by Richard J. Kaufman
Really, I had no intention of buying a new speaker system. I only sat down one Sunday afternoon to design the perfect subwoofer with a combination passive/active crossover network and a unique method of bass equalization.
After a few hours the design was ready, leaving me with just one problem: I now need a speaker system with very specific characteristics, such as second-order bass roll-off, and high damping. My present speakers just won't do. I find myself haunting stereo sound rooms on weekends, but I'm not satisfied with what I've heard so far.
YOU'VE GOT TO HAVE FRIENDS
I work with a friend I'll call the Engineer. When I tell him I'm looking for new speakers, the Engineer introduces me to George, a leading salesman at a downtown electronics emporium with high-end aspirations.
I tell George I want to listen to the Snell KIIs and compare them to another speaker that was the most impressive of the dozen or so I'd heard uptown. (I mention the brand name only because much of what follows is so specific to this particular speaker. This is not meant as an endorsement of the Snells--though they are, indeed, fine speakers. Bear in mind my needs were very specific, if not peculiar. You will not be constrained as I was. By all means listen to Snells, but listen to other speakers as well before you buy speakers.) At the mention of the Snells, George says "Good! Let's escape from mid-fi hell," and he leads the Engineer and me to the first sound room towards the back of the store. (There is a series of sound rooms, off a hall leading back through the block-long store, each with more expensive equipment than the last. Someday I'll visit the last room and attain audio nirvana, but it won't be today.) The Snells blow away the speaker that had been my first choice. The Engineer concurs. Being a man who considers it unethical to make a sale with out working at least a little, George suggests we listen to a new British import at a slightly higher price. The price range is already more than what I'd intended, but what the heck? I wait as George fiddles with the sound levels till the speakers are matched, so we can hear the real differences between them. Then we listen. Both speakers sound good. I find the British speakers seem "sweeter," more musical . .. they also have a peak in the mid-bass due to a Thiele-Small alignment with a Q of 0.9 or so.
The Snells sound as if they have a Q of 0.7, for a B2 alignment. (Later measurements confirm this.) This makes the Snells' bass sound tighter and more natural, and the low midrange less chesty and resonant. Also, bells and cymbals have a treble edge that is utterly captivating. I find the Snells are more analytic, but also, I think, more accurate.
The Engineer finds the British speakers more analytic and the Snells more musical. He is surprised I prefer the Snells, though they would have been his choice. He used to have a job in a CD factory checking the final product before it was "pressed," so I think that makes his opinion professional. I am pleased we agree, though I would have been happier if our subjective reasoning were the same.
I tell George I have a few more speakers I want to audition but promise to call him. That weekend, I listen to some more speakers in midtown. Then I call up George and make an appointment to listen with him again.
THE ENTHUSIAST
This time I bring my friend the Tube Enthusiast. The Enthusiast does not own a CD player, and electro static speakers and tube amps are his preferred equipment. He can also pinpoint even the most minute defects and distortion in an otherwise functional system. He's the one who first remarks that the channels are re versed or that the system seems to be playing in mono. He's the one who notices that a tweeter isn't functioning while other audiophiles are commenting on how sweet and smooth the sound is.
We listen. My unspoken decision to purchase the Snells is confirmed by what I hear. But the Enthusiast is not enthusiastic. "It's as if reality is here," he says, pointing straight ahead, then gesturing from side to side, "and one speaker is to the right, the other to the left, of what you want." I ask him to be more specific. "The British speakers are tubbier, and the Snells' tweeter has too much zing." I know the Enthusiast's horn-loaded electrostatics roll off sharply as they approach 12 kHz, so I politely attribute his criticism to an unfamiliarity with true treble material, and announce my intention to purchase the Snells. George offers to let me audition both pairs of speakers in my own home--he likes the British boxes-but I refuse: I know I want the Snells.
"I'm sure you'll be happy with them," says the Enthusiast. "And if you decide they have too much treble, you can always put a penny over the tweeter." He does know how to put the worm of doubt in my ear.
Figure 1---Adding a foam ring about your tweeter can greatly improve stereo
imaging and the smoothness of the treble response.
REARRANGING AND REDECORATING
It is a week later. Only twice have I rearranged the furniture in my office/listening room. Now I can sit comfortably at the proper distance, mid way between the speakers. I am quite pleased with the sound on most material, but sometimes there is too much treble. I am not imagining it.
The treble seems at once too bright, yet deadly accurate. With measuring equipment and laboratory-quality microphones, I find that the speakers are flat as far as half an octave beyond the highest frequency I can hear, but they still sound too bright. How am I to ex plain this paradox?
My listening room is too live, no doubt. I get a bigger rug from another room. It helps the room, but there is no effect on the perceived brightness of the tweeters. I idly contemplate placing acoustic foam on all vertical surfaces. My wife has been very tolerant about rearranging furniture, but I know she'll draw the line at covering the glass on the French doors with textured foam. I am not really very eager to do this anyway.
I have an inspiration. Instead of treating the room, I'll treat the speakers. After a trip to the hardware store to purchase some foam weather-strip ping, I remove the grilles and make foam circles about the tweeters. The idea is to block the dispersion of treble sound, limiting it by means of the foam barrier to a cone of sound 45° off-axis to the speakers (see Fig. 1).
It works! The unpleasant edge to the treble is gone! I sit back to enjoy my handiwork, and make a surprising discovery. The imaging is uncanny. Each instrument and voice takes on an exact physical and solid presence at some point between the speakers. The image is extremely stable and never wavers. With some material the sound even seems to come from the right of the right speaker or to the left of the left. Such imaging is not uncommon with electrostatics, but only twice be fore have I heard it so pronounced with dynamic loudspeakers. I get a pair of ancient Advents (the original model) I'd been using with my TV, hook them up to my system, then treat the tweeters with the foam tape. Their imaging is improved. Not as good as the Snells', but much more defined than before. I note the same effect with a pair of the original EPI 100s my wife had when we got married. What have I discovered? I spend the next few weeks modifying friends' speakers. In every case they prefer their speakers with the foam treatment. Where the image is very well defined to begin with, there is less improvement because there's less room to improve. Even so, no one has removed the foam rings. Hundred-dollar boxes or $4,000 towers--it doesn't matter. Everyone agrees the sound is better. I return to the stereo store, and George lets me put foam rings on his demo speakers. We listen. "I like it" he says, then calls in his manager and makes him guess what's going on.
They must like the sound, because the foam rings are still on the demo model when I stop by some weeks later.
I have brief fantasies of going into manufacturing, but the device is not patentable. Without a functioning distribution network, imitators would flood the market before a profit could be made. Nor would a manufacturer want to pay royalties or buy such a simple idea. Yet I can still share my discovery with the audio community. A magazine article would help finance my continued experiments.
There is one last test before sharing my discovery with the world. I call up the Critic. The Critic makes his living from sound, reviewing and passing judgment on music and equipment. If he believes in the tweeter rings, they are ready for the world.
THE CRITIC
I've gotten an invitation to visit the Critic in his home, despite having revealed that I intend to modify his speakers. I've come with my wife because we have a dinner party that evening in the neighborhood, and as a former musician she's interested in my audio activities. "I'm just afraid that after listening to your system, he'll go out and buy a couple of hundred dollars worth of records," she says to our host. "I'm afraid you'll just have to watch him carefully," the Critic replies, his eyes twinkling.
His listening room is deceptive. It seems small and narrow, but it's not.
Figure 2--Adding a felt blanket to your speaker system's baffle eliminates
edge diffraction effects and improves imaging.
It's a converted sun porch that runs the length of the house. It's nearly 11 feet wide, and about three times that in length. What makes the space acoustically difficult is the glass windows that run the length of one long wall. They reflect every sound, creating the same overly bright effect I had been struggling with. On the positive side, the length of the room helps sustain an illusion of spatial depth to the sound reproduction that I have seldom heard on other systems.
"What would you like to hear?" the Critic asks. The problem will more likely be that he has misplaced and can't find our selection, than that he doesn't have it. The walls are lined with CDs, many with Japanese and German labels. Low shelves under the windows hold hundreds of vinyl records. We listen and listen: Vivaldi and Bach; Rolling Stones; Cajun Zydeco, and religious choral works. Somewhere in all this music I remember to get technical and pry the grilles off his speakers. I carefully place and position the foam around the tweeter, forming a ring.
We continue to listen. The Critic is showing off rare and unusual polycarbonate (that's what CDs are made of) he has collected in his travels. We try a few vinyl records, too, then switch back to CDs. Reluctantly, for I am enjoying just listening, I steer the conversation back to the sound quality: Specifically what improvement the foam tweeter rings have made. "I think it's better," the Critic demurs. "Have you tried felt?" "Felt?" I wonder.
"Felt" says the Critic. "Cover the baffle with carpet felt. It will eliminate much of the baffle acoustically. Write about that, too, and I think you've got an article." We listen some more, and talk about food and houses and children and running and knee injuries, and promise to get together sooner rather than later, then say goodbye.
On the way to the dinner party my wife insists on stopping at a mall, where she buys a couple of hundred dollars worth of CDs. She is looking for something the Critic played for us, a European issue of some Viennese choral group, not likely available in this country. She looks anyway, and keeps finding other recordings she wants. I help her and surreptitiously add a few boxes to the pile she has already put in my arms.
FELT
The next weekend I ask the man at the local carpet store for a yard of scrap carpet felt. He finds a left over piece. I offer to pay for it, but he gives it to me when I tell him it's for a magazine article on improving stereo equipment. I do have to give him advice on buying equipment. He already knows what he wants, and I reassure him he is making the right choice.
There is enough felt for several speakers. This is a good thing, be cause it is not obvious to me how to fit the felt to the speakers or how to attach it or cut speaker holes of the right size in the right places: A false start or two seems likely. The material is about % inch thick with an embossed pattern on one side and loosely pressed-together fibers on the other. It shifts shape slightly as it is handled, so it isn't easy to work with. Through careful measuring I somehow get the holes in the right place on the first try. I've had the fore sight to cut the sides slightly large and the holes slightly small so I can trim for an exact fit. (See Fig. 2.)
The next problem is attaching the felt to the speaker baffle. I double up heavy-duty masking tape, stick it to the baffle, and then .press the felt on. Before I can listen, the felt begins to sag. I don't want to use glue because the result would be irrevocable. I notice the speaker grilles are made from an elastic cloth stretched over a wooden frame. Perhaps if the felt were trimmed to fit within the frame, pressure from the cloth fabric would hold the felt in place? It works. I trim the felt for the other speaker, then sit down to listen.
I expect the results to be exactly the same as with the foam rings-but no.
The precise imaging is there, as are the clear but not harsh highs. To de scribe the effect in terms of perceived sensation rather than by measurements, it seems the sound is slightly smoother, not so harsh in the upper midrange. Perhaps calling it a liquid clarity, rather than a dry clarity, conveys the sensation. In any case, the improvement is subtle but obvious, offers everything the foam rings did, plus a little more. (The felt is still on my speakers. It sagged after several weeks, covering the tweeter and destroying the highs. I 'affixed it to the grille frame with two pieces of Mylar strapping tape at the top. Hanging in this fashion, it has held its shape ever since.)
WHY IT WORKS
A search of literature turned up several articles in the Journal of the Audio Engineering Society concerning imaging in speakers, as well as distortion caused by cabinet shape and interference between adjacent drivers.
For clear stereo imaging, the frequency response of a pair of speakers must be precisely the same, differing by no more than half a decibel. Unless a speaker has properly mitered or rounded corners and edges, there will be a waver in the frequency response curve. Instead of a straight line, the response becomes a line of peaks and valleys. The location of these peaks and valleys varies with listening angle, and this is sufficient to interfere with a speaker pair's imaging capabilities.
Another source of response irregularities is the cavity of the woofer's cone.
A second requirement for precise imaging is that a speaker's off-axis response match its on-axis response.
The Snells I purchased have an 8-inch woofer and a 1-inch dome tweeter. The crossover is at 2.7 kHz and is very sharp. The immutable Laws of Physics dictate that an 8-inch speaker with a 2.7 kHz input signal, where the wave length of sound is less than two inches, will beam its sound forward, whereas a 1-inch tweeter will diffuse its sound in all directions. Clearly, the on- and off-axis response of this or any other two way speaker system will not be the same in the vicinity of the crossover point. Three-way systems should have an easier time of it, if their crossover points are chosen so the lower range speaker isn't beaming at the crossover frequency.
This is why, I think, the foam rings create such precise imaging. They limit the tweeter's dispersion in the cross over range to more nearly match that of the woofer. The rings also eliminate reflections from the side walls of the room, reducing the perceived brightness of the sound, without affecting the direct-arrival sound, so the perceived frequency response is still accurate.
The felt probably does a better job of attenuating sound at lower (but still high) frequencies than does the foam.
By eliminating edge-diffraction effects at these frequencies, it gives a further (if slight) improvement, while still pro viding the benefits of the foam rings.
MAKING YOUR OWN
If you wish to experiment with your speakers, I would suggest starting with foam. I used 3/8-inch thick Foam King weatherstripping tape. Other brands, materials, and thicknesses should also work. Best results seemed to be had when a line between the top inside edge of the foam and the base of the tweeter formed a 45° angle to the front panel of the speaker system (Fig. 1). With my speakers, which use a Vifa tweeter, the internal diameter of the foam ring was just slightly shy of 3 inches. No great precision seems necessary, and considerable variation still gives good results. The felt carpet padding is about the same thickness (it's hard to mea sure a material that deforms as you handle it), so you can make the tweeter holes the same size as the interior of the tweeter rings, if you want to continue to experiment. Another option, which I have not as yet tried, would be to cover the edges of the speaker cabinet with a very thick adhesive foam material, say 1 1/2-inch thick. Such thick weatherstripping, with an adhesive backing, meant for sealing air conditioners in windows, can be found in some hardware stores. It should be easier than using felt.
I don't know of a simple and effective, yet reversible, method for attaching felt, unless your speakers have a frame and stretched grille-cloth system that happens to hold the felt in place.
Hot-melt glue comes to mind as a possibility, but I would fear damage to the drivers if one were to slip during the application process. Perhaps a system of fasteners such as Velcro would work, if enough fasteners were used to support the weight of the felt.
WHY DON'T MANUFACTURERS DO IT?
Some manufacturers have used rounded or beveled edges on their cabinets to eliminate diffraction effects. This greatly increases their fixed costs. The speaker becomes priced out of its market category, since for the same expenditure a larger cabinet can be made or better quality drivers can be used. The edge-rounded cabinet is hard to sell on the sound room floor. People still go for 10-inch woofers over 8 inchers, and zingy treble over accurate sound. Foam rings are inexpensive enough to avoid this price problem. Some manufacturers have used them, though none I know use my recommended configuration to control lateral dispersion as well as to eliminate edge diffraction. Speakers equipped this way have enjoyed no great particular success; perhaps the look of plastic foam turns buyers away.
I also don't know of a currently produced commercial speaker that uses a felt blanket on the baffle. These materials may be looked on as too shoddy, for people make audio purchasing decisions with other senses than their ears. Marketing is everything. I would be pleased if readers' experiments' with foam and felt eventually led manufacturers to include them in their products, for it would be a definite improvement in sound for everyone.
References
Vanderkooy, John, "A Simple Theory of Cabinet Edge Diffraction," Journal of the Audio Engineering Society, December 1991.
Kates, James M., "Loudspeakers Cabinet Reflection Effects," Journal of the Audio Engineering Society, May 1979, reprinted in Loudspeakers, Vol. 2.
(adapted from Audio magazine, Sept. 1992)
Also see:
Perception and Geometry [by Richard C. Heyser; June 1977]
Listening and Experience (Aug. 1992)
= = = =