A good place to start is to describe what an HLCD speaker is and is not. There's the debate over whether or not HLCD tower speakers sound good at close range. Some do sound better than others. But the primary objective of an HLCD is to sound good at 80 feet. If the HLCD performs as it should at 80 feet there is little question that it won't sound musical nor pleasant at 5 feet. There is nothing tranquil about being 'in' the boat with HLCDs firing in your ear. The human auditory system is most sensitive to a pain threshold between 1 kHZ and 5 kHZ. Generally a horn tweeter in an HLCD abruptly and dominantly enters into the picture in the heart of our most sensitive region. Given that the HLCDs are overhead and heard off-axis by occupants in the boat, this serves to subdue some of the strident aspects of an HLCD speaker at near-field. While some will continue to argue this point, you'll see why close range listenablity isn't likely with an HLCD speaker as you read more.
HLCD stands for 'horn loaded compression driver'. An HLCD tweeter is constructed with a powerful motor structure and oversized high frequency diaphragm that is coupled via a compression chamber to the reduced throat of a horn. A phase plug is used at the mouth of the compression chamber to create an equal path length from all parts of the larger diaphragm so that a flat pressure front is produced at the throat of the horn. The reduction creates a very high degree of force or compression as its called. The waveform under considerable pressure travels down a long horn that flares to the large mouth of the horn. The horn is an acoustical transformer that matches the transducer output to the compliance of air. The length, diameter and progression of flare are all complex and crucial ingredients in the design of a horn. In the process the high frequency speakers enjoy a significant increase in efficiency and resulting output. In the end a horn allows a small diaphragm to move a large mass of air.
There is a lot of knowledge and complex mathematics involved in modeling and building an effective horn tweeter. As applied to tower speakers, one horn cannot cover a bandwidth exceeding three octaves and maintain considerable output plus power handling. Every option is either going to leave a gap at the top end or bottom end of the horn's range. A good designer would conceal these deficiencies. Not many speaker builders really understand the dynamics of horns.
Sometimes this horn tweeter is mounted beside the midbass driver(s) but in most HLCDs the tweeter motor structure is behind the midbass motor structure and the long horn travels down the center axis of the midbass driver. Because the HLCD is coaxially mounted and because the technology was borrowed from the pro sound industry, wakeboard HLCD tower speakers are often referred to as Proaxials.
Many coaxially-mounted horn tweeters do not have fully developed horns and lack some of the bandwidth enjoyed by larger horns which are mounted beside the midbass driver(s). In contrast, the side-by-side configurations lack some of the coherency as you move off-axis. Neither approach is without some degree of compromise.
In order to understand why an HLCD offers a unique solution to the demands of projecting 80 feet to a wakeboarder you first have to understand the unique challenges.
There are multiple factors that affect the upper frequencies when projecting over a long-distance behind a wakeboard boat.
- The non-linear absorption rate of air, which is multiplied with distance
- Air turbulence caused by wind and chop.
- Masking caused by the engine exhaust plus close and long range sources of ambient noise. Masking is most prominent at coincidental frequencies.
- Diffraction of upper frequencies off the irregular surface of choppy water.
Again, all of these factors have a non-linear impact. Its impossible to predict the exact impact and how it varies over the audio bandwidth given that these conditions are different with each boat and on any given day. Technically, even humidity and air temperature can have an impact plus a large discrepancy between the air and water temperature can play a role, although temperature and humidity have a negligible effect in this context of much larger issues. So the best that a tower speaker manufacturer can do is approximate what the average listening distance and volume will be and what the average environmental conditions will be. You can't compensate for everything when the acoustical landscape is forever changing.
Another challenging factor is that sound loses about 6 dB of output as you double the distance. It takes four times the power to make up for a 6 dB loss. A speaker is likely to lose at least 25 dB over an 80 foot distance plus another 4 or 5 dB of high frequency absorption plus the losses due to noise masking and other environmental conditions.
For speaker manufacturers, this application is loaded with design conflicts.
Here is a rough illustration to show the type of power that's needed for wakeboarding with a HLCD speaker. We could easily get carried away with countless qualifications but I'll keep it simple with analysis flaws intact.
Let's start with a tower speaker that produces at least 90 dB average output (not peak) at one watt at one meter. From that we'll extrapolate that it will take 5 watts to maintain 90 dB at 5 feet while overcoming the masking effects of normal ambient noise, wind and exhaust. 90 dB is a good level considering its well over the sound pressure level of normal speech and should be plenty loud enough for the music to be intelligible. Four identical tower speakers driven by identical power would sum for an additional 6 dB, or a collective 96 dB output. Sound pressure level is reduced by 6 dB with every doubling of the distance which translates to four times the amplifier power to maintain the same output. So at 80 feet it would take an estimated 600 watts to produce 90 dB (plus or minus 3 dB) which easily keeps us in the realm of intelligibility. That's 150 watts per speaker which is on par with the average thermal power handling of most HLCDs. And, we know from experience after installing numerous tower systems, that 150 honest watts per speaker is a good number to produce articulate music at 80 feet and without overdriving the tower system.
There's no question that the HLCD treble sections, with their extraordinary efficiency and output, can easily get it done. The biggest challenge is for the more conventional midbass driver to keep pace with the horn tweeter and play high enough into the midrange to smoothly splice with the higher output tweeter. Creating the right midbass driver and a crossover to seamlessly join the two sections is filled with design contradictions. A serious improvement in one parameter can mean with certainty a considerable compromise in another. A horn tweeter is going to have a limited span (not much more than three octaves). So the midbass driver must reach upward into the upper midrange spectrum while also being expected to produce upper bass from within a small displacement pod. These goals are definitely conflicted with physical laws.
While it may be a poor analogy , it's not a lot different from the challenges of designing a tire. The softer rubber with more grip wears out quicker. If the tread pattern provides the ultimate performance in a very specific application (like dry traction) it may be dangerously inadequate for another (like rain). So the final design has to make some compromises.
Here are a few examples where taking the design of a midbass driver to one extreme can hurt other design objectives.
- A lighter mass cone and surround may be more efficient. The downside to a light duty surround is it may not stand up to the rigors longterm of the marine elements and the intense pounding that tower speakers take. A lighter cone is likely to have a higher resonance which will result in less midbass production from within a smaller pod. A thinner underdamped cone is likely to be acoustically transparent and let the sonic reflections through that are created within the small pod, particularly at wavelengths that are coincidental with the pod's internal dimensions. It doesn't make sense to compromise so many performance areas just to gain a couple of decibles in the S.P.L. race.
- A smaller voice coil will generally be more efficient. The fact that the HLCD horn travels down the center axis of the midbass driver dictates that the voice coil bobbin must be of a minimum diameter. A larger voice coil will have greater thermal capacity and greater power handling. But a midbass driver with an inordinately large voice coil will have significantly more coil inductance which limits its high frequency extension and ability to smoothly splice with the HLCD.
Again, the speaker engineer is faced with conflicting parameters and must prioritize or find what he considers to be the best balance. The engineer who is overly occupied with a singular objective is likely to make a speaker that sounds bad. There's no magic pill. You can't cheat physics. And we haven't seen any radical changes in technology or materials in the last several decades that could significantly impact tower speakers.
A separate midbass driver with a hard dust cap or continuous cone has more surface area versus a midbass driver that has a coaxially-mounted tweeter horn passing through its center. As a result the independent midbass driver would presumably produce better midbass than would the coaxial (or ProAxial) version. And again, as usually there will be trade-offs with the phase coherency of side-by-side positioned midbass and tweeter drivers versus the coaxial product. Every advantage inversely produces a disadvantage in some other respect.
A larger midbass driver has greater leverage in an open-field environment. It's no big secret that surface area rules. But there is a limit to how big of a tower enclosure (pod) you can market. Most people are not going to accept a couple of 5-gallon tupperware containers on their tower. There's a question as to whether you want to dominate the tower and boat appearance or compliment it.
You could probably squeeze 10-inch drivers into many of the pods that currently contain 8-inch drivers but it wouldn't make for a good overall speaker. Any off-the-shelf prosound 10-inch driver is intended to mate with a larger horn than tower pods will facilitate and this 10-inch was also intended for a huge bass-reflex enclosure. If the midbass driver is efficient enough to compete with an HLCD then it cannot thrive in a comparatively tiny tower pod.
For midbass extension there has to be a positive ratio between the driver's surface area and the pod's internal displacement. There is an unavoidable relationship between bass extension, enclosure displacement and efficiency. As a matter of physics, the sum of these elements are fixed so as a designer you can improve one characteristic of a speaker but at the detriment of the other elements. If you design more performance into one area you unquestionably compromise one or both of the other performance areas. You could even make a moderate improvement in two elements but with disastrous consequences to the third area.
There is a similar and interdependent relationship between low frequency bandwidth and high frequency bandwidth. It's impossible to build in extra performance at one extreme without sacrificing the other. A direct radiating midbass driver is restricted to not much beyond a five-octave bandwidth. It has to reach the lower cut-off point of the HLCD. Stretching the bandwidth of the midbass driver will add serious flaws. There's going to be many trade-offs with larger midbass drivers that do not justify the benefits.
Placing an oversized driver in an undersized enclosure results in an elevated Qtc and an inordinately higher resonance, which can increase the bass roll-off while creating some distracting midrange aberrations. To solve the midbass extension issue a speaker that produces well in a smaller enclosure has less midrange top end and falls off radically in efficiency, which is not what we're looking for in a tower speaker.
Also, the larger midbass cones usually exhibit a more pronounced 'break-up mode' at the top end of their midrange response. This abnormality is lower in frequency and closer to the heart of most musical fundamentals as the speaker size grows. To some degree this mode is found in all midrange drivers, its just more pronounced and harder to control or conceal with larger drivers. Since a horn is usually only effective over a three-octave range, we're highly dependent on the midbass driver for a lot of upper midrange (at least up to 2 kHZ in larger speakers with massive horns).
At the upper end of a midbass driver's response, its polar pattern begins to increasingly narrow and it tends to beam with very little off-axis radiation. The larger the driver, the worse the effect.
Also, you have to consider the size of the motor structure it would require for a highly efficient 10-inch driver. The extra weight of overly-massive tower speakers will eventually create serious problems with the tower joints and connecting hardware. Over a broad range of towers, 75 total pounds is the estimated limit before you'll experience issues. The collars would have to be either massive or very expensive (machined from stainless steel versus aluminum) to safely support the extra weight.
Crossovers are a major challenge for the designer of any speaker but are particularly difficult in trying to reconcile so many contrasting objectives. Off-the-shelf crossovers are not an alternative. HLCDs are normally packaged with larger midbass drivers in large bass-reflex enclosures. A crossover that is designed to mate drivers in a typical prosound application is not going to be able to effectively address the mismatch in drivers within a wakeboard tower speaker.
You can use a crossover to totally tame an HLCD speaker but why would you want to? An attempt to turn an HLCD speaker into a sound quality speaker would only cancel the S.P.L. and projection advantages inherent in an HLCD.
You can over-engineer the passive crossover. Even though the crossover can be designed to create a seamless transition between the midbass driver and the HLCD, the complex crossover can have enough insertion loss to lose considerable efficiency. But who wants an inefficient tower speaker?
So for a crossover to be effective in a tower speaker it must be relatively simple with fewer power-consuming components. This is not an easy task considering all the problems a designer would like to resolve with the crossover but cannot without driving down the efficiency. Here are just a few of the issues that a crossover would normally rectify.
A crossover has to:
- Equalize and splice the amplitude response of two distinctly different drivers.
- The crossover has to manage an opposing and transitioning phase rotation of the high and lowpass sections for a coherent joining of two drivers.
- Speakers don't have a linear impedance. The impedance of a midbass driver is rising just at the point of the intended roll-off. The rising impedance serves to nullify the sloping effect of the crossover. The tweeter has a nasty resonant impedance peak typically an octave below the intended -3dB point. With a changing impedance curve we have a changing amplitude response and a changing phase response. These variations cause a re-biasing of the crossover network values. A crossover must address the reactive nature of speakers.
The above is just the short list of crossover design considerations. Do all HLCD manufacturers have the technical wherewithal to build good crossovers? Nope. Some do it on purpose, some accidently get close and others just take what their importers send them.
ADDING DECIBELS WITH SIMILAR VERSUS DISIMILAR TOWER SPEAKERS
Decibels are logarithmic units and can't be added like other numbers. If you have one source producing a sound pressure level of 65 dB and you add a second 65 dB sound source, you don't have a level of 130 dB, you have a combined level of 68 dB (plus 3 dB). Two equal 85 dB sources will combine for an 88 dB level (plus 3 dB). This seems easy enough when both values are the same, but the conclusions are much less obvious when the levels are different from each other. When the levels of two sources with equal amplitude are combined the overall level increases 3 dB or the equivalent of doubling the power. But when the two levels are dissimilar the combined output is less than 3 dB.
|Sound Source One
||Sound Source Two
||Combined Output||Additional Gain|
|100 dB||100 dB||103 dB||3 dB|
|100 dB||97 dB||101.76 dB||1.76 dB|
|100 dB||94 dB||100.97 dB||0.97 dB|
|100 dB||91 dB||100.51 dB||0.51 dB|
For instance, if a 100 dB 8-inch HLCD is combined with a 90 dB 6 ½-inch conventional coaxial you'll gain less than a ½ dB. That's a terrible waste of speaker, power and particularly investment.
It would normally take ten times the amplifier power for the lesser speaker (90 dB) to equal the louder speaker (100 dB). However, the lesser speaker is likely to dynamically compress and the amplifier driving it is likely to clip long before levels can be matched. If there is an initial difference in efficiency then it will be hopelessly inefficient to make up or correct for the difference. In fact, its likely to require much more than the theoretic power difference (2 x power = + 3 dB) to achieve a balanced output. In the end whatever investment was made in the secondary speaker has to be considered a terrible waste to obtain less than ½ a dB. A small increase in investment to upgrade to symmetrical speakers would have been highly cost-effective. The effectiveness of various combinations will vary but its unavoidable that systems with dissimilar speakers are generally a bad approach.
There are exceptions to every rule. There may be an application where asymmetrical and smaller speakers are mounted to the outside of the tower and aimed outside of the wake in order to significantly broaden the dispersion pattern. In this particular scheme the highly directional polar patterns of the inside and outside speakers are not shared and are not going to combine well anyway from an S.P.L. standpoint. So, this would be a permissible exception because there is a different objective versus maximum S.P.L. and maximum projection.
Another example of an exception is when two different speakers are purposely used, like a small ProAxial and separate midbass driver, in order to obtain a more balanced or higher sound quality combination. In this case the objective is something other than the ultimate S.P.L.
A 1 dB change in amplitude is generally acknowledged as the smallest discernable increment by the human auditory system. So if the independent outputs of two speakers are different by more than 3 dB the summed increase in output is too minimal to be beneficial. This supports why its important to use similar speakers, positioning and power but also places significant importance on the correct tuning process.
On a different tangent but still connected to this issue, there are those tower combinations that use similar type (HLCD for example) and similar size but dissimilar speaker models in some respect. In one instance, the high frequency section or horns may constitute the only significant difference. Since 50 percent of all musical fundamentals are between 200 and 600 hertz and fall within the sole domain of the midbass driver, tuning should be performed with the treble section partially attenuated. This technique will yield a more balanced average output and will produce more collective output throughout the spectrum that counts most. This is in contrast to tuning the multi-speaker tower system based on its dissimilar and peak output. The acoustic laws of 'adding decibels' support this approach. This approach will not only produce more average output but will generate a more coherent overall response and improved sound quality. The treble attenuation is only applicable during tuning and lifted before normal use.
Also, the summing calculations assume that all speakers are of equal type so that the phase response is identical. When you start putting together creative combinations of dissimilar speakers, the phase response is random. A speaker configuration of dissimilar size, type and efficiency with random phase responses is going to combine for an ever more disappointing tower system. Usually the most coherent sounding system is the simplest.
Comb Filter Effect
Tower speakers are generally mounted in a horizontal line. We have to take the mounting provisions that the wakeboard tower gives us. A horizontally oriented array is particularly poor in its off-axis performance because of the comb filter effect. This can be described as:
The various speaker's output sums and subtracts irregularly as we move from center to any position off-axis. At the root of this effect is a condition where the distance between the speakers is a multiple of a frequency's wavelength. Directly on axis, the output of two speakers will be in-phase. Off-axis, at some point in space, any frequency can arrive at the listener out-of-phase. As we move from center to outside, the out-of-phase cancellations will fall in the domain of different frequencies because we've altered the frequency pathlength to the listening postion and we've altered the frequency pathlength in relation to multiple side-by-side sound sources. The negative comb filter effects are certainly compounded as you add more speakers in more complex configurations.
Wakeboard Tower Open-Air Sound Transmission
The following guidelines are only generalizations as there are a number of applicational circumstances that will vary the outcome such as wind and exhaust noise. For instance, its hard to predict the environmental noise created by a particular boat with a specific ballast at a specific speed and distance on a given day. Plus, sound does not travel and dissipate in a linear manner. In other words, different frequencies are affected differently over different distances.
Sound emanates in a non-linear polar pattern, which means all frequencies are not dispersed equally. The pattern of dispersion changes as you move off-axis and the pattern narrows as you increase the distance. So, different frequencies are transmitted differently depending on the range and degree off-axis.
Masking effects by competing wind and exhaust noise impact coincidental musical frequencies the most with less impact on alternate frequencies. Acoustical masking can be described as one sound (noise, for example) that is close in frequency to another sound (music, for example) and when 6 dB louder it masks the lower amplitude sound.
Loudness usually decreases by 6 dB as you double the listening distance. This means that you would need four times the power to maintain the same amplitude at twice the distance.
|Distance||Amplitude||Power required to maintain the same amplitude||Equivalent power in increased decibels|
|5 Ft.||90 dB||5 watts|
|10 Ft.||90 dB||20 watts||6 dB|
|20 Ft.||90 dB||80 watts||12 dB|
|40 Ft.||90 dB||320 watts||18 dB|
|80 Ft.||90 dB||1280 watts||24 dB|
The above example is predicted on a single speaker only.
Here's a table to show how the absorption rate through open air is non-linear. The following attenuation rates are taken at 100 feet.
|100 Hz||0 dB|
|500 Hz||.1 dB|
|1000 Hz||.2 dB|
|2000 Hz||.3 dB|
|3000 Hz||.5 dB|
|4000 Hz||.8 dB|
|5000 Hz||1.1 dB|
|6000 Hz||1.6 dB|
|7000 Hz||2.1 dB|
|8000 Hz||2.6 dB|
|9000 Hz||3.3 dB|
|10000 Hz||4 dB|
|11000 Hz||4.8 dB|
|12000 Hz||5.6 dB|
Ideally a wakeboard tower speaker provides the accentuation over a specific bandwidth and the correct slope to offset the attenuation shown in the above table. The above table is an illustration of just one of many circumstantial conditions of the wakeboarding environment.
The human auditory system perceives sound by different mechanisms over different spectrums of the audio bandwidth. Not so coincidentally, sound is propagated with different transmission characteristics over different areas of the audio bandwidth.
Low frequency energy is absorbed to a very small degree but the intensity is rapidly reduced over distance as it is dispersed in an omni-directional manner. In an open-field environment, having only one reinforcing plane, plus the fact that tower speakers are positioned well above the plane or water, low frequencies will dissipate too quickly to reach the wakeboarder with any real intensity. There's never going to be enough driver surface area, pod displacement or power to attain adequate leverage to get serious bass at wakeboarding range, at least from tower speakers.
However, a tower speaker with very good midbass extension can deliver plausible midbass at wakesurfing range.
High frequencies because of their shorter wavelength and resulting narrow polar pattern, will behave differently and for different reasons. The obstacles in correcting the delivery of high frequencies are much easier to solve.
Acoustic Perceptions Related to Non-Linear Speakers
Some brands of HLCD speakers have been labeled with a specific deficiency and it kind of catches on where people have an expectation and when hearing a speaker for the first time the experience seems to reinforce a preconceived notion. The bias can be for a number of reasons. Buy oftentimes this same criticism may apply to all HLCD speakers, regardless of brand. And, sometimes the conclusion is wrong.
For example, most people conclude that the midrange 'honk' that accompanies many HLCD speakers originates from the horn where in reality it most often is a product of the upper 'break-up mode' of the midbass driver. They're just not trained to audibly recognize what particular bandwidths sound like and they're not knowledgeable in crossovers and the ranges of given low and highpass drivers. Its easy to see a big horn and blame the horn.
Another speaker may get the reputation of having weak midbass performance. But in reality the midbass output and extension may be equal to or better than other speakers thought to have better midbass.
R.T.A. measurements often contradict our psycho-acoustic perceptions and conclusions. There are some psycho-acoustic factors and masking effects of the human auditory system which dictate that a few accentuated frequencies tend to dominate our perception and that these significant non-linearities, particularly peaks, can effectively mask near-frequency areas that in actuality are not attenuated. Or, accentuated areas can also conceal deficiencies. There is often a difference between acoustical measurements taken with highly sophisticated electronic equipment and how the human auditory system perceives and interprets the same stimulus. This is especially true with the non-linear nature of HLCD tower speakers.
You might not hear the lower midrange warmth when an upper midrange 'honk' is considerably louder. You might not be able to tell that a horn tweeter does a poor job of reproducing frequencies above 10 kHZ when the horn tweeter is brutally bright just one octave away.
In my experience, its just a few tonal aberrations that tend to dominate our auditory perceptions and form what we judge to be the overall character of a particular speaker.
There are always going to be unavoidable trade-offs in prioritizing and designing a true high performance tower speaker. Sound quality versus amplitude. Midbass extension versus efficiency. Projection versus dispersion. Other objectives might include sonic adaptability, security, mounting flexibility, cosmetic integration, size, weight, cost, just to name a few. An improvement in one element almost certainly means a compromise in another element or multiple elements. The challenges of designing an HLCD are far greater than the more balanced objectives in building a conventional speaker because of the need for such an extreme in amplitude from what is still a relatively small speaker as compared to typical prosound insdustry products.
The "best" tower speaker doesn't necessarily have to be the loudest or the most accurate. For some the "best" might mean the most balanced design or the "best" might be the speaker that best addresses a particular need for one individual who is more willing to sacrifice other attributes.
Conventional speakers cannot compete with how HLCDs project with intelligibility at 80 feet. When describing an HLCD as intelligible or articulate I mean that the rider can recognize a clear melody and even understand lyrics (provided they're clear to begin with). By comparison multiples of conventional speakers may still be heard at wakeboarding range but without the necessary amplitude and high frequency distinctive qualities. Most conventional tower speakers will sound like they have a wool blanket thrown over them at 80 feet. Its really hard to get too critical or particular about individual performance characteristics when you see what an HLCD is expected to do. Comparing an HLCD to a true sound quality speaker is pointless because SQ speakers can't come close to what HLCDs do so well.
Send us an email if you have questions about this article
Hear HLCDs in Action. -Video-