectives, all competing and interacting with each other — conditions which are often beyond the capabilities of even the most experienced team of engineers! Typical applications include determining the control systems required in the active suspension systems of road and racing cars, in satellite guidance applications, multi‑cylinder internal combustion engines, and even in the systems designed to maintain the stability of aerodynamically unstable aircraft, such as the Stealth Bomber.
As far as loudspeakers are concerned, the application of this technology to a low‑distortion bass driver is now a proven success. However, using it to enhance the performance of mid‑range or high‑frequency drivers is much more difficult. The physical separation of microphone and diaphragm introduces a phase shift even at low frequencies, but in a mid or high driver, phase shifts could easily exceed 360 degrees, leading to ambiguity and instability. However, the Meyer R&D team are currently investigating alternative methods of coupling a pressure transducer to drive units. Watch this space!
Most audio engineers think of audio equipment as systems into which energy is put, converted and output. One favourite method they have of analysing the efficacy (or otherwise) of a given audio system is by measuring frequency responses — the amplitude of signals at specific frequencies — at the output. A system is deemed to be 'accurate' if it is maximally flat across the full audible frequency range. However, there is much more to 'accuracy' than that, and the often neglected, but crucial, element accompanying frequency response is the phase response. Systems can exhibit identical frequency responses but will sound completely unlike one another if their phase responses are different!
Although analysis in terms of frequency and phase (the 'frequency domain') is a very convenient means of visualising the workings of an audio system, another equally valid approach is to consider the system in the 'time domain'. Instead of measuring performance with a series of continuous tones at different frequencies, a brief transient click, or impulse, can be used instead. If an impulse is input to a non‑perfect system, its output will be a distorted impulse: the impulse response. Much like white noise, an impulse contains all frequencies at the same time and its analysis therefore provides information on frequency and phase response.
The manner in which the impulse is distorted will be unique to a particular system and reflects precisely its characteristics. It may be harder for non‑engineers to interpret, but it is a very fast and accurate way of measuring and evaluating a system. The fact that an impulse looks remarkably similar to a single sample from an analogue‑digital converter is the reason why this technique is so prevalent in the engineering of digital systems, but it also lends itself very well to the analysis of loudspeakers. This is because its instantaneous one‑shot nature allows the loudspeaker to be measured accurately in a room, without having to worry about reflections or standing waves. There is insufficient acoustic energy to set up standing waves and room reflections can be ignored provided they arrive at the measuring microphone after the direct signal impulse has decayed.
The X10 has a dynamic range in excess of 116dB and boasts a frequency response flat within 2dB from 23Hz to 17kHz (with substantial output down to 15Hz). The system has been designed to provide peak levels of 136dB SPL with a continuous rating at an ear‑splitting 126dB SPL. Measuring roughly 30 x 31 x 21.5 inches (hwd), the X10 is suitable for either soffit mounting (ie. recessed into the control room wall) or free‑standing, its response being matched and tuned to its environment during installation using Meyer's SIM system.
Each X10 contains a two‑channel amplifier chassis at the rear incorporating a line‑level crossover along with the servo control and time alignment circuitry. A 1200‑Watt amplifier powers the low frequency speaker with a 620W amp driving the compression driver. Both operate in Class A mode up to about 40W to ensure the lowest possible distortion, after which they move into Class (the power rail for the output devices switches to a higher voltage to meet the demands of signal peaks).
The system noise floor is just 20dBA and, since the amplifiers have signal‑to‑noise ratios over 138dB, the audible noise floor is effectively defined by that of the monitoring source — only with the highest‑quality 24‑bit sources would the amplifier's noise floor start to become a factor. In other words, the X10 is capable of accommodating the full dynamic range of any audio source without obscuring low‑level detail or clipping on transient peaks at natural acoustic levels — clearly essential characteristics for a truly professional monitoring tool.