The initial words below were written by Vladimir Dorta. They are an excellent commentary on subwoofers! His words are written much better than I could place them. They address some of the discussion concerning subwoofers. I will add additional commentary at the end of his article.
MINIMONITORS AND SUBWOOFERS
I have written this article for newcomers to high end audio who want to know more about high fidelity sound reproduction. The following intends to help them make an informed decision at lower initial expense and also show a logical system upgrade path.
“For years, subwoofers have been a peripheral and niche product. With the rise of audio/video systems however, we have seen the increasing application of subwoofers for movie-like bass reinforcement. As a category, subwoofers are about to go mainstream.” This was written a couple of years back in Audio Shopper (wwwaudioshopper.com). There has been a big jump in subwoofer technology during the last seven years or so and today there are several high-quality subwoofers on the market priced between $900 and $2500, from companies like Bag End, REL, Revel, Velodyne, Paradigm, PBS, Hsu Research, NHT, Energy, etc.
Optimally installing a subwoofer means biamping (externally connecting the subwoofer through an active electronic crossover that has both low-pass and high-pass filters). Most modern subwoofers are “powered”, that is, they have all that’s required already built in one box: driver, crossover, amplifier and switching facilities. Biamping requires either pre-out and amp-in line-level connections or dedicated subwoofer connections on a Home Theater processor. Today, with the ever increasing popularity of home theater, most amplifiers, receivers and processors do have them. For stereo-only fanatics, there are some inexpensive and very fine integrated amplifiers (Creek, for example) that also have them. Unless the subwoofer is specifically designed to use them (REL for example, but this brand only makes bass-augmenting subwoofers, not analyzed here), do not use the high-level (speaker level) connections some subwoofers offer, nor connect the satellites without a high-pass filter. Doing this will negate two main advantages of adding a subwoofer to your sound or video system. But please bear with me and keep on reading.
The Subwoofer Bonus
Contrary to what most people assume, a subwoofer’s main advantage is not the boom-boom it is capable of producing. When you add a quality subwoofer, you get much more than just increased volume capability or recreation of the missing foundation of the music (the bass frequencies). In fact, a really good subwoofer will never boom, nor will it introduce low frequency sounds when they are not part of the music that’s playing, and certainly will never give you the hit-the-stops, 90% distortion, “one-note bass” that passes for bass reproduction on most car audio systems today. A good subwoofer’s main benefit is a double bonus: its ability to reproduce the original ambience (the full acoustics) of the recording venue and to liberate the satellites’ midrange and treble drivers from having to reproduce bass frequencies. Both benefits are of crucial importance.
Adding a seemingly small range of low frequencies (the lowest one or one and a half octaves, that is, 20Hz to 60Hz) does not seem to be a big thing in itself. But bass being the foundation of the music and containing most of the ambience of the recording venue, the total experience of reproduced music can only be apprehended when both the foundation and the ambience coexist. When the total ambience of the recording venue is reproduced intact, you receive a musical experience that cannot be described in words. Once you have experienced true deep bass as part of a total sound experience, you can’t go back to anything less. The best way to prove this is to disconnect the subwoofer while listening to well-recorded symphonic music. The soundstage shrinks and the new sound seems a poor imitation of the previous one. And this is true even when there is no bass in the music!
How To Make the Most of a Compromise
All speaker designs are compromises and the end results always involve trade-offs. Any speaker can only be as good as its theoretical model: a point source (sharp imaging and soundstage, small scale reproduction), a horn (big, dynamic, “they-are-here” sound) or a planar dipole line source (big, open and reverberant sound, not so sharp imaging). A minimonitor is the closest approximation to the point source ideal (both drivers in close proximity) but cannot overcome the laws of physics that require big drivers to move the air needed to reproduce bass frequencies. A subwoofer crossed over at a conveniently low frequency would add those missing bass frequencies. In this age of high quality subwoofers and minimonitors, anybody willing to go the extra mile and experiment with placement would also reap other benefits from subwoofer-satellite systems, as we will see.
A Higher Quality Speaker
Minimonitors (for the sake of a definition, small two-way speakers mounted on floor stands), have small cabinets which can be more rigid and have narrow front baffles. By being inherently more rigid than big speakers, minimonitors have less cabinet resonances, and by having small and narrow front baffles they cause a lot less diffraction of the sound waves. Panel resonances and baffle diffraction create spurious sound sources. Minimizing both is fundamental for correctly reproducing the original recording soundstage and the image of the performers within that soundstage. Small high-quality speakers are well known for their imaging and soundstaging abilities, but a less well publicized ability is their “disappearing act.” Sound appears not to come from the speakers themselves but from a wall of sound located between and behind the speakers in such a way that the sound appears to be totally disconnected from the speakers, an illusion that greatly increases the realism of the reproduction. Two-way speakers also have certain advantages over three- or four-way designs: Almost by definition, a small two-way better emulates its theoretical model, the point source, because the two drivers are very close together, not spaced all over a large baffle. Secondly, almost all speakers designers agree that it is easier to design a good two-way speaker than a good three-way one. That’s the reason why almost all the classic speaker designs have been two-ways: Rogers-BBC LS3/5A, Spendor BC1, KEF 104, Advent, Dynaco A-25, Spica TC-50. One of the big advantages (and the main limitation too) of a two-way is the wide frequency range that is reproduced by its bass-midrange driver, in other words, the absence of a crossover in the critical midrange. A three of four-way speaker has to have a crossover smack somewhere in the midrange, with all the phase shifts, lobes and discontinuities it entails. The main limitation signaled above has to do with power handling. A small 6-inch bass-midrange driver cannot handle the power needed to reproduce low frequencies at high volume levels. But this limitation of a two-way minimonitor is handily overcome by biamping it with a subwoofer.
There is also an advantage of cost. It is much cheaper to build (and therefore purchase) a small high-quality speaker than a similarly good big one. The main unit cost on any speaker is the enclosure, and to make big enclosures sufficiently rigid means a lot of work and expense. True, small speakers need heavy stands which increase their total price, but not by much. Please resist the seemingly logical temptation to place them on top of the subwoofer boxes. If you do that, everything I have said up to now will be thrown out the window. Resonances will be coming back with a vengeance by virtue of the vibration transmitted from the subwoofer box to the small speaker placed on top of it. The small baffle will not be there anymore and the resulting diffraction will play havoc with imaging and soundstaging. No more “disappearing act” due to the fact of now having a big speaker that keeps telling you: “hey, look, I am here!”
Having Your Cake and Eating it Too: Optimum Imaging and Optimum Bass
Physically separating the midrange-treble source of the minimonitor from the bass source of the subwoofer makes sense in other important ways too. In a typical room, optimum placement for imaging almost never corresponds to optimum placement for bass. Normally, most speakers image better when placed at different distances to and well away from front and side walls. The reason is that early reflections (those arriving at the listener’s ears less than 20 milliseconds after the main sound) destroy the stereo illusion by behaving as spurious sound sources. Ample separation (3 or 4 feet) from walls is most important with small speakers, in order to get all their imaging potential out. On the other hand, in most rooms the deepest, smoothest and most powerful bass is normally obtained at a room corner. As Tom Nousaine says, “The main problems with rooms and speakers at low frequencies is a relative scarcity of room modes. This leads to ‘holes’ in the response at certain locations. The corner will excite all possible room modes thereby making in‑room response as smooth as it can possibly be.” This means that you can place the minimonitors for optimum imaging and the subwoofers for optimum bass. There are so many possible combinations that you will almost certainly find good ones for both imaging and bass. You can never do this with big floor standing speakers, period. With satellites and subwoofers, you get the room to help instead of getting in the way.
Great News: You Don’t Need Two Subwoofers!
There are still more advantages. Normally, in typical rooms and with the latest 15" and even some high-quality 12" subwoofers like the Hsu HRSW12a and Paradigm 2200, only one subwoofer is needed for strong bass down to 20Hz. This is because a top quality subwoofer’s driver and enclosure are both designed to exclusively reproduce bass, and placing it at a corner will give additional room gain. A second reason is that that low bass being omnidirectional, stereo bass cannot be distinguished from mono bass below 100Hz. Bass drivers on floor standing speakers, on the contrary, are normally very compromised (or else the speaker would need to be really big and expensive). This is due to a high crossover frequency from bass to midrange (300 to 400Hz) and a relatively small woofer driver diameter (10 or 12 inches) needed to cover several frequency octaves, plus the compromise of a manageable enclosure size. Remember too, that the floorstanding speaker would have to be located some distance away from walls to give decent imaging. The result: not so deep or powerful bass, a compromise in every way.
A Low Crossover Point
The optimum crossover point between bass and midrange drivers is at or slightly below100Hz. This was established years ago by Peter J. Moncrieff of International Audio Review while testing for distortion on typical bass/midrange drivers. Below this crossover point, Doppler and frequency-modulated distortions increase dramatically.1 There is another advantage to a 100Hz crossover point. Below this frequency bass notes are omnidirectional, as explained earlier, which means the subwoofer is not identified as a sound source and can then be positioned almost anywhere in the room. Bass would still seem to come from the main speakers’ soundstage, not from the subwoofer. Martin G. DeWulf, in his review of Gallo Reference speakers in Bound for Sound, 1998, says the following: “Testing has borne out that the human ear is most sensitive to phase shifts and amplitude changes in the range of 100 Hz to 400 Hz; that’s two full octaves. The first person to bring this to my attention was speaker designer extraordinaire, Bud Fried. For that reason, Bud tried to stay away from crossovers in this range.” Almost all three- and four-way speakers have a crossover somewhere in this range.
Biamping with a low crossover point also causes less cone movement of the midrange driver (it doesn’t have to move a lot to reproduce bass frequencies), which makes for a clearer, lower distortion midrange sound. This is immediately apparent at first hearing. The midrange being the most important part of the frequency range, the entire sound is now perceived as more detailed, relaxed and dynamic than before. This, in combination with the better imaging and soundstaging, is worth the whole trip.
Cheaper or Better Amplifiers
The fact that the midrange driver is not trying to reproduce low bass frequencies also helps in choosing electronics for the minimonitor speakers. Subwoofer biamping increases the satellite’s sensitivity by up to 4 times, so you can either use less amplifier power (less expense), or purchase a better amplifier at the same price point, or have more volume potential with the same amplifier. Lastly, with most separate active crossovers, you can change bass and midrange crossover frequencies and also stagger them slightly to correct bumps or valleys in the response due to room modes or speaker problems.
A Free Lunch, Then?
Not really. As is said before, every speaker design is a compromise. In this case, however, the disadvantages are not very significant, provided you are ready to experiment with subwoofer placement and do a bit of work with a sound level meter (Radio Shack’s model is a good and cheap one) and a test CD (Stereophile sells two or three different ones). A good recording helps too. My favorite is Oscar Peterson’s We Get Requests, Verve 810047, specifically the third track, You Look Good to Me. With this recording, bypassing the subwoofer should result in no apparent change in the music (upper bass below and around 100Hz).
The biggest problem is how to get the deepest and smoothest bass. Generally, the best place to put a subwoofer will be at a front wall (the wall behind the satellites) corner. Second best position is generally some place along a side wall, slightly away from the front wall corner. One very good method of knowing when you have the deepest and smoothest bass is switching the listening and subwoofer positions. Place the subwoofer on the couch where you’d normally sit to listen, while playing a continuous range of bass tones or low frequency music, and move yourself (easier than pushing a heavy box) sitting down along the front wall and thereabouts. The sound level meter will surely help a lot here.
The second problem to be solved is phasing the two audio sources. Sound waves from both the satellites and the subwoofer must arrive at your ears at the same time. When not in phase, bass tends to sound “slow” and seems to lag behind the sound of the satellites. Most quality subwoofer crossovers have a phase switch or knob. There are several methods to adjust phase. My favorite is the following:
1. Invert speaker leads on BOTH satellite speakers.
2. Play Test CD track of exact subwoofer-satellite crossover frequency.
3. Move (or switch) phase control until LEAST bass is heard.
4. Reattach satellite speakers in phase.
1. Doppler and IM distortions are produced when a speaker driver moves back and forth a relatively long distance to emit a specific low frequency sound while, at the same time, it is trying to move a short distance to produce a higher-frequency sound. Imagine a 12-inch driver reproducing 40Hz and 400Hz at the same time, or a 6-inch bass-midrange driver reproducing 60Hz and 1kHz sounds, and you have the idea. The so-called Doppler effect is the change in sound (due to a shift in frequency) emitted when a mobile object (the woofer in this case) is moving in the direction of or away from the listener. The typical example is the big change in sound of the whistle of an approaching train when it passes through.
I could edit Vladimir's remarks above, but this might remove some of their original spirit. However, I believe there are some gaps in his comments and further explanation is required.
First, the issue of biamping is only effective if the full range signal is not applied to the monitors. In most situations this is not the case. While there is normally a high pass low level (via RCA jacks) on most plate amps this high pass circuit might not be very clean. Certainly subwoofer plate amps are not built to the same high standards of a good hifi amp. Theory would seem to suggest that it would be easier to make a clean crossover at the low level, but theory doesn't always meet application.
I suspect that the current genre of standard plate amps don't have anything "special" for the high pass crossover circuit. Many folks tout the advantages of active crossovers and that they eliminate the larger active components present in the common passive crossover. These same folks conveniently omit the fact that a crossover is still necessary. The larger passive components have simply been replaced by smaller active component. Smaller components aren't necessarily better - especially when integrated circuits are used. Additionally, Ed West, a phenomenal engineer created a super duper electronic crossover simulator for the MB-1 (Murphy Blaster 1) speaker and shared it at the Washington DC DIY 2001 gathering. Ed's electronic low level crossover and Dennis's high level crossover measured virtually identical and sounded virtually identical too. The net result was no difference/improvement in sound quality using a low level electronic crossover and biamping. Ed's active crossover was monumental, but the resulted in no audible gain. Active crossovers DON'T sound better than passive crossovers. Given a plate amp, the very humble active high pass proved detrimental in my experience.
Using the subwoofer high pass also necessitates very long runs of RCA interconnects between the preamp and the power amp. Depending upon your preamp these long runs of RCA interconnect wire might be problematic. I have an Anthem CD-1 with a fairly strong output stage hooked to a Creek passive preamp and then a Bryston 3B-ST. This normally works wonderfully since the Bryston has a high 50k ohm nominal input impedance. I hooked my subwoofer plate amp between my preamp and my power amp without using the high pass crossover and the results were very disappointing. With the additional 12' of Monster Cable interconnect destroyed the response under 150hz and made everything else sound quite tizzy. Maybe with a very stout preamp and very good interconnect there would be improvements. I don't know. I do know that hooking the subwoofer between my preamp and power amp destroyed the music. This was not very audible with the paper cone MB-1 speaker, but very audible with the 1801's. I believe this is partially due to the cleaner sound of the 1801 and partially due to the deeper bass of the 1801.
There are many arguments about where in the frequency range to avoid a crossover. Most seem to agree that a midrange should cover 300-3000hz as a minimum. Above Vlad mentions a remark from Peter Montcrieff. Here I will digress.
The primary marketing premise of Peter Montcrieff's InifiCap is that his capacitors use a special end coating that attaches the lead to the foil of the capacitor. Through this method Peter is able to achieve superior signal transfer from the lead to the foil of the capacitor due to an "INFINITE" number of connections. Solen and Bennic use this method of termination too. It is nothing special, and Peter Montcrief's comments reflect grotesque marketing self-aggrandizement.
Bud Fried has a valid point concerning an area to avoid with a crossover. There are comb filtering distortion artifacts present at all crossovers between drivers regardless of slope. Given that the fundamentals of of the human voice are quite low, a lower crossover point surely would be better. There seems to be much adieu about where NOT to put a crossover. The extreme of this line of thought arrives at the full range driver. I believe multiple drivers and crossovers are a preferable compromise. I'll also admit hearing a very good speaker at CES with a dome midrange likely crossed at about 1khz. Vladimir addresses the issues of Doppler distortion above, and I believe this is quite valid. There is also the issue of motor design.
Midrange drivers are generally intended to operate down to 200-300hz. There are motor advantages to this higher frequency application. The most notable issues are motor related. Since the voice coil doesn't have an extended travel the voice coil can be designed for greater thermal power handling. Perusing midrange drivers quickly reveals that midrange drivers have better power handling than their midwoofer brethren. This is a driver design compromise between midrange and midwoofer drivers.
There are also cone size advantages. Small cones are generally limited by air displacement on the low end of their frequency response. Large cones are generally limited by flex and directional response patterns at the high end of their frequency response. Large coned drivers generally do better at lower frequencies due to their ability to move air. Getting a driver to 100hz is a full octave lower than 200hz. The theoretical impact is that a driver designed with a more restrictive lower range limited will likely be more effective above that region. For example, a midwoofer such as the SEAS W18 does quite well from 40hz to 2.8khz in the 1801, but would not be a wise choice if the crossover could be 300hz. At 300hz there are midranges that will easily operate to 3.5k and reduce the load on the tweeter. The Accuton C79 would be a better choice.
I make these comments to illustrate the many compromises present when choosing a driver for a midrange or midwoofer application. These comments aren't intended to be exhaustive, but to further illustrate the grey area in driver selection.
The above remarks might seem to indicate that I am very negative about subwoofers. Such is not the case. My comments are stated to assert the other side of the issues regarding Vladimir's remarks.
Here are some additional remarks concerning my thoughts about subwoofers and the 1801 specifically. These remarks were initially made at the Harmonic Discord site. I fear this site might be short lived, and believe these remarks are worthy of reading.
I must admit that I am not a "subwoofer guy". I am much more of a "classical
music guy". I will admit to some the theoretical understanding and experience of
the issues regarding subwoofers, but will also allow that there is a BUNCH of
information that I do NOT understand regarding this issue.