Audio Speaker FAQ

What should I listen to when evaluating speakers?

The most important thing is to listen to recordings that you *know*. Any good salesman will play you recordings that highlight strengths of a particular speaker. Do not be embarrassed about bringing a stack of CDs with you to the hi-fi shop.

Do not spend valuable listening time switching brapidly amongst dozen pairs every 3 seconds. If you are shopping at a quality store, the dealer will, from the description of your room, your size requirements, your musical tastes, and your budget, be able to show you a couple of speakers that should be close to what you want. Spend several minutes listening to each. When you think you're close, don't be embarrassed about spending half an hour or more listening to the speakers. You're going to have them in your home for a lot longer, and some speakers will cause "listening fatigue" after a short time. Make sure you really like them before you hand over money.

One thing to try is well recorded "Spoken Word" records; most people have a very good ability to tell when a speaking voice sounds unnatural, even if they've never heard the person speaking live. But beware, there are some people with very unusual voices (consider Tom Waits, for instance). If you play an acoustic instrument, find something that features that instrument solo, or in a small group; make sure it really sounds like it should. Because almost everyone has heard a live piano, it's a possibility, but remember that pianos are difficult to record well. But a good piano recording can be very revealing.

Blues, jazz, folk, or 'easy listening' music with simple instruments and a female vocalist is also usually a good test. Many well recorded female singing voices provide a very good test of a system's response. Try something simple and soft, which will let you hear any extraneous noises coming from the system; and something complex, with lots of instruments all happening at once, to make sure the system doesn't become 'muddy' when things get complicated. And, of course, try a few of your favorites, and see if you like what happens with them.

If a sales person suggests some music to listen to, it may be that it favors a particular speaker. Sales folk tend to suggest things which sound great. Anything you own and like is good, because you know it and are happy to listen to it carefully. No matter how good the recording, if you don't like Opera, you won't listen to it as carefully as your favorite, scratchy, 1940's rhythm and blues.

Most important is to listen to something you are familiar with. Even if a recording is flawed (and what ones aren't?), how is it different from your normal setup? Some of the most important differences are "Gee, I never heard that instrument before!"

What should I listen for when evaluating speakers?

When comparing two speakers side-by-side, during an AB comparison, be extremely careful to match the levels before evaluating. A slight level difference can make one speaker sound better, even though the difference may not be perceived as a level difference. Some claim that you will be influenced by a difference of less than 1/2 dB!

First and foremost, the sound should impress you as natural. If you listen to vocals, close your eyes and try to picture someone singing in the same room with you. Does it sound realistic? Likewise with instruments. You selected recordings of instruments that you like and have heard live. Do they sound like what you remember them sounding like live?

Your very first impression should be something like "what nice sound". If your initial gut reaction is "gosh, what a lot of detail", the system is likely to have emphasized treble (even if not be much as the effect can be subtle) which is often interpreted by beginners as "more detailed". You'll probably find that annoying after a while. If your first reaction is "hey, what powerful bass!", then the system is probably bass-heavy; which is not ideal for most listeners. The most common mistake for beginners is to buy a system with REALLY powerful bass, because it sounds "impressive" at first. After a while, though, you'll get tired of being thumped on the head by your music. Acoustic guitar shouldn't have a flavor of bass drums to it.

Note that this is not to say that good bass and treble aren't important. But your first sense of a speaker which might be reasonable for you should be that the music is 'all there', and that it comes together as good music, without one particular part trying to dominate the whole. Sit back and listen to it for a bit. You should be able to pick out the individual instruments if you want. They shouldn't force themselves on you, and you should also be able to hear the music as a single piece, the sum of its parts, without feeling like each of the instruments is trying to grab your attention away from the others.

You should check how things sound with the amplifier turned up, and also with it turned down to a fairly low volume level. Some speakers which sound very nice at low levels begin to sound confused, like they can't cope, when turned up. On the other hand, some sound nice loud, but sound thin and bodiless when you turn them down a bit. With the spoken word or female vocalist, listen for "sibilance", the 'sss' at the end of 's' and 'z' sounds. It shouldn't be there, for recording engineers try hard to avoid it by microphone placement, pop filters, etc. Some speaker types characteristically can't play very loud; their operating principle prevents it. Whatever you hear, do some auditioning at the maximum volume you anticipate ever wanting.

It is acceptable and sometimes desirable to switch the amplifier to mono to evaluate naturalness. If speakers are mounted and connected sensibly, this can be a good test of both the room and the speakers. The centered image should be rock-solid dead center, and not move with different instruments or voices or with different loudnesses. If it isn't perfect mono, it will be nearly impossible to create a good stereo.

Generally, a speaker in a large box is capable of producing low frequencies at higher volumes with more efficiency than a small box, but that doesn't mean that a good small box can't have great bass, it just can't be as efficient nor play as loud without distress.

Good speakers can "recreate a natural stereo sound stage", placing some instruments to the left of the left speaker, some sounds in the middle, and some to the right of the right speaker. Poorer speakers fail at one or more of these tasks.

Why use a subwoofer? Will it help? One or two?

One reason to consider a subwoofer is to add bass to an otherwise acceptable system which is bass deficient. A second, more technical but as important, is to move the lowest frequencies to a separate driver, and thereby reduce a kind of distortion caused by overworking the woofers in the existing system; it's called nonlinear mixing of different frequencies, or "intermodulation distortion". A third is to increase the power handling ability of the system and its overall reliability. All are valid reasons, but simply adding a subwoofer isn't actaully simple.

To improve the sound of a good speaker system, a subwoofer must "integrate smoothly" into the system, extending the low frequency performance (ie, bass) without causing peaks or dips. Many subwoofers have a crossover filter that goes between your amp and your main speakers; it sends the lows to the subwoofer and the rest (higher frequencies) to the existing speakers. This may damage the perfect sound of a good system, it may sound similar, or it may sound better. And there is more than one way this can be done -- do it wrong and you may damage electronics or speaker drivers. Read the manual...!

Many otherwise good small speaker systems have a bass peak at resonance, which attempts to compensate for the absence of lower bass. Like it or not, this is a common way to make a small system sound acceptable. not everyone listens mostly to clavichords and flutes. If the small system is done well, the improvement you will get from a subwoofer will not sound dramatic, but still be real and, to many, significant. For home theater systems, which are set up to handle low frequency sound effects, subwoofers are necessary in most cases. If you're putting together a home theater system, you will have to as well since even very good high fidelity speakers won't be satisfactory on modern sound effects tracks.

Correctly designed, and sensibly tied into the main speaker systems, a good subwoofer will enhance the sound of a good small-box system. Done wrong or haphazardly, anything is possible. Even a fine large speaker system might benefit from careful addition of a subwoofer. However, the better the original system, the more likely it will be that a subwoofer will do more harm than good, except for those sound effects.

Low frequencies are less directional than high frequencies, so some people say that only one subwoofer is required for good sound. This is true to some extent, but not completely. For home theater systems, the sound effects channel (the '.1' in the theater electronic systems) needs only one subwoofer, of course. Depending on the arrangements made in the electronics (mostly this means the decoder circuitry) you may or may not be able to use the subwoofer for ordinary music through your home theater system.

There are a few reasons for getting two subwoofers. Some feel that you need two subwoofers to accurately reproduce the stereo image, no matter how little low-frequency stereo information there is. Certainly if there is any low frequency information in a channel, you may want to relieve the main speaker system on that side from handling it. Others feel that two subwoofers are much easier to set up in a room, less likely to excite standing waves in the room (a problem in most rooms), and give smoother sound.

A third reason is that two subwoofers can produce twice the low frequncy energy (more or less this is loudness, though at low frequencies our ears don't handle loud in quite the same way they do such things as voices or most instruments). Finally, even though subwoofers produce very low frequency sound and very low frequency sound provides few directional clues, many subwoofers also have output at 100 Hz. Sound at 100 Hz is weakly directional, so two subwoofers will give a slightly better stereo image than one. Assuming, of course, that the two are sensibly placed.

Finally, though most source material rarely contain any music with stereo components below 50Hz, there may be some noise component with low-frequency, out-of-phase, noise. Pipe organ 'chuff', for instance. This unusual noise might add a sense of space to a recording if it is reproduced by a system in which the woofers are very far apart.

nevertheless, for ordinary music, it is still true that a single good subwoofer, correctly added to a system will help the sound but two will probably help more.

How do you connect a subwoofer to a stereo?

Many subwoofers contain their own amplifier and a crossover filter before it. For these, a low-level preamp output is usually needed; feed it into the subwoofer amplifier's input and also into the main amplifier.

For other subwoofers, which are set up for a speaker level input, just run them in parallel with your main speakers, or combine them into your system with your own bass amplifier and crossover. However, NEVER connect speaker level signals to an input (anywhere on any equipment) which requires a low level (also called a 'line level') signal. You will cause damage, instantly when you turn the amplifier providing the speaker level amplifier on.

Some A/V receivers have an electronic splitter specifically intended for use with subwoofers. If you have one of these, you will either want a separate amplifier for your subwoofer or an amplified subwoofer.

Always consult the manual which comes with the subwoofer and the amplifier(s) you are using. Proceed with caution.

What do I need for surround sound?

In recent years, "Surround Sound" has meant many -- mostly incompatible == things, most embodied in one or more products. Many current mass-fi receivers have "Surround Sound" buttons that do little more than change the imaging in some way.

More recently, "Surround Sound" has become synonymous with the surround decoder systems produced by Dolby Laboratories, DTS, SRS, and others. Dolby Surround is the best known and most widely used; unfortunately for many of us, it comes in several flavors. One is a kind of 'passive surround' (which simply decodes the 'phase' information in a recording and routes it to rear speakers); years a go this technique was often called Hafler decoding. The more complex Dolby system is called Pro Logic. The Pro Logic system uses computer circuitry to route directional information to the appropriate speakers, sually five or more.

Generally, one needs at least two more speakers beyond the main stereo pair. Advanced Pro Logic systems such as those from Lexicon and Fosgate can accommodate several more speakers beyond the two additional ones (usually placed in the rear). Often one can find Pro Logic systems with two front, two rear, two side, as well as a center channel speaker for dialogue. Plus a subwoofer signal for movie sound effects.

I was just approached (accosted?) by a couple of kids driving a van that said they had some GREAT speakers to sell. Am I getting ripped off?

Almost certainly, yes. The speakers these people sell are almost never what they describe. They are not used in studios. There might be one or two DJs out there that use them because they can't afford anything else, or got fooled. They are essentially never overstocks, and probably NOT HOT!.

Are they good speakers? They're no better than the big boom boxes you find in $400 rack systems in department stores. They are worth no more than what the kids paid for them ($100/pr). Usually, the kids are sold the speakers by a distributor with a truck full of them and told to go out and sell sell sell.

The speakers often have names like "Acoustic Monitor DB IV", "Acoustic Linear," "Pro-Poly," "Audio Reference 4350", "Omni Audio", and so on. They all "feature" things like "liquid cooled 3" tweeter", poly-cone 12" woofer, fantastic (but effectively impossible) frequency response, 98 db/watt sensitivity (extremely unlikely), and so on. The brand names are often quite similar to reputable firms you may have heard of, but different enough to delay law suits. Remind you of fake handbags and jewelry?

These speaker are made by a couple of manufacturers with the intent of being sold exactly this way. They cost the kids in the van about US $100 a pair, and the kids are given minimal (or no) training about what kinds of stories to use, what parking lots are the most likely to generate sales (department store parking lots near colleges in September...).

Anything over the US $100 the kids paid is over priced. Adn there's no guarantee, nor can you get replacement parts for almost any of the speakers the kids have for sale.

What speakers should I consider in the $XXX/pair price range?

This is probably the most commonly asked question on rec.audio, and also the most difficult to answer. The market keeps changing with new models and discontinued old models, everyone has different tastes, and no one has the time to listen to even 10% of the products available in any country so reviewers are of limited help. Also, many products (good or bad) are only available in specific regions or countries.

If you really want recommendations and are willing to listen to the opinions of others, check the past few issues of one of the high fidelity magazines. Stereophile is one, and there are several others. Although they are strongly biased towards expensive gear and have assorted other biases they do apporve of some very good equipment in their frequently-updated list of "RECOMMENDED COMPONENTS". Other choices are Web based, many of which are international. TNT Audio is based in Italy, has English language versions of many reviews and includes lower cost equipment than most of Sterophile's. Audio Ideas Guide is similar, but based in Canada. A Google search for speaker reviews will turn up more; there are several. Note that some are very opinionated, some are unreliable, some are poorly informed, some are very good.

Can you build better speakers than you can buy?

Some people can build better speakers han you can buy. They are either experts, golden ears, extremely well equipped, inspired, or some combination of the above. Most drivers used in commercial speaker designs are also available as raw drivers from their manufacturers (several in Scandinavia, for some reason). Only very large speaker system vendors can order enough drivers to permit a proprietary design at any reasonable cost.

Some companies have plans available to entice you into buying their drivers: examples include Audio Concepts, Audax, Dynaudio, Focal, KEF, Lowther, and Scanspeak. Several audio fan Web sites have produced and built designs (for some reason horn speakers are particularly popular). Other sources of plans are from individuals: MurphyBlaster (an economist and violinist who designs, builds, and modifies speakers for fun), Humble Homemade Hifi from Holland (Tony Gee has had a blast building speakers of most every type and size), and Keith Kidder has several designs, including two or three subwoofers, on his site (he also includes some thoughts on acoustical treatments for rooms, some of his design and test equipment and links to other designer sites). Your success rate with plans will probably be very good IF your cabinetry skills are also very good and IF you follow the plans precisely. If you deviate, anything is possible.

Stereophile has published three different plans designed by Dick Olsher which are similar two-way ported systems. A recent one of these was in Stereophile Nov '90, pages 94-127. Audio Magazine published a plan called "The Pitts" by Ken Kantor, in Audio, Nov '88 pages 65-71 continued in Dec '88 pages 73-77. This plan is a two-way sealed box. Wirelss World (now Electronics World) published several of Dr Bailey's pioneering designs for transmission line speakers in the mid-80s.

I have built one published design and one manufacturer's design. I believe that both met my expectations. They took me a long time to build, taught me a lot, were fun projects, and sounded good when finished.

I also believe that a commercial system which cost what my parts cost will never sound anywhere near as good as the one I build. If you consider $2/hour for my time, however, building the speakers was financial suicide.

Designing your own system can be even more a can-of-worms, and for those without the skills and commitment, should be left to those with either a strong stomach, a very forgiving ear, infinite resources, excellent guidance (available from a some nearly fans, most likely -- consider the Boston Audio Society, for instance).

How can I improve the sound of my speakers?

The easiest way to change the sound of your speakers is to change where you put them. Ideally, stereo speakers (and the front two speakers in a home theater set up) should be located at ear level, in front of the listening position, and equidistant from that position. It's then a matter of fiddling with:

• the angles (some speakers are 'beamy' at high frequencies on axis (toe out or in a bit) others should be directly facing the listening position)
• the distance apart (you ears and brain require some separation, but not too much, so they can integrate the information coming from both speakers)
• the distance from the listening position (too far away and all you can hear is what is called the reverberant field -- not recommended)
• the distance from the wall (this affects low frequencies).

Just moving the speakers around in the room or putting them onto stands can make a major difference.

The other modification which is easy to do is to arrange your listening room to match your speakers, your ears, and your preferred music. What you actually hear is a mixture of direct sound (coming straight from the speakers) and reflected sound (having bounced off a wall, the ceiling, or the floor). Your ears expect certain relations between these; it's a matter of psychoacoustics which is only beginning to be adequaltely understood. So changing the amount and location of absorbtive materials (wall hangings, drapes, ...) will change the reflected portion of what you hear and may improve the sound dramatically. It is, in any given room, a matter of experimentation and consideration and the best solution will not usually be found quickly.

Other than that, actual speaker modifications can be a can of worms, or can produce very subtle changes, which you might prefer. For example, you might improve a speaker by carefully and tightly adding some cross braces from left to right, top to bottom, and front to back. Wooden battens (perhaps 1" x !' or slightly larger) may be enough. This will stiffen the cabinet and reduce speaker cabinet wall vibrations, which hurt sound quality in most cases. Alas, this will be most effective with lower-cost and poorly built speakers.

Along similar lines, some claim success putting lead sheet, mass loaded bitumen/asphalt sheets (think sound deadening treatment for auto body noises) or epoxy putty on too thin panels in the speaker to damp out resonances. You can also try doing something similar to the thinner parts of the speaker "basket" or frame.

Still another "tweak" is to add (or remove) sound deadening material (felt pads on the inside walls of the speaker) or absorbent fiber material to the interior volume (Bonded Acetate Fiber or fiberglass (itchy!) or long fiber wool (with insect repellent added)). Instead of felt pads some advocate sand-filled latex coatings on the inside walls of speakers. Others advocate ceramic tiles held in place with "thinset".

Still others rave about commercial products like AC Glop, Acoustic Magic, and Bostik Sheet. However, the people who rave about these products are often the same people who sell them.

Any change along the lines of adding felt, cross-bracing, or putty will have subtle effects on the sound.

For the brave at heart, you can replace old or cheap drivers with better ones, but the results of this one change can be very dissatisfying if you happen to get the wrong type of driver for that application, and may never sound right. And there will likely be changes required in the crossover to match the new driver. Speaker system design is still somewhat of a mixture between science and art.

Whatever change you try, don't "burn your bridge" home. Be sure that you can undo whatever change you did -- just in case. Many tweaks to good speakers, no matter how well thought through, will correct for one flaw, but create others, or correct a flaw that the designer had cleverly used to his advantage.

How can I replace/re-cone my old speakers?

The best chance of success is to buy an identical replacement driver from the manufacturer of the system, and install it exactly the way the original was. No new air leaks, no reversed wiring...

Second choice is to buy the exact same driver from a distributor or the driver manufacturer. This is sometimes difficult because it is hard to learn exactly what driver the system manufacturer used. In addition, EVEN IF the manufacturer used stock speakers, they might have used matched pairs or selected speakers by hand for an exact set of specific characteristics.

There are companies that rebuild drivers, but they charge quite a bit. I have heard $75 per driver. This is rarely done for anything but very expensive commercial drivers. Speaker manufacturers will sometimes sell owners the materials they need to repair a speaker. If you are handy with delicate things, it is worth a try.

In addition to speaker manufacturers, there are companies which sells rebuild kits for approximately $30 per pair, containing new foam, a special glue, and instructions. If you have a blown or distorted voice coil, this still won't help. A few netters have used rebuild kits from this company successfully. Note, however, that the surroudn material is part of the driver function and substituting a foam surround for a rubber one (or vice versa) is not likely to be satisfactory.

Contact:

Stepp Audio Technologies
PO Box 1088
Flat Rock NC 38731 USA
800-747-3692

Two other vendors of speaker repair parts are:

Parts Express (sells 8", 10", 12", & 15" repair kits)
340 E First St
Dayton OH 45402-1257 USA
513-222-0173

Simply Speakers
P. O. Box 22673
St. Petersburg FL 33742 USA
800-767-4041 or 813-571-1245

Also check out: http://www.decware.com/surround.htm and http://www.les.safety.net/stepp.html for directions on replacing speaker foam.

Some speaker manufacturers have very good warranties. Electro-Voice (US) warranties all its professional products for life. KEF (UK) has a similarly broad warranty on their speakers. Contact the manufacturer first.

What computer programs can I use to design speakers?

There are many useful programs available, but none are complete without a good knowledge of speaker design. Further, you will NEED to supplement any program with hand tweaking for the best sound. Finally, no simulation program is never useful without accurate driver parameters, and the parameters which manufacturers provide are often imperfect, so many good designers strongly recommend your own lab measurements. This will require a quality microphone (ideally calibrated). The Loudspeaker Design Cookbook (see 12.9) explains how to measure a speaker, and also gives enough theory to feel confident with a good program. You can get a lot done with a simple spreadsheet and the equations in such books.

All that said, there are several programs available for various aspects of the speaker design problem. One Web site with multiple o(most open source) programs is the Free Design Software Consortium.

Can I magnetically shield my speakers for use near a TV?

You probably will need to buy speakers that are made with an integral magnetic shield. Magnetic shielding is usually done by either shielding the speaker magnet or by cancellation of the magnetic field very close to the magnet, or both. Shielded speakers are NOT built by lining the enclosure with metal. While it sounds like a good idea, it doesn't work.

A common magnet shield is a mild steel cup around the magnet. This is the cheapest shield, and is usually fairly ineffective. It also will interfere with the speaker's critical magnet parameters, so this type of shield can hurt speaker performance by affecting the magnetic field and reducing the magnetic flux density in the gap. This will likley reduce efficiency and affect the speaker's low frequency performance.

Cancellation is done using a reverse-polarized magnet glued to the back of the main magnet. If done correctly, it can almost completely cancel the rear stray field. In some cases it can also increase the magnetic flux density in the gap, which may or may not be change the driver's performance helpfully.

What are fluid-filled (fluid-cooled, ferro-fluid) tweeters?

These tweeters are built almost exactly the same as other tweeters. They look and act almost exactly the same, too. The only difference is that they have a small, controlled amount of a special fluid inserted into the gap between the magnet and the voice coil.

One big effect of adding this fluid to a tweeter (or to any speaker) is that it makes the voice coil capable of hadling more power because the heat has an additional way to dissipate instead of melting or warping something. This means that the speaker can have a lighter voice coil, for better performance, or a higher power rating for the same voice coil. The other big effect of this fluid is to add mechanical damping. The frequency response and transient response of the driver will change, possibly for the better.

In addition, this fluid may help center the voice coil, may lubricate the voice coil, and may help keep dirt out of the gap. This fluid will not increase the magnetic field, concentrate the magnetic field or otherwise change the magnetic circuit. But it will not cushion impact if the voice coil bottoms.

The fluid used for this purpose is often called "ferrofluid". It consists of sub-microscopic particles of magnetic material suspended in special oil (geenrally a silicone type). The particles are small enough that Brownonian motion keeps them evenly suspended. This fluid stays in the gap because of the strong magnetic pull of the magnet. There is some debate over whether these fluids can dry out with time. Manufacturers claim that the oil used is non-volatile.

It is possible to use ferrofluids in mid-range drivers and woofers. However, as tweeters tend to have the most fragile voice coils, tweeters have the most to gain from ferrofluid. There are various different fluids on the market, some of which have characteristics tailored to tweeters, some to woofers, etc.

It is very risky to blindly add fluid to a driver. It may not be compatible with the adhesives used in the driver, may not be practical with the particular driver structure, and is impossible to remove. Permanent driver damage is possible.

Should I use spikes under my speakers? Pennies under the spikes?

Spikes prevent speakers from rocking. They also couple the speaker directly to the floor. Spikes will pierce carpet. Some spikes will damage carpet. Most will just put a small hole in the carpet which is invisible. Putting a heavy speaker directly on carpet will cause a permanent mark on the carpet. Spikes can prevent this.

If you have a pretty hardwood floor, then spikes will definitely damage the finish. A rigid disc under the spike will distribute the load and lessen the damage. Any coin should work fine. Using a coin will not change the speaker/floor interaction. Do not use a coin with a carpeted floor. Alternatives to spikes for wood floors are Blu-Tack and similar products. (see 12.19)

If your floor is extremely rigid, then the spikes will make the speaker more rigid. If the floor is more conventional, such as a suspended floor or a wooden floor over joists, spikes can have a positive or negative effect, depending on the resonant characteristics of the floor/speaker system.

The counterforce resulting from a forward cone motion in a speaker may try to move the speaker backwards, but spikes will have little or no effect on this. Most audible effects from spikes are due to changing the coupling of the speaker to the floor, ideally so it will be less likely to pass vibrations to the floor or to increase resonace on its stand. Some argue that in most cases, spikes will have no audible effect at all. Try it for yourself. Very expensive spikes rarely, if ever, do anything more than homemade or less expensive spikes. There are no magic spikes.

How do you couple speakers to speaker stands?

Ideally, your speakers should sit flat on the speaker stand or floor. They shouldn't see-saw back and forth if nudged.

One good way to accomplish this is to use a small dab of putty under each corner of the speaker. There are a few common putties used for this, but all share the properties of being very elastic and staying flexible indefinitely. These putties are inexpensive, removable, and reusable.

A bad way to do this is to drill holes in the speaker cabinet, or screw into them. The reason is that thee may be air leakswhich will affect the performance of the speaker or structural damage (many speaker cabinets are made from MDF, which is not suited to most fasteners like screws).

Try either Blu-Tak, which is available in the UK from office supply stores for cleaning typewriter elements, Faber Castell UHU Hold-It, which is available in the US from office supply stores for holding up pictures, DAP's Fun-Tak, which is sold in hardware stores for holding up pictures, or Pritt Buddies.

What is a Sealed, Ported, Bass Reflex, Acoustic Suspension, Bandpass, and Coupled Cavity Speaker? Which is better?

All are "direct radiator" enclosures, so called because the sound produced directly from the driver (the "radiator") is emitted toward the listener without the assistance of a contrivance such as a horn.

• SEALED BOX:

The simplest direct-radiator system. The rear of the driver sees a sealed enclosure, and none of the rear output of the driver contributes to the sound output (except for panel resonance and rear sound reflected back through the cone; these should be treated to minimize them). Depending upon the size of the cabinet, stiffness of the mechanical suspension vs how 'stiffness' of the air in the enclosure (and that's a function of the size of the box and its stuffing), you can have either the equivalent of an Infinite Baffle enclosure, in which the mechanical suspension is the dominant source of system stiffness (the box will be large; or an Acoustic Suspension enclosures, in which the air in the box is the dominating stiffness (the box will be small).

Sealed boxes tend to be the lowest efficiency systems for a given box size and bass cutoff frequency, and the smallest.

• VENTED ENCLOSURES:

Also the same as Bass Reflex, Ported, or, essentially, Passive Radiator. Here, an aperture in the box provides a way for the rear output of the cone to contribute to the total output of the system. However, it only contributes over a narrow range of frequencies. In fact, in a properly designed system, the front output of the cone is reduced at the same time the output of port increases, so the port DOES NOT ADD to the output of the woofer, it REPLACES the output of the woofer at these frequencies. This, if done properly, can significantly reduce distortion, increase power handling at very low frequencies, a region that can be difficult for drivers. Proper design of a bass reflex system is harder than some other enclosure types as the driver must match the enclosure closely

Vented systems can be up to 3 dB more efficient than a sealed box system that has the same bass cutoff frequency and size.

• BANDPASS:

These are compound systems in that they have at least two enclosures: one on the front and one on the rear of the driver. The enclosure on the front, which looks remarkably like a vented box (because it is), acts as a low pass filter, and can couple the output of the woofer more efficiently to the outside. They have several useful advantages. For example, the front enclosure can be used as a very effective acoustic crossover, filtering out mechanical noises generated by the woofer, something no electronic crossover can do. For very low frequencies, such an acoustic crossover can be far less expensive and more easily designed than an equivalent electronic crossover. These designs are tricky and it is very easy to get wrong. In addition, they are typically effective only over a narrow band of frequencies and so are best suited to subwoofers which need cover only 2 octaves in some cases.

They are called "bandpass" because the combination of the rear enclosure and the driver form the high pass portion while the front enclosure forms the low pass section. Making the bandwidth of the system narrower raises the efficiency of the system.

• COUPLED CAVITY:

A variation of bandpass and vented systems, they are the results of a designers attempt to solve specific problems. They consist of two or more rear enclosures, each coupled to the next by a vent. Each enclosure/vent combination is another resonant system, and the combination is essentially, a high order, multi-tuned resonant system.

Generally, these systems have quite complex response and are difficult to design. No comprehensive theory on their operation exists like that for sealed, vented and bandpass systems.

What is the best material to make speaker boxes out of? Why?

An ideal speaker cabinet material would be very stiff and probably dense, so that it would not tend to move with variations in box air pressure. It would also be very well damped, so that if it ever does deflect from air pressure, it will come back to the original position without resonating. It would also have a very high resonant frequency (supersonic), so that low frequency box air pressure would not cause it to resonate. An attractive material is preferred, and additional credit is given for a material which is easy to cut, glue, and finish. A great material would be cheap, too. Finally, it would be nice if the material were light, because we all have to move our speakers sometimes, and it's hard to appreciate good speakers with a sore back.

With all of those attributes, many conflicting, it's clear that there is no best material. However, there are many materials that have enough of the above attributes to make excellent speaker cabinets. Yet each has advantages and disadvantages.

In the list of good speaker box materials below, letters are used to indicate which attributes the material possesses.

S = Stiff
D = Damped
H = High Resonance
A = Attractive
M = Machinable
C = Cheap
L = Light

• MEDIUM DENSITY FIBERBOARD (MDF):

SDMC This is the most practical material for quality speakers. It is harder to find than plywood, but most lumber yards can special order it. It cuts very nicely and has a smooth surface. It takes veneer very well. However, bring a helper when you pick the stuff up. One sheet is very heavy. MDF is harder on tools than common wood, but easier than particle board. This is the material that many great speaker makers use. US $45 for a 4'x8'x1" sheet. Density: 50 lbs/cu ft.

• POLYCARBONATE (LEXAN):

DM A clear or solid-color polycarbonate box can look strikingly good. However, this is not a cheap material. To locate it, look in the classified directory under PLASTICS. US $400 for a 4'x8'x0.5" sheet. Density: 75 lbs/cu ft. Acrylic (Plexiglass) is cheaper than Polycarbonate, but weaker andmore poorly damped (not recommended).

• MARBLE:

SDHA One challenge with marble speaker enclosures is cutting holes for the drivers. A carbide bit on a router will work, but it will dull quickly. Marble is also difficult to glue, so bracing is difficult. But it sure is pretty when you're done! US $25 to $45 per 1'x1'x1.25". Density: 160 lbs/cu ft.

• PLYWOOD SHEETS SPACED AND FILLED WITH SAND OR LEAD SHOT:

SDAMC If you have time on your hands and want a great impractical box, try this. Make a simple box out of common plywood. Then glue cleats on the outside of the box to space the outside plywood from the common plywood. Glue hardwood-veneered plywood to the cleats and pour sand or lead shot into the spaces between the cleats. It won't be light, but with the filler, it will be extremely well damped. In addition, if you use strong cleats and glue well, the box will be extremely stiff. One person used different size Sonotubes as an alternative to plywood, and filled the space between them with sand. Be sure to sterilize the sand in your oven before putting it in the box. Wharfedale famously made speaker cabinets this way for some years.

• ALUMINUM SHEETS SPACED AND FILLED WITH ALUMINUM HONEYCOMB (Aerolam):

SDHL Airplanes use this material for flooring. Next time a plane crashes in your neighborhood, see if you can get the wreckage for your next speaker project. You can't get a better, light-weight material. Celestion has exploited this for some great products. If you're really ambitious, you can make your own sandwich out of high-quality plywood faces and a thick honeycomb core. You will probably need an epoxy to glue the honeycomb to the plywood. A home-brew sandwich will probably be easier to cut and glue than Aerolam.

• FORMED CONCRETE:

SDHC There are tricks to working concrete, such as to cast braces, rebar, and steel-wire right into the mix. Also, some concrete is better damped than other. Remember to oil your concrete forms so that they can be removed. Most concrete speakers use an MDF front panel, but you can pour one if you use cardboard tubes or plywood rings to mold the concrete into the shape of a speaker cutout. Alternately, you can make a common veneered plywood speaker box and cast concrete inside it for stiffening.

Any box can be improved by making the walls thicker, by bracing the walls, and by stiffening the walls. The stiffness of a material goes up as the cube of the thickness, so a slightly thicker material is much stiffer. A thicker panel will also have a higher resonant frequency because the stiffness goes up faster than the mass.

Consider lining the inside of your speaker with ceramic tile, attached with thinset mortar. You can get tile remnants cheaply. They are easy to apply and can be added as an afterthought to an imperfect box. However, be sure to attach all braces before tiling, because it is hard to attach anything to tile.

Also consider bracing any weak parts of the box. For example, all joints will benefit from a wooden cleat. The back of the box will benefit from stiffeners where the speaker terminals are attached. Most importantly, brace the front panel, or make it out of a double thickness of material.

What size fuse or circuit breaker should I put in my speaker to protect it from damage?

Most modern speakers consist of a box containing more two or more drivers interconnected through a network of inductors, capacitors, and resistors. One fuse or circuit breaker in series with that array can't possible protect all drivers.

Conventional circuit breakers are a very bad choice for speaker protection. They add series resistance, series inductance, and lousy electrical contacts, all tending to degrade performance. Moreover, breakers have a trip characteristic that does not match the damage mechanisms of speakers.

Fuses are a better choice, but still are not very good. This is because speakers have complex thermal behavior. Loud playing will warm up the voice coil making it more sensitive to damage. No fuse takes this into account correctly. A fuse will do a better job of protecting tweeters, but is still not perfect.

If you want to protect a speaker with a fuse, use the lowest current, fast-blow fuse which will not blow during normal listening. This may trip prematurely in a very loud passage, or may degrade sound quality, but it is your best bet for fuse protection. For a woofer, start with a 1 Amp fuse and work up. For a tweeter, start with 100mA and work up.

There are also cheap tweeter protectors available which contain a light bulb and a resistor potted in a small tube. They work pretty well, and if you reduce the tweeter network's series resistance by a few tenths of an ohm, they are not terrible for the sound. But they are audible and not failsafe.

Why are speakers labeled + and - or Red and Black?

Speakers make sound my pushing and pulling at the air with the motion of their cones or diaphrams. When a positive voltage is applied to the red or "+" terminal on a standard speaker, it causes the cone to move outwards and push air.

If you have two speakers side by side and one cone moves out while the other moves in, air will move between the two speakers but not much sound will escape. The two cone motions will cancel each other. So when you have two speakers close together, it is vital that they be wired "in phase", with positive voltage going to the "+" terminal of both speakers at once. You can do this by wiring the speakers in parallel or series. In almost all cases, parallel is preferred. If wiring speakers in parallel, the "+" output should go to both "+" terminals and the "-" output should go to both "-" terminals. If wiring speakers in series, the "+" output should go to one "+" terminal. The other terminal ("-") should go to the second speaker "+" terminal. The other terminal ("-") of the second speaker should go to the "-" output. See the FAQ section on amplifiers for more on series and parallel connections.

There is one exception to the + to + rule. A crossover filter will have lumped reactance, and that reactance may be such that the tweeter should be wired out of phase with the woofer when used with that crossover. This matters a great deal with such crossovers.

Even if speakers are not side by side, it is good to wire them in phase. For very low frequencies, speakers systems 15 feet apart are effectively 'close together' and the same cancellation effects mentioned above apply. For higher frequencies, the effects are more subtle but still important. One symptom of wiring speaker systems out of phase is that the stereo effect is absent or erratic. Instead of a main sound seeming to come from the center, the sound of the lead vocalist, for example, may seem to come from outside the room. Other odd effects are also possible.

So when in doubt about entire loudspeaker system boxes, always wire "+" to "+". The situation is less fixed within the enclosure amongst the drivers and crossover.

What is the best "stuff" to fill a speaker cabinet with?

The following discussion will focus on practical facts on speaker cabinet stuffing and on sealed systems. Theory is limited help in selecting speaker stuffing. Vented system do share a few of these same issues and will also be mentioned, but the goals and physics of stuffing a vented box are different than those of a sealed box.

NHT speakers use polyester fill. Some use a Danish polyester that mimics the properties of fiberglas very closely. Excluding this special poly, there are two kinds of polyester available: pillow stuffing, and audio-spec polyester.

Forget common pillow fill. It's cheap and easy to get. If you use enough, it will damp the midrange, and that's a lot better than an empty box but it has little effect on lower frequencies. One responsible study of thsi indicates that for stuffing mounted in the middle of the enclosure, smaller fibers are better at absorbing low frequencies than thicker fibers. Since low frequencies are most important in absorption schemes, this gives something of a criterion of choice. How quite to meaure fiber thickness is not so obvious. "Mountain Mist Polyester Fiberfill" from Stearns Technical Textiles is a common, inexpensive material that is said to perform as well as audio-spec polyester. Stearns also sells "Fiberloft Premium Grade Polyester" to some speaker makers. Mountain Mist is a coarser fiber than Fiberloft, but both are the same composition. We have no information on differences in acoustic properties between Fiberloft and Mountain Mist, but Fiberloft makes softer pillows and costs more. Both are available from these chain cloth stores:

• Cloth World
• Hancock Fabrics
• House of Fabrics
• Jo Ann Fabrics
• Minnesota Fabrics

For more information, contact:

Stearns Technical Textiles
100 Williams Street
Cincinnati OH 45215
513-948-5252 or 800-345-7150
http://www.palaver.com/mountainmist/
E-mail: stearns@fuse.net

For lining the walls of a vented enclosure to reduce internal reflections, or filling a transmission line to absorb the back wave, highly absorptive wool or fiberglas are ideal. However, these materials may not provide the desired results in a sealed system. They will provide more reflection absorption than polyester, but the latter is quite good in this regard in the critical midrange. In a sealed system, you want damping and isothermal conversion. (Author's note: I have tried "all-out" efforts using fiberglas lining and polyester fill to achieve the best of both worlds. I found little practical benefit over polyester alone.)

Most professional designers agree that practical experience, combined with trial and error is the best way to get optimum stuffing material, quantity, and method for a given design. This is why good designers routinely experiment with fill in the development of a new system. If you are designing a system that differs substantially in shape or volume or source impedance (passive crossover) from one of known reference, you will need to experiment to get best performance. this is generally a nuisance, as drivers must be removed and replaced, stuffing removed and replaced through the driver holes, ...

Adjusting the filling is the last step in getting bass right, and is used mostly to fine-tune the system Qtc and resonance. As increasing amounts of polyester are added to a sealed box, the resonance and Q gradually go down. This can be shown mathematically to be due in roughly equal parts to the effects of simple resistive damping and isothermal conversion. At some point, a minimum is reached, and further material reverses the trend by taking up volume. An experienced designer can find the optimum amount of fill in a few trials by monitoring the impedance versus frequency curve as stuffing is added or removed. you probably will require a much larger program of experimentation.

Filling at the middle of the enclosure also has the important effect of reducing internal reflections, to reduce standing waves and comb filtering. Filling at the surface of interior panels is by definition at a motionless node for any standing wave in the cabinet and so will little effect. However, the amount of filling has comparatively little effect on this.