Side-stepping the Side Wall Issue

Room treatment can be a controversial subject in most homes . . . from an aesthetics viewpoint, that is. What can you think about doing without inviting an argument?

People are pretty well aware that the room can affect tonality and bass, and that there are things one can do to the room and with speaker placement and EQ to help control these things. That’s interesting stuff, and even though this article is about imaging improvement, the things one does for imaging can affect tonality, and positively so in my opinion. FWIW, they will have no effect on bass.

Imaging. Our brains take any reflections arriving sooner than about 10 ms after the direct sound, convolve it with the direct sound, and try to use the combination for determining the sound’s location.

Mixing incongruous local sound with the recorded sound detracts from the recorded image and reduces its naturalness because the conflicting timing information can't be fully resolved. It’s one of the main reasons why you can almost always tell when you're hearing a recording and not live sound, no matter the potential for reality in any particular recording.

It can be more positively thought of as superimposing a local image on the recorded one, and is certainly not objectionable to everyone. It’s a prime factor in what Harry Pearson called soundstage. (That term seems to get used in other ways, too, and there are ways of synthesizing a soundstage with signal processing.) In my opinion, because it imposes a similar sound on all recordings, it gets old. I like to hear mainly what’s on the recording.

Rear Wall. For reference, if your rear wall (behind you) is less than about 5.5 feet [1.7 m] from your seat, you’ll be getting a reflection from back there that’s early enough to affect imaging. To keep imaging as close as possible to what is on the recording, a close reflection should be at least partially absorbed, or if the room is fairly non-reverberant already, then diffused.

How? It doesn’t take thick, expensive, commercial room treatments for this. Reasonably heavy drapes or a decorative rug will generally do it. Alternatively, very simple panels of fabric-faced, 1” thick fiberglass board will make a substantial difference at low expense.

Also, the entire wall doesn’t need covering, just the area that’s maybe 20° or 30° wide when viewed from your seat at head height. For most speakers, the same angle applies both vertically and horizontally.

It’s not important to go wider for multiple seats, because small problems are not so obvious from listening positions that are off center, although it can help. As sort of a positive aspect of a small room, if the back wall is really close to you, then the area needing treatment is really small.

It’s important not to overdo the absorption, because a room without at least some reverberancy will sound dead, or at least very dull. Also, the stereo effect relies on having some local ambiance to be enjoyable. It's just first reflections coming sooner than 10 ms after the direct sound that mess it up.

If a room is already has enough absorption, the solution is to add diffusion to the reflection areas. Diffusion scatters the sound in all directions without reducing reverberancy. Commercial diffusers of various types are available from many makers, but even a bookcase or CD rack can help, as long as the CD's or books are not placed neatly, but at random depths of insertion.

Chairs. Something important to avoid is listening in a high-backed chair. Even a pillow right behind your head will cause problems, but a soft pillow behind your head on a high backed chair is better than no pillow on a high backed chair.

Side Walls. Think about the difference between how far you sit from your speakers vs. how far you sit from the side walls. If this difference is less than about 11 feet [3 m], then side wall reflections will detract from the imaging.

It is obviously not likely to be avoidable in typical rooms. In a setup that puts the speakers about 8 feet [2.5 m] from the seat, the side walls will have to be something like 11 feet [3 m] from the seat – it’s not gonna happen in many rooms.

Here are a couple of diagrams to illustrate the problem:

For side walls, the area to treat is the one where you’d put a mirror if you wanted to bounce a flashlight beam from your seat off the wall and onto the speaker. In fact, this is a good way to find that area, if you have a mirror that you can place on the side wall.

Sidestepping it. Of course, the best kind of room treatment to talk about with your spouse is probably no treatment. JansZen speakers are designed to limit the amount of image-affecting sound hitting the side walls as much as possible, so side wall treatments will make so little difference that there’s no real reason to bother with them.

As an aside, the simplest way for me to have done the design would have been with a beamy transducer, which is easy with electrostatics, but that creates a teensy problem known as head-in-vice listening.  My designs control dispersion in such a way that there is minimal change in frequency response over a pretty wide area, like a few feet side-to-side at 8 feet away. Beyond that, the higher frequencies drop off quickly. This minimizes wall splash and maximizes imaging in any room without creating an overly narrow sweet spot.

Aside from the imaging issue, it happens that tonality problems and slap echo both tend to be reduced by the same placements of absorption as will intercept first reflections and improve imaging, although they might require more. Bass modes are a different animal, although absorption added for simply reducing bass decay times will also reduce overall reverberancy and can thus help with tonality and slap echo.  
 
As one last note about improving imaging, sitting very close to any pair of speakers is a sure way of increasing the proportion of the direct vs. room sound. Desktop computer speakers tend to image well for this reason, but it’s obviously not a great setup for more than one person at a time. 

For those who are accustomed to a speed of sound figure of 1 foot per millisecond, I'd like to point out that this is an approximation. The speed of sound in air is just about 13.6” per ms at STP, which is why I cite 11' as an ideal path difference to get a 10 ms delay. In metric, the delay is very close to 3 ms per meter.

Front wall back wall: Does it matter at all?

Speaker placement near the front wall (behind the speakers) can seem to increase early reflections, and I think I know why. It starts with the visual cue we get when the front wall is well back from the speakers. This facilitates the auditory illusion of image depth, or at least does not contradict it. There’s a temptation to attribute this increase in perceived image depth to a reduction in first reflections from the front wall.

Placement out into the room, however, actually increases the strength of first reflections, and has little effect on the secondary ones. 

Except in the case of dipoles, there is no rearward sound from speakers except at bass frequencies. Since bass frequencies don’t serve any purpose in sound localization, there are no relevant first reflections from the front wall. Any reflections from that wall are at least secondary after first reflecting from the back wall, or from a side wall and then the back wall. The back wall is usually the prime culprit with directive speakers, and the side and back walls for those with wide dispersion.

In fact, moving speakers closer to the front wall will increase the distance to the back wall, and this is a change that can only serve to benefit the image by decreasing the strength and lengthening the timing of early reflections from the back and side walls. Also any additional distance that you move a speaker toward the front wall moves it away from the back wall, and this will increase the distance the reflected sound travels to your ears by double the change in distance. The effect is similar to stepping back from a mirror and seeing your image move away twice as much as you have actually stepped and getting twice as small.

To put speakers near the front wall, they must be designed to allow this, however, which mainly amounts to having an appropriate amount of low bass, namely not too much, because being near the wall will reinforce the low bass. JansZen speakers have woofer controls that are useful in this situation. Of course, they must also be a non-dipole design, which ours are.

Electrostatic Speaker Myths

Not all electrostatics make you pay for supreme clarity, naturalness, effortlessness, and neutrality

Certainly, electrostatic speakers are not the usual thing, and their reputation for being quirky has been well earned (not by JansZen speakers, though). On the other hand, lots of things that might seem universally true about them ain't necessarily so. Here's a list:

Electrostatics need special amplifiers. Full range electrostatic speakers generally do present an amplifier with such a low impedance and highly reactive load that not all amplifiers will work with them. There are a couple of examples made here in the USA. One even makes its own special amplifier for its speakers. The low impedances are the result of driving a large electrostatic speaker's inevitably large capacitance over a wide range of frequencies, and the capacitance itself can cause some amplifiers to oscillate. 

Hybrid electrostatics like ours, on the other hand, with a cone woofer handling the bass and lower midrange, has a low capacitance, and is thus easy to drive. Our zA1.1 is a 6 Ohm speaker, and our zA2.1 is a 4 Ohm speaker, and neither one's impedance drops below 3.5 Ohms at any frequency.

JansZen speakers do not need special amplifiers.  

Electrostatics need big amplifiers. This one is true to some extent, but also true for any speakers with sensitivity in the mid-80 dB range. Ours are, however, especially good at making natural sound at low volumes, and if one has a 25W/ch amplifier, one will still get up to 100 dB peak levels from a pair of zA2.1's in a medium sized room. 

Our zA1.1 and zA2.1 have 85 dB and 87 dB sensitivity, respectively, and can handle 120W continuous power. If one wants to get the absolute maximum available sound level from them, a 120W/ch amplifier will reach the maximum steady state volume, and a 240W/ch amplifier will reproduce brief transients by taking advantage of the speakers' headroom capability.

We're somewhat busted on this one, depending on what you consider a big amplifier

Electrostatics sound better with tube amplifiers. This is a matter of personal taste/opinion. Our speakers work well with any type of amplifier. Since they use cone woofers, the bass is naturally tighter with a solid state amplifier, but not everyone likes that. Whatever sonic signature or lack thereof an amplifier might have, it will be well represented with our speakers. 

You can use any type of amplifier with JansZen speakers. 

Electrostatics are unreliable. True for some others, but not ours. There are quite a few weaknesses in other designs that lead to reliability issues after a few years:

  1. Inadequate protection against the formation of unintended high voltage pathways. This can lead to clicking, popping, hissing, or squeaking sounds. Soot and dust from the air can collect due to the constant presence of static electricity. Our design prevents dust collection as much as possible, and prevents whatever debris is inevitably attracted from coming into contact with any of the high voltage parts. 
  2. Inadequate or nonexistent protection against arcing. When the stator electrodes are not insulated, or are insufficiently insulated, sparks can jump through the membrane that vibrates to make the sound, eventually causing it to tear and stop working. Our electrodes can withstand extreme voltages without allowing a spark. The stator wires will glow blue and make tons of ozone, but arcing will not occur, even when amplifiers that are much more powerful than recommended are used. 
  3. Loss or alteration of membrane tension. To work properly, the membrane that vibrates to make the sound must have the right tightness. There are several ways that it can change, such as being stretched too much when the speaker expands from high temperatures during transport, but then not recovering completely, or losing of adhesion at its edges, or a phenomenon called "creep," where a polymer gradually relaxes over time. There is a particular problem with one design, where to start out working properly, all the tension has to be applied in the vertical direction, and none in the horizontal. Over time, tension can accumulate in the horizontal direction, however, either from stress relief from high temperatures, or creep that transfers some of the tension from the vertical to the horizontal, and the result is that the membrane gradually saddles back toward the rear stator and eventually collapses onto it.
  4. Failed bias supplies. Electrostatic speakers require an electrostatic charge on the membranes that vibrate to make sound. Because our electrostatics are not handling low frequencies, they do not need very high bias voltages, so our bias supply can have a minimal number of components -- no power transformer, no oscillator, no active circuitry -- which makes it inherently reliable. 
  5. Failed signal step-up transformers. Electrostatic speakers need high voltages to vibrate the membrane. Again, because our electrostatics are not handling low frequencies, they do not need very high drive voltages, and this means that the signal transformers endure far less stress. 
  6. Loss of membrane coating. Electrostatic speakers require a microscopically thin conductive or semi-conductive coating on the membrane to distribute the bias charge. There have been designs that inherently allow or cause the coating to gradually disappear. Our coating is not only stable, but is protected by an additional membrane that is laid over it, which is unique to our knowledge. 

Janszen speakers are reliable, and carry a rather long, 5 year warranty. 

Electrostatics make weird noises on their own. True for some, no doubt, but not ours -- see above about unintended high voltage pathways. 

JansZen speakers are as quiet as any.  

Electrostatics don't make enough bass. To get good bass from a purely electrostatic speaker requires a lot of surface area, meaning the speaker has to be very big. There are two reasons for this: 

One is that to make a lot of sound at low frequencies, you've got to be moving a lot of air, but the membrane that vibrates to make the sound cannot move much, so the speaker needs a lot of area to make up for that.

The other reason is that at low frequencies, an electrostatic must be operated as a dipole, open at the back, because it doesn't generate the force needed to pressurize the air in an enclosure. (Note: Cone speakers waste most of their energy and force accelerating their massive cones, coils, coil formers, and suspensions, so pressurizing air in an enclosure is hardly any extra effort for them. Electrostatics have practically no mass of their own, and operate with inherent delicacy, using most of their energy moving air to create sound waves.)

In a dipole, since low frequencies are omnidirectional, much of the low frequency sound goes out the front and in the back, and vice versa, canceling out much of the low frequency sound.  

We take care of the problem by using cone woofers, and we do it in ways that different from the others. 

JansZen speakers produce full, deep bass.  

Electrostatics mated to cone woofers have a noticeable discontinuity. At one time, woofers were not good enough to reproduce sound as well as an electrostatic, so there was a reason to bother making huge electrostatic speakers. Although electrostatic speakers still outshine all other types for clarity, low distortion, and effortlessness, woofers have come a long way -- most are pretty good these days, and some are just excellent.

As long as the mating is implemented in a way that allows a seamless transition to the electrostatic portion of a speaker system, there's no reason to bother making a full range electrostatic speaker anymore. This seamlessness, however, is a trick that other companies don't seem interested in applying, perhaps because boomy bass is a popular thing, even if it doesn't match up with an electrostatic's speed. 

In addition to taking good care of the cone/electrostatic transition, we also have ways of getting more than the usual amount of bass from a compact, sealed enclosure. Set up properly, our zA1.1 and zA2.1 speakers will be down only 3 dB at 30 Hz in many rooms. 

There is seamless continuity between the cone woofers and the electrostatic elements in JansZen speakers.  

Electrostatics are not loud enough to sound dynamic. A pair of zA2.1 can produce 108 dB peaks. If this is not loud enough for you, then for you, this one might be true. Most people, however, find our speakers surprisingly dynamic. This is at least partly because our speakers sound just as clear, full, and revealing at low volumes as when played loudly, and do not need to be blasted to "come alive".

JansZen speakers are "dynamic."

Electrostatics are unsuitable for popular music. This one has to do with the one about not being loud enough and the other one about not having enough bass.

Simply put, JansZen speakers can rock. With external subs, or in the case of the zA2.1A-HP, without them, they can even give you a trance club experience.  

Electrostatics are hard to set up. Unlike ours, most electrostatic speakers are dipoles, with just as much sound coming out the back as the front. Unless the speakers are about 1/3 of the way out into the room, this causes interference effects as the reflected sound interacts with the direct sound. In any case, it also creates other challenging setup issues. For some, the sense of envelopment from sound bouncing all over the place is worth the added coloration and lack of imaging, but this is not how we see things. 

Our speakers are monopoles, i.e., the sound comes only from the front. Their set up is pretty much like with any other speakers. In fact, because the tweeters are planar, their directivity reduces wall, floor, and ceiling splash, which lets them go near side walls without creating excessive brightness. It also makes their imaging spectacular. 

JansZen speakers are are about the same to set up as any other speakers.  

Electrostatics are "beamy." This refers to the need to sit in the very tight sweet spot that some planars produce. The larger the speaker area, the tighter the beam, and the less latitude one has for listening position. The general solution is to use a small transducer, but that reduces loudness.

In our case, we make the area that produces the treble narrower than the area that produces the midrange. We can do this without losing loudness, because an electrostatic speaker gets louder at higher frequencies. The sound from the treble and midrange areas overlaps and combines to provide about a 20° width at 10 kHz, and then drops off quickly farther to the left or right.

This is called controlled dispersion, and creates a fairly wide sweet area, while restricting wall splash. It also means we do not have to add an intrusive electrical network to compensate for the increased loudness at higher frequencies. As with any stereo setup, there is one place where the imaging is centered and most convincing, but the tonal quality is maintained over an area that's wide enough for two or three people. 

JansZen speakers let you relax in your seat, and you can listen with others, if you want, just like other speakers. 

Electrostatics sound like headphones. This is an interesting one, and is related to the reputation for being beamy. Extremely beamy electrostatics or other planars can sound like headphones, even to the point of sounding to some extent like the sound is in one's head.

Our speakers, on the other hand, have the immediacy of headphones, but the sound is very convincingly three dimensional, out and away from you, with height, width, and depth, floating in the air. As is one of the usual goals in high end audio, the speakers themselves become very difficult to locate. 

JansZen speakers offer the advantages of headphones without the drawbacks.  

Planar magnetic (magnetostatic) speakers are electrostatic. It's easy to mistake a planar magnetic for an electrostatic. Examples of magnetostatic planars are the AMT, the isodynamic or quasi-ribbon tweeter, the ribbon tweeter, or the full ranger, such as from Magnepan or Apogee. Both types do vibrate planar membranes to make sound, but the operating principles are entirely different. 

A planar magnetic has metallic foil bonded to its relatively heavy plastic membrane to conduct the current that generates its dynamic magnetic field. This works against a static magnetic field from magnets in the speaker frame, and thus vibrates the membrane. A true ribbon has just the foil without the plastic membrane. In both cases, the metal foil is relatively heavy, and presents a mechanical load that significantly influences its motion, much like a cone speaker, although not as severely. Its prime attribute is its planar nature, which lets it operate as a true piston, unlike a cone speaker, and thus produce relatively uncolored sound. 

An electrostatic speaker applies a static charge to a very much lighter membrane than those found in magnetostatic planars. The stationary electrodes produce a dynamic electrical field that works on that static charge, thereby vibrating the membrane.

An electrostatic membrane therefore starts out very light, and does not have any mass added to it, at least, not enough to create a mechanical load. Its load is almost purely acoustical, i.e., the air itself. This lightness makes it far better able to respond to transients and also avoid continuing to vibrate after a transient has ended. 

The differences between planar magnetic and electrostatic speakers are important, and while planar magnetics can produce excellent sound, electrostatics can do it even better.  

In short, JansZen electrostatic hybrids are as friendly and trouble-free as any other kind of speaker, and are even easier to deal with in some ways. They'll give you the very clearest sound possible, with no worries about how it's being done. 

Speaker Setup -- Minimizing the Allison Effect

Allison Effect Peaks and Dips, and How to Avoid Them

The Allison effect is named after Roy Allison, a prominent speaker designer during the 20th century high fidelity craze. He is known for having figured out how some of the prominent peaks and dips in the bass range of the frequency response were caused by constructive and destructive interference of sound waves after reflecting from the floor, ceiling, and near walls.

He noticed that when a nearby surface was a quarter wavelength from the middle of a woofer, the response had a dip at the corresponding frequency. This happens because, when the round trip is a half wavelength, the reflected sound is out of phase at the woofer, and the two cancel. In reality, for several reasons, the effect is spread out in frequency, so the dip is modest, rather than a full cancellation, but still significant, generally about 3 dB. 

Along the same lines, when a nearby surface was a half wavelength from the middle of a woofer, he noticed that the response had a peak at the corresponding frequency. This happens because, when the round trip is a full wavelength, the reflected sound is in phase at the woofer, and the two add together. As with the dips, the effect is spread out, which reduces the heights of the peaks.  

The effect is double when two surfaces are about the same distance from either or both speakers. This is the usual thing, because a symmetrical setup is the usual thing, where both speakers are the same distance from the front wall, side walls, and floor. If four distances are about the same, such as when the side wall and floor distances are the same for both speakers, then it's quadrupled. When a room mode also happens to coincide with an Allison effect, then the frequency response can be quite terrible, indeed. 

The cure is to make the speaker placements asymmetrical.

First, they must be different distances from the side walls, which means setting everything somewhat to one side, rather than centering the system along a wall. The speakers must also be different distances from the front wall, which means rotating the system away from that wall somewhat. If you offset the system one way and rotate it the other way, then at least the listening position can remain more or less centered. Finally, if possible, and most challenging to our sense of what looks right, the speakers should be at different heights. It looks a bit off kilter, but sounds much more natural, and less like you're hearing speakers in a room. 

I've created a simple spreadsheet, called JansZen - Allison effect minimizer l.xlsx for MS Excel, and JansZen - Allison effect minimizer l.ods for Open Office Calc (find them here)that makes it easier to determine a good set of positions. It does not include room mode effects, which are much harder to model, or room gain, or overall boundary bass lift. Of course, if you find that a set of placement distances makes a room mode effect worse, then you can modify one of the settings to circumvent that, and go again. 

To use the spreadsheet, first enter the room dimensions and speaker separation distance [in feet] into the green cells at the left.

Then start entering speaker distances in inches into the green cells in the Distancecolumn in the table. The blue cells are figured by the spreadsheet. Each distance is from the center of the woofer. If your speakers have more than one woofer, use the middle of the span between them.

Below the table, there's a bar chart of speaker distances, and the goal is to make them all different from one another. Try for as even a slope of bars as possible. The following example has two pairs of equal bars, because the speaker heights were set equally. 

The spreadsheet does not tolerate having two distances exactly equal, so if two are the same, such as floor distance to each speaker, then add a small amount, like 0.1", to one of them. This will allow the sorting routine to function without affecting the results.  

Here's a model using the example distances from the spreadsheet. It doesn't look as out of whack as you might expect. 

Have fun, and enjoy your new freedom from room effects. 

Controlled Dispersion vs. Beaming

Off Axis Listening to Controlled Dispersion Speakers (like ours) vs. Beamy Speakers 

You might have heard talk about how the best positions for critical listening to directive speakers is fairly limited. If so, this might seem to put the kibosh on such speakers for people who want to use them for background music, or listen from another room. Perhaps this is a real concern with some speakers, but not with ours, and here's why:

The sweet spot or sweet area is indeed the place where the image is most convincing and centered. That's actually true for wide-dispersion speakers, too. But when you hear our speakers from another room, it sounds like musicians are playing live back in the room where the speakers are. There's no image, of course, but the realism is striking due to the fidelity of the initial sound and how the walls in the speaker room uniformly mix and distribute it before it exits the room.

One zA2.1 owner first came to us on the basis of the sound he heard from the hall outside our exhibit room at an AXPONA show in Chicago, when the room was packed and no seats were available. He heard them later in our demo room in Columbus before he ordered, but he arrived already impressed. 

Off axis in the same room, there’s less treble, but this can be a good thing for the usual off-axis situations, namely parties, working, cooking, etc., when one has other things that are more important to pay attention to. We have several clients who tell us they play them all day long in their offices or living areas while they work, read, or cook, although of course they do their critical listening in the sweet area. 

This brings me to the point where I distinguish our speakers as having controlled dispersion, and not being “beamy,” that is, they don't require you to sit in one exact spot to get a good frequency response. In our case, controlled dispersion creates a listening area that’s two or three people wide with reasonably uniform frequency response, at 10 kHz and below. In effect, rather than having laser beam dispersion, you could think of them as having more like floodlight dispersion, illuminating an area with broad spectrum sound, rather than a pair of ear-sized spots. . 

It's worth noting also that even well off-center, the image remains partially intact, merely shifting off center. The reason is that JansZen's are quasi-line arrays above about 500 Hz, so the mids and highs are dropping off relatively slowly with distance compared to the case with a point source. This means that, even when seated directly ahead of the left or right speaker, you still get enough sound from the other speaker to create an image, although it will be shifted toward your side.

If you decide after all that it will be very important to have wide frequency response at a party, then this is handled by turning the speakers so they face the walls, and perhaps turning down the woofer level by one click, which is possible on the zA2.1. The room boundaries will then disperse the highs evenly. 

Neutral Sound

An engineer who used to to develop mass market speakers commented recently on my favorite audio forum that it is practically impossible to sell a neutral speaker [for home use, anyway]. This is interesting to me, of course, because my designs are neutral sounding, i.e., meant to leave the recorded sound alone, as well as I can possibly manage.

My father, Arthur A. Janszen, believed that there's a cyclical variation in popular taste between neutrality and coloration, and that during swings that favor coloration, there was particular praise heaped on accentuated bass and treble (i.e., suppressed midrange). I think his belief was really just a hope that neutrality would someday come into fashion, which it seems to me never happened and never will

Privately, retailers have always had derogatory terms for the suppressed midrange sound, but it is what impressed and still impresses people in the showroom, whereas neutral speakers sound relatively dull during immediate comparisons to those designed for showroom impact. If customers got sick of the sound after a while, so much the better, because they'd be back looking for purchases that might improve it.  

40 years ago, midrange de-emphasis was described as "smooth" sounding in the press, a term that annoyed my father, as did any terms that seemed like they might mean something positive, but actually referred to departures from fidelity, deliberate or otherwise. He once took the editors of Consumer Reports to task for using the term "warm" in a review of one of his designs.
 
One of the things that I believe is proving an impediment to placing JansZen speakers with dealers is that they are neutral sounding. 

On the other hand, this makes them especially suitable for home trials, where people can have a chance to appreciate their true merits and ultimately fall in love with them. This love is especially true for women, and I've heard more than once from men whose wives of many years began sitting down to listen for the first time ever, after JansZen speakers were in place. Most others have commented that their wives like them better than any others that have been in the house. 

This is not just because they're neutral sounding, I must add, however. In addition to realistic tonal balance, they have vanishingly low distortion and no peaky midrange and treble colorations that plague all dynamic speakers and hurt women's ears. 

And if you're a woman reading this, whose husband sat down and listened with you for the first time after you bought neutral sounding speakers, I don't mean to leave you out! It's just that most audiophiles are men, and everyone who's bought our speakers so far have been men. I'd really love to see our speakers catch on among you, and wish I knew how to find you out there. If you can give me a clue, please do. 

Room Correction -- Alternative to Room Treatment and Speaker Setup?

Room correction. With that name, you might think it solves a room's acoustical problems. What actually happens, when everything goes well, is an automated equalizer, or a person wielding a manually operated equalizer, undertakes to mask the problems. I think a more accurate term is room compensation. My opinion is that room correction should be a last step after the room and speaker setup have been adjusted to require the least amount of equalization possible. 

I don't mean to contradict that one can get marked improvements in frequency response flatness from equalization of rooms or speakers, and flatness is of course a key element of natural sound. On the other hand, at least for masking room or speaker problems, EQ and especially automated EQ can have drawbacks:  

  1. Equalization treats only symptoms, not the underlying problems, and does so only partially. Altered frequency response is just one of several symptoms. Each symptom can arise from any of several possible room problems. Most importantly, a room with frequency response anomalies tends also to have time domain problems that degrade imaging, transient response, and accurate reproduction of recorded ambiance.
  2. Equalization is unable for example to treat the transient response impact of room modes, but simply puts less energy into them; the modes are still there and they still resonate.

  3. Equalization has the intended effect on room mode masking for just a single seating position, since room modes are not evenly distributed around a room. Systems that purport to compensate multiple positions do so by not compensating any single position as well as when targeting a single position. When set for a single position’s particular combination of modes, a slight movement in any direction will move the listener into a spot that will have been made worse by the EQ.

  4. Equalization tends to reduce the naturalness of the sound. This is because the effects of the room may be technically problematic, but in fact sound perfectly natural when excited by a reasonably flat speaker, but not so much when the speaker’s innate response has been altered by room EQ.
  5. Equalization potentially degrades the transient response, because it can implement sharp adjustments and thereby introduce ringing that can sound nasal, cupped hands-like, hooty, etc.. Although most automatic systems are smarter than that, it is impossible for an automated system to know the difference between every adjustment that makes physical sense and those that would be detrimental.
  6. Equalization potentially overdrives the signal chain when trying to fill holes in the room response, or adding some bass lift to compensate for pushing down individual bass peaks, and in that case reduces dynamic range.
  7. Equalization puts one at the mercy of the A/D and D/A conversions and circuitry in the equalization device. As far as I know, none are up to the latest high resolution standards (384 kHz/ 32 bit PCM or 11.2 MHz DSD). If the device includes a DAC for file playback, high resolution files that are up to these standards may be degraded; if you have an analog setup, or have a practically perfect output signal from a high grade DAC, the signal may be degraded.
  8. Some devices introduce audible hiss. 
  9. It might be awkward and complicated to operate. Some automatic units test the owner’s patience with a practically endless series of long, slow frequency sweeps for making brute force response measurements, instead of an MLS or other brief interrogation, or have a graphical interface smaller than a business card, or can be operated only by way of a remote control the size of a crispy cracker, with 32 equally sized and shaped buttons on it, no tactile feedback, and slow or unreliable system response.

That said, dynamic range issues aside, equalization of the peaks caused by room modes for a single seating position will generally sound better in that seat, as long as it is only the peaks and not also the troughs that are equalized, and the overall bass response is then lifted with a shelving filter. When rises and troughs are relatively shallow and broad enough to span a couple or more octaves, both can be equalized to beneficial effect without adding resonances that degrade transient response. Brightness caused by room reverberancy can be rolled off to beneficial effect. Peaks caused by the Allison effect can be equalized out over a pretty large area of the room, but the Allison effect can be minimized with an asymmetrical speaker setup, to be discussed in detail in another posting, thereby reducing the amount of EQ needed. 

Not to go negative on room compensation without supplying an alternative, I must point out that it's possible to reduce the effect of room modes with strategic speaker placement, among other physical things one can do.

True room correction, that is, treatment of the underlying sickness, is done using bass traps, absorbers and diffusers. You can get a good part of the way with just absorbers. Of course, doing these things right is not automatic, but you can count on a genuine improvement in the sound. They also tend to become a significant feature of the decor, which might not be desirable, but carpets, drapes, and furniture with cloth covered cushions are very helpful. I'll try to make this the topic of a future posting.

I absolutely must also mention that the simplest of EQ options, namely good old fashioned bass and treble controls, can create broad spectrum adaptations to room sizes and conditions with the greatest of ease, not to mention fix the sound of odd recordings, and they don't degrade transient response.

Unfortunately, tone controls fell victim long ago to an extreme interpretation of the high end audio simplicity doctrine. As a result, manufacturers found they had to provide at least an illusion of signal path simplicity, and must have welcomed the side benefit of reduced costs along with the perception of higher value.

There are thus many who are of the persuasion that tone controls and EQ in general are not good things, so I can't finish without mentioning that many of these same people will make exceptions for certain other forms of equalization.

One is the playback of vinyl records and tape, which are rather heavily equalized to compensate for shortcomings in their natural frequency response and dynamic range, and whose equalization must be undone with reverse equalization in the playback circuitry. Cables, amplifiers, phono cartridges, and so on also make most of whatever impact they have based on the way they alter the frequency response, which makes them very expensive and unpredictable forms of EQ. Tone controls are inexpensive, harmless, and predictable. 

Of course, there is increased adoption of the far more complex and pitfall-laden room compensation devices discussed above.

I'd be very happy to hear that someone who might benefit from room treatments has taken the above to heart and started experimenting with them. Physical room correction is quite an interesting undertaking, and does not necessarily have to comprise unattractive items that an arbiter of decor will object to. 

What's All this Arcing Stuff, Anyhow?

No sparks in my house, are there dear?

An arc is a spark that keeps on going once it gets started. You want that in a welding rig for melting metal, but it's not so nice in a speaker. At a minimum, arcs punch holes in the membranes that vibrate to produce the sound, eventually causing it to tear and stop working. 

Arcing is thought to be a general risk for an electrostatic speaker, and it is for some, even under what you might think are ordinary conditions. For at least one brand, there's a cottage industry dedicated to their consequent upkeep. Those speakers need electronic voltage limiting, which is a definite improvement over arcing, but has some drawbacks: transients are clipped off or compressed when playing at high volume, which sounds “pinched” and might cause an audible crunch as the protection circuit kicks in, arcs can still occur under some conditions, and some protection circuits can overload and damage an amplifier.

Some electrostatics use insulated electrodes. When done right, this will prevent arcing under any and all practical conditions. 

JansZen electrostatics use insulated electrodes that prevent arcing up to at least twice the voltage (four times the power) that causes ionization. One key to making this work well is that, with this amount of physical protection, the ionization that occurs with a sustained overload can occur safely, and merely makes the electrodes glow blue (which is pretty, but you can't see it through the grill unless you turn off all the lights and go right up to the speaker, but it will sound terrible and the ozone it makes will stink). Anyway, this makes the electrostatic portion of the speaker get quieter, which throws off the tonal balance and lets the owner know to turn it down. The reason it gets quieter is because ionized air is electrically conductive and thus shorts out the bias charge on the membrane. After turning down the volume, the sound will immediately go back to normal. Our speakers also have other forms of protection that will kick in if the owner endures or perhaps enjoys the muddled sound during an extended period of extreme overload. 

Here's the good part: when an overvoltage condition is brief enough, there isn't time for the air to ionize. With transients from hammer strikes, string plucks, etc., you get the full sound of that transient, even though the speaker is theoretically being vastly overpowered at those instants. This increases realism when these transients occur during periods of already loud play. As a result, we recommend amplifier power that is above what is needed to produce the maximum sustained sound level; this gives you distortion-free headroom and thus increased overall fidelity, without danger to the speakers. For the models zA1.1 and zA2.1, this is in the 250W/ch/8 Ohm range. 

ESL Mythbuster – A Double Dose of Double Do

Dipoles create the best ambience, and electrostatics are always dipoles?

What's a dipole? A dipole speaker puts about the same sound out from both the front and back. More specifically, the sound from the back, called the backwave, is 180° out of phase with the front wave. A bipole does the same, except the backwave is in phase.The backwave adds what is known as “air” in audiophile terms, a sense that there is sound all around the listener. This is generally an enjoyable thing, but not all air is the same.

Getting some air. There are two basic ways to get air. The usual way, as with a dipole, is to increase the amount of sound that is reflected around the room, to the point that it is a quite significant portion of the total sound that reaches the ears. I call this the added sound approach.

All added-sound ways of getting air have drawbacks, and dipoles have a couple of their own. Beginning with dipoles specifically, one drawback is that to get the intended results, the speaker must be placed far from the wall behind it, at least six feet (two meters) away, and ideally about a third of the way into the room, which takes up a lot of space. Another is that the backwave interferes with the front wave, and also interferes with its own reflection from the the wall behind it. This creates peaks and dips in the frequency response called comb-filtering that alters the recorded sound. Alteration like this that adds or subtracts sound without distorting it is called coloration, like adding tint to white paint.

Added-sound in general causes a smearing or blurring of the stereo image. This is because the precise timing of the sounds from the speakers is no longer distinct after getting mixed with the sound reflecting from the walls, floor and ceiling. Also, using the listening room's reverberancy to add air mixes this reverberancy with the recorded ambience, and the two are practically never the same. This confuses the brain's perception of the recorded ambience, making it seem less natural and real, although of course many find it still enjoyable.

Dipoles everywhere. I don't know of any non-JansZen electrostatics being made today that are not dipoles.There is even a contingent of dipole cone speaker makers and DIY'ers. Dipoles tend to look cool, they add air, and for electrostatics, they are relatively inexpensive to build. The partial exception is QUAD, whose speakers have absorbent material in back that removes a good bit of the treble from the back wave. Dipole operation at bass frequencies can reduce room mode excitation, but has great difficulty producing deep bass, and like all dipoles, relies on placement well out into the room. 

JansZen electrostatics, however, have never been dipoles, with the exception of only a couple of briefly made models in the1980's that had quasi-dipole tweeters (lens in front, partial absorber in rear), and of course the venerable KLH Nine of the 1960's, which was necessarily a dipole because of its electrostatic woofers.

Why not dipoles? Because the “air” is already in the recording, and it's the real thing. The recording engineers' art includes faithfully recording the ambience of the performance space, or in the case of compiled studio recordings and electronica, creating the most authentic possible ambience by various means. Any attempt to add air in the playback space is bound to throw this off.

The trick is to faithfully reproduce the recorded ambience, including all the faint and subtle phase cues, while minimizing the contribution of the room's own acoustics. This can be done most naturally with electrostatics, about which JansZen staff are the world's foremost experts.

Engineering vs. Engi-hearing

I'd say that when it comes to speaker design, pure engineering is a great place to start and a disastrous place to finish, and conversely, a lack of engineering is a disastrous place to start, from which there is no way out. 

Some designers express a personal belief with no engineering basis, creating speaker enclosures for instance that are meant to be instruments in themselves and resonate like violas, and at the far opposite extreme, others seek for example a uniform power response, both ultimately at the cost of what I call natural sound. Floyd Toole, a prominent acoustical device researcher and engineer, falls somewhere in between, for me an interesting case, wise enough to trust his and his subjects' ears, yet unable to resist the urge to pick a number for an idealized roll-off specification, although taken with a grain of salt, his 1 dB/octave might be a nice suggestion for where to start. 

What do I mean by natural sound? I remember once, REG made the point on his forum that a good musical performance or a good stereo recording cannot be defined, i.e., like enlightenment, if you can say what it is, then that is not what it is. It's also a matter of taste to some extent. Likewise, I can't define natural sound, except to say it causes a reaction in me and those whom I observe that is comparable to experiencing live music.

I can wax on about what one of my designs does technically, and have done, but truth be told, there could be speakers that do about the same things that I have included in my own designs, but might not sound natural [to me]. My point is that when it comes to some designs, the actual sonic outcomes can be hard to infer from visual observation or from reading the designers' writings or specs, so if possible, give them all a listen, an extended one if that can be arranged, before making a decision.

As you do this, bear in mind that some types of sound reproduction will sound odd because they are. Some that are interesting or exciting at first will grow tiresome, eventually, by always sounding the same, always changing what was recorded in the same way. Some might sound odd at first, but grow on you as you get used to hearing something very close to what was recorded, each recording that you thought you were familiar with opening up new experiences for you -- odd at first because this is not the usual thing. Of course, I believe my designs are in this final category. 

Welcome to the Blog

Welcome to the first installment of the “JansZen Blog.”  If you love music and realize or suspect that excellent sound can increase your enjoyment of it, then this is for you. It’s not so much a high-end audio or audiophile blog, and it's not a commercial for JansZen. You might even think of this as audio for poets.

Read More