I’d like to explain the logic used to reach the conclusion that a LLT subwoofer (large and low tuned ported) offers the best overall package of performance per dollar and discuss what this type of design requires to meet the stated goals.
When comparing the trade-offs of different subwoofer alignments, ported subs are typically known for a reasonably flat frequency response to ~tuning, relatively high output level capability, and to many people, the inability to offer detailed, articulate, or tight bass. What’s troubling is that it seems most don't really bother to examine why this is - instead, differing designs of the same alignment get lumped together in stereotypes. Several different adjectives for sound quality can be used, but what it really comes down to is accuracy, or reproducing the signal that is sent to the sub without adding or taking anything away. There are numerous types of measurements that can be conducted on a subwoofer, but when you start talking about real world subwoofer usage in room and not 4pi or 2pi environment measurements, the numerous performance aspects can be aggregated into in room frequency response and non linear distortion at the desired playback level. All other performance parameters are really just a function of these two in room. Theoretically, the most accurate subwoofer would be one that is flat in room from 1hz to well above your crossover point and can play at any required output level with no non linear distortion. This is what we would ideally hope to achieve. That said, there is a population of enthusiasts who prefer a less accurate representation – a colored playback. From my findings, this typically means an early roll off in room and high THD levels. The combination of high THD levels and a de-emphasized low end will result in artificial emphasis in upper bass, and this is often described as "musical" or "punchy" bass. One should stick with the sound quality characteristics they prefer, but for the sake of hoping to achieve the most accurate recreation of the original material, we will ignore this population. Along the way, I will look at things from a cost effective point of view.
So what causes some ported subs to have poor accuracy? In the range where port becomes the primary source of output, some negative side effects accompany the benefits. Transient response becomes degraded due to a phase shift between the driver and port as well as the seemingly necessary use of a high pass filter in what should still be the sub's operating range. The port itself may contribute to some audible nonlinearities (output compression, chuffing, first port resonance) if it does not have enough cross sectional area, and since most ported subs are designed to measure flat anechoically, when they're used in room, the resulting FR often has a large hump in output down low, which usually results in what is described as “muddy” or "slow" bass. When the driver is asked to produce output below the tuning frequency, the port works against the driver, robbing it of output, and the resulting roll off is 4th order, or a decrease in output by 24db/octave. The addition of a 2nd order (or higher) high pass filter combines with this 4th order roll off to create a very steep rolloff. The higher the order or steeper the roll off, the worse the transient response will be, as there will be lingering energy when a low note gets played. Ideally there would be no roll offs, filtering, or crossovers in our audio systems, and a single driver could reproduce 1-20khz effectively, but currently that’s not a realistic goal.
Additionally, most ported subs are tuned in the 25hz range, and tuning this high fails to provide enough solid extension to cover the infrasonic range which has become commonplace in nearly all of of today's blockbuster films. This relatively high tuning combined with a typical 12" driver leads to increased excursion use and distortion in the 30hz range, and skyrocketing distortion below tuning. While the bulk of music doesn’t contain infrasonics, capability in the infrasonics is still beneficial, as the negative effects that occur as result of port use typically affect frequencies up to an octave higher, sometimes even more than that. This means that a ~25hz tuned sub could be coloring bass reproduction up to 50hz.
Those are the bulk of the ill effects associated with ported subs, and they explain why - based on most commercially available ported subwoofers – some people have reached less than great conclusions about ported subwoofer sound quality. So what can be done?
One idea has been to eliminate the use of a port and its side effects and just go sealed, and that does solve a few problems, but the lack of port output creates a set of compromises all its own in nearly all rooms. With a sealed sub, assuming a rigid and well built enclosure, only the driver is responsible for output. The amount of air that needs to be moved to reproduce low frequencies increases exponentially as frequency lowers. This means that excursion demands will also increase exponentially if we want reasonably flat response from a sealed subwoofer, as the cone area of the driver remains constant. As the cone starts moving more air by utilizing more excursion, the backward stroke of the driver starts to pressurize the enclosure, and at some point, based on the internal volume of the enclosure and driver parameters, cone movement is actually resisted. Prior to any equalization, a sealed sub will tpically have a 2nd order roll off usually beginning around 40-60hz, again based on enclosure size and driver parameters. This 12db/octave roll off wouldn't be too bad if it didn't begin naturally until ~20hz, but even that still wouldn't be ideal. The solution that many people use is equalization to boost the low end or cut the top end (the results are the same either way), resulting in the driver outputting more balanced output levels across a wider frequency range. To overcome the spring force in the enclosure when lower frequencies are being reproduced, more power is needed. The more power that is needed to reach adequate output levels, the more heat that is generated in/on the voice coil of the driver. The more heat that is in/on the voice coil, the more THD increases.
Since driver excursion demands are much less at higher frequencies than lower frequencies at a given output level, and small enclosures (typical of most sealed subs) decrease low frequency sensitivity, the THD relationship between upper and lower bass is polarized. By boosting the low end with EQ, you end up with an exponentially disproportionate rise in THD. You are also essentially limiting the sub by its weakest point, which is based on how much power the EQ'd sub can use before reaching excursion limits too early. Headroom is traded for more extension, which isn't a bad trade if you have plenty of headroom to begin with, but along with the extra extension, THD levels are exponentially increasing, relatively speaking. A beefy sealed design will also require the purchase of a very beefy amplifier and some means of EQ – this takes away from the cost effectiveness. What you're faced with is either a poor in room FR in average to large-sized rooms, disproportionately high THD in lower frequencies, or the need to build multiple sealed subs with multiple beefy amps $$$
To help reduce some of these sealed problems, drivers with very high power handling and very high excursion limits have been produced. The higher power handling will reduce the overall amount of distortion, but not the relationship in distortion, so distortion at low frequencies will still be disproportionately higher than at higher frequencies. The increased cone travel of high excursion drivers can result in BL nonlinearity distortions unless a low distortion motor technology is used. Ultimately, you can get a driver with a low distortion motor technology, high excursion limits, and very high power handling to overcome most of the issues with a sealed alignment, but now you’re looking at spending a pretty penny. And as you’ll soon see, these top notch drivers can actually be put to better use in a different alignment, extracting more performance for your dollar.
An infinite baffle subwoofer overcomes some of these issues by using an extremely large enclosure that makes the driver free to move with little to no resistance from air spring, only the driver’s own suspension. The low end sensitivity is increased, resulting in a more naturally linear FR (lower system Q), keeping a slightly better relationship between frequency and THD. The increased low end sensitivity means that less power is needed, but output is limited to the amount of power that can safely be used without exceeding excursion at the lowest usable frequencies. Because of this, IB is a better choice than traditional sealed in regards to accurate bass reproduction, however, headroom is often a legitimate issue. Due to the lack of an air spring, the power requirements need to be low enough to prevent the sub from reaching over-excurson at some low frequency, and this limits output capability. This can be overcome by using multiple drivers, but it is a very inefficient use of drivers and not very cost effective.
Assuming you want efficient use of your driver(s) and the best performance for your dollar, you will want to take a second look at ported. Ported will give a big boost of clean output around tuning, so you can focus that output where a sealed design starts getting stressed, keeping THD low and maintaining solid output. All you need to do is overcome the issues I mentioned with traditional ported subwoofers.
How can we make transient response, low end excursion utilization, negative port effects, and in room FR better? Since they are all significantly affected by the tuning frequency of the enclosure, the answer is to lower the tune beyond what is typically used. Using a very low tune keeps the responsibility of output away from the port over a larger range of bass frequencies, allowing port output to come into play only when the driver starts showing signs of becoming stressed. A logical goal is to make the driver solely responsible for output down to at least ~30hz, as the frequencies above that are where the bulk of musical bass lies and where accuracy is most critical. Using the one octave estimation, this would ideally put the desired tune at no higher than 15hz. The lower tune means that the 4th order roll off (which occurs below tuning) occurs at a much lower frequency as well, and a high pass filter and its accompanying negative side effects can be avoided. Standard DVD players, receivers, preamp processors, EQ units, amps, and other electronics in the signal chain have output roll offs in the single digits. Some equipment is better than others in this regard, but I’m not aware of any non-modified signal chain that will be flat to low single digits. This aggregated roll off acts as a layer of protection below the sub's operating range so that you aren't sending the sub a full strength signal from the amp at extremely low frequencies.
Here are measurements of the low end rolloff of a few pieces of equipment in a typical signal chain. The first is a collection of units measured into the single digits by Ilkka - the others (an Emotiva CD player, Behringer amp, and Classe DVD player) only extend to 10hz with minimal rolloff, but one can tell from the slope that the attenuation will be significant by 5hz.
With very low tuning, the negative effects associated with port use don't get a chance to play a role in the audible range, and the bass quality in a proper design is a direct representation of what the chosen driver is capable of. The more capable the driver, the better the accuracy will be. The ill effects of the port are pushed down into frequencies so low that they basically become non issues, as those frequencies are mostly tactile as opposed to audible - our sensitivity to those frequencies is extremely low, as shown by hearing sensitivity curves:
As you can see, sensitivity decreases exponentially in subwoofer frequencies. If driver distortion is minimal, the primary issue that could potentially affect frequencies up to an octave higher or more is the transient response. This can be attacked two fold. First
, the lower tune pushes the degradation of transient response down into lower frequencies, where sensitivity is decreased. This can be observed in spectral decay measurements of a subwoofer using multiple tuning options. Here is a SVS 20-39 PC+ in 20hz tune:
Notice how high in range this degradation reaches. Now here is the same measurement of the same sub under the same circumstances with the tuning lowered to 12hz.
As you can see, the degradation in transient response has been pushed down far, and now, the audible region appears to be totally unaffected. Second
, due to no high pass filter being used in a LLT, the overall scope/range of this degradation is reduced as well, due to a 4th order roll off as opposed to a 6th or higher order roll off. To illustrate this point, here are some measurements of how long it took a 20hz tuned ported sub to reach full amplitude with a 20hz sine wave – first using an amp with a discrete high pass filter built in at 18hz (Right Channel, bottom measurement):
And then again, same circumstances, but using an amp with no built in high pass filter:
There are obstacles that accompany using a really low tune - nothing comes for free. The port issues can potentially become amplified, as we're now looking at pushing much more air through the same port area - compression and chuffing can become major problems - and since the port actually has to be longer to create that lower tune, the first port resonance can be low enough as to be audible based on the subwoofer's operating range. In addition, sensitivity in the musical range can take a hit, and the driver output doesn't match up well with the port output - this means the FR is no longer flat. How can this be alleviated? Use a larger enclosure.
The larger enclosure will increase the low end sensitivity, allowing you to keep a flatter response into lower frequencies. By fine adjusting the enclosure volume and tune, you can mold an anechoic FR that is flat throughout the bulk of the bass range with a shallow roll off starting at ~25-30hz, extending down to the tuning point, with the 4th order roll off occurring somewhere below that. This shallow roll off area is key, as most rooms typically yield about 4-8db of room gain/octave. The lower the sub can extend while maintaining an anechoic FR with this shallow roll off, the more solid and flat the in room response will be. This is the main reason LLT owners have such extremely flat and low extending in room FR measurements - measurements that are almost too good to be true without the use of EQ boosting below ~25hz.
The larger enclosure also enables you to use a much larger diameter port - this reduces audible chuffing and output compression limits and essentially eliminates them as a concern altogether, as the limits will often be well below what the larger port allows. Add a rounder to the outer layer of MDF to act as a DIY semi flare, and the audible chuffing limit is a thing of the past. Some have even built their own MDF flares to use on the internal port opening as well.
I’d highly recommend you read this
informative study on compression and audible chuffing limits with ports using various diameters and flare radii by Bill Collison – also download his handy calculator to use as a supplement to normal simulation programs. With a larger enclosure, for a given length of port, the diameter can be bigger while keeping the same tune, making it easier to keep the first port resonance at least above 190hz, where it won't be a problem with a typical 80hz crossover. The larger size of the enclosure also allows you to have lots of clearance between the interior port opening and any surrounding planes, whether that is an MDF wall or brace, a cardboard sonotube, or even the driver itself. You want to have at least as much clearance as the diameter of the port itself in all directions from the openings. Another benefit of the large enclosure is a lower system Q, or higher damping. The higher the damping, the more accurate the bass reproduction is - this ties back into the deep and relatively flat extending response. Here is a brief explanation of system Q by Collin Miller of Secrets of Home Theater and HiFi from his essay "How a Hole in the Box Works - A Big Dig into Bass Reflex":
Damping is described mathematically as a Q value. In loudspeakers, it describes the ratio of energy storage at resonance, to dissipation. The electrical properties have a Qes. The mechanical properties have a Qms. When both are combined to the cabinet’s Q value, a system Q is derived. High Q systems store energy for a long time, resulting in a longer delay and higher amplitude at resonance, while low Q systems get rid of energy quickly, providing a shorter group delay and lower amplitude at resonance.
The entire essay can be found here
- it's a good read. Now again, some prefer a colored sound – in this case that would mean a high Q.
The increased low end sensitivity (less enclosure spring force) of a larger enclosure means less power is needed to achieve the same output levels at low frequencies than in a smaller enclosure. So this time, we are trading some higher frequency headroom for low frequency extension like before, but we're not getting the disproportionally high rise in THD like a traditional sealed sub due to the port output playing a role. We are also getting much deeper, stronger bass. Some have inferred that a really low tune causes higher THD levels above tuning at equal output levels because more excursion is needed to reach those same levels. This is not an accurate assessment. In the higher frequencies, when port output isn’t coming into play, output is solely a result of air displacement from the driver, just like having the enclosure sealed. Because of this, the same amount of excursion will always result in the same amount of output – all that is changed with larger enclosures and lower tunes is the sensitivity, or how much power is needed to achieve the needed excursion. In addition, since all of the negative effects of traditional ported design have been pushed deep down into lower frequencies, and less power is needed, you gain large advantages in reduced distortion.
To visualize this concept at its most extreme, model two enclosures of the same size using the same driver, one ported and one sealed. Feed them both with the same amount of power and tune the ported sub to ~2hz. From about 7hz on up, they behave identically in just about every performance parameter – FR, group delay, output, and excursion, and the port velocity from the ported design won’t yet have come in to play. What does this mean? This reinforces that when port output is not in play, a ported and sealed sub behave virtually identically. The key is to utilize that port output while minimizing the noise distortions and transient response degradation as much as possible so that they are inaudible, reaping all of the benefits without suffering the negative side effects - having your cake and eating it too. The LLT design allows you to do just this. The resulting natural FR works better in room than any other alignment in terms of achieving a flat FR naturally, and for the usable extension you are getting, the THD stays lower than sealed at any given output level, as the sealed sub isn’t getting any relief. Unless one is using enough drivers to get them flat with low distortion to 1hz at their desired playback levels while sealed (very inefficient usage), there is always room for improvement with a port.
There has to be a catch in being able to get all these positives with seamingly no negatives, right? The size of the LLT has to be large, sometimes very large for proper performance. ~300 effective liters for most 15" drivers is common, and ~650 effective liters for 18" drivers. Not everyone is willing to live with this.
What makes a LLT different from an EBS? A LLT has to meet certain minimum specifications to ensure a certain performance, whereas the EBS definition is broad and encompasses almost any design.
- Tuning <=15hz
- Cylindrical port diameter >= 6” for 15" drivers and >=8" for 18" drivers
- Port opening clearance equal to at least the port diameter in all directions
- First port resonance >=190hz
- No discrete high pass filter
- FR with a ~4db/octave rolloff between the tuning frequency and the lowest room node
- Minimal resonance enclosure
When finalizing your design, you may often find that adding an additional X liters lets you extend 1hz deeper or gives you 1db more headroom - that may be the case, but keep overall performance balance in mind as well. What does that extra volume do in relation to excursion demands and max headroom from the amp you have chosen? How about port velocity? I recommend going big, but you'll eventually reach a point of diminishing returns, and past that, performance will actually take a hit. In terms of desired FR, this picture should give you a good idea of what you are shooting for.
In terms of power, recent measurements performed by Ilkka and the realization that real world playback does not consist of continuous tones have caused me to reconsider my previous amp limiting guidelines. I used to think that one should aim for an amount of power that straddles the rated xmax of a driver in the given enclosure as seen in simulations without exceeding it. This guideline was faulty for three reasons. 1) The xmech, or mechanical limit of a driver's suspension, is much higher than the xmax, usually 1.5x or more. This means there is freedom for more excursion at the cost of nonlinearities, but that extra cushion of excursion is better put to use than having to clip your amp, as the resulting signal sent to the voice coil can be damaging. 2) Simulations show excursion used with a given amount of power based on a continuous playback of each frequency, like a sine wave. This is not an accurate representation of how the driver will behave with real music or movies, as few if any have extended continous tones. Excursion use will be less than the simulations predict with real material at a given power level. 3) A driver's suspension, even if very loose, will tend to resist movement more as it approaches its limits. This acts as another layer of protection.
Because of this, I now recommend using as much power as is economically feasible, with the only caution being fear of frying your voice coil. If you are constantly pushing your driver to extreme playback levels, you will need to be weary of this happening. It's better to put your money towards a second sub than it is to put it towards an immensely powerful amplifier, as the better amp will only get you so far without risk of damage.
That's it! For questions, please use the LLT Explained Discussion Thread