REGISTER NEW USERLOST PASSWORD? WELCOME, Logout

Servo bass

Paul's Posts — 23 December 2011

By

Servo bass

I’ve written a lot about bass and its importance in your system in these posts.  It’s a subject I’ll probably continue writing about – so important it is to the music.

As my readers know I am always surprised when I find a system without benefit of a subwoofer.  There’s perhaps one or two loudspeaker systems in the world that don’t benefit from a sub and my guess is most people don’t own those systems.

So if most great systems have subs (and they should) then why is it there aren’t many great subwoofers available?  I haven’t been studying subwoofers of late but to my knowledge there are very few truly high-end subs out there.  High-end to me, by the way, means they are either very well EQ’d or better still, servo operated.  What is a servo?  It’s feedback on the woofer.

The problem with all loudspeakers is they have relatively high distortion and are not flat – woofers being the biggest culprit of them all.  Measurement folk turn a blind eye to real performance standards when it comes to loudspeakers and proclaim 5% distortion and flatness of better than 3dB as magnificent.  Imagine an amplifier with those specs.

A servo solves both flatness and distortion failings to a pretty impressive degree – at least an order of magnitude when designed correctly.  To add a servo to a woofer, one must attach what’s called an accelerometer to the woofer and then use that device as an active part of the system making real time corrections based on a comparison with the input signal.  Accelerometers range from expensive to very cheap and all have about the same performance.  Devices we used when I was designing such things were less than a dollar.

Does adding a servo increase the complexity of designing an active woofer?

Heck yes, but since when did hard have any bearing on doing what’s right?

email Servo bass Forward to a friend and help us engage more readers

Get new and fresh stories like this each morning by joining the folks reading Paul's Posts. Click here

Related Articles

Young turks Get that mole! Treat the cause or the symptoms? Jelly rolls

Share

About Author

Paul McGowan

Paul McGowan is the CEO and co-founder of PS Audio Inc. a Boulder Colorado design and manufacturing company of high-end audio products and services. McGowan has been designing and building high-end products for nearly 40 years. Hobbies include skiing, music, hiking, artisan bread baking, kick boxing and cooking. He lives in Boulder Colorado with his wife Terri and his 4 sons.

(26) Readers Comments

  1. DanielM

    December 23, 2011 at 3:01 am

    Hi Paul,

    Maybe I am a little bit off topic, but regarding bass issue, what is your opinion about Gradient Revolution Active speakers and their philosophy with open baffle bass section below 200 Hz and active crossover?

    Daniel

    • December 23, 2011 at 8:25 am

      I haven’t listened to that system but generally open baffle bass is a double edge sword – one that usually comes out on the losing end if not done properly. Because there’s no box the woofer is very quick and works well but of course you get major cancellation modes having both the back and the front of the woofer exposed. Bob Carver did this in his Amazing loudspeaker but he EQ’d the system to compensate – sort of.

      A servo accomplishes the same quickness without all the problems.

  2. lokidog67

    December 23, 2011 at 3:32 am

    Paul, I recently added a REL R-328 to my system. I don’t think it has a servo. However, it blends well with my Joseph Audio 23XLs. My listening experience has been raised to another level. Besides improved bass,
    higher frequency details from the Josephs are more audible. The sound stage is deeper and wider. I’m no engineer, but I do have fairly good ears for an old musician. There may be measurable distortion, but the musical presentation sounds better to me with this “non-high end” sub added.

    • December 23, 2011 at 8:26 am

      REL makes one of the best subs. No, they don’t servo it but they do EQ it which is good as well.

  3. demeter

    December 23, 2011 at 4:15 am

    Hi Paul,
    I know very well what you are writing about: I (still) own the Genesis 200 speakers with the servo bass woofers, which are, to my knowledge, designed by you.
    Those loudspeakers are not easy to set up in the room, but if done rightly, you have such a sound(stage), that you cannot thing to replace them with other loudspeaker systems anyway.
    So, thanks a lot to you!

  4. Soundminded

    December 23, 2011 at 7:31 am

    Servo control/negtive feedback is one of the most powerful tools available to an electrical engineer. No other technique comes close in many important regards. Every modern industrial control system relies on it. The difference between servo control and open loop control is the difference between steering a car with your eyes open and with your eyes closed. However it is also one of the most difficult and complex techniques to master and execute successfully. There are about as many equations in a text about servo control as there are in a calculus text. Small wonder many amplifier designers badly execute negative feedback designs and throw their hands up claiming the whole idea isn’t any good. Incorrectly applied this powerful tool can wreak havoc, even create a spontaneous oscillator.

    When properly applied, servo systems increase bandwidth, flatten frequency response, reduce nonlinear distortion. They work by driving a system not by the input signal but by the difference between the input signal and a portion of the output signal, in other words it is a self correction tool. Automatic room correction where a sound system adjusts its frequency response through equalization based on the output of a pink noise generator and the input of a microphone can be thought of as a one time servo system. Most servo systems operate continuously.

    Some of the problems of executing a servo woofer include finding linear accelerometers, properly mounting the accelerometer, two series connected differentiating or integrating circuits circuits, their combined non linearities affecting the performance of the servo loop, and proper gain and FR of the reverse path circuit. The criteria for stability of the overall gain is that gain must be less than one when the phase angle reaches minus 180 degrees or the system will oscillate. One pioneer of servo woofers is Velodyne. From their name I presume they use a velocimeter rather than an accelerometer which requires only one stage of signal integration or differentiation. Another was Infinity’s models 1 and 1a Servo-Static subwoofer/satellite system. I don’t think their later IRS systems used servo control.

    Servo systems applied to subwoofers skillfully will correct speaker cone motion such that the motion of the woofer is a closer analog of the input signal than it would be otherwise. It can not correct for room resonances, cabinet resonances including ports, poor speaker placement, limitations of power handling or maximum cone displacement and cone breakup. It will not make a 4″ woofer sound like an 18″ woofer. There are practical limits to its application. It is best used on high end systems where other means to improve performance have reached their point of diminishing returns but performace is still judged inadequate. For these reasons servo control systems are rare and usually expensive.

    Servo woofers optimize peformance according to Newton’s second law of motion as applied to forced oscillation. This equation is presented in every college physics text, mechanics text, and probably most calculus texts as it is the quintessential second order ordinary differential equation, possibly one reason Newton developed calculus in the first place. As with all measures to improve performance, the engineer must ask himself if there are cheaper better ways to achieve the same result? The best applications of acoustic suspension design suggest that there are, the parameters of Newtons equation being optimizable for any desired FR. (Thiel Small is a cookbook shorthand for applying this equation.) That optimized acoustic suspension design is isn’t more commonly used speaks more about the limited knowledge of speaker designers than it does about which is the the most cost effective design approach to achieve a desired performance goal.

    • December 23, 2011 at 8:27 am

      Yup. I spent nearly two years designing the servo systems and it’s a lot of work and calculating but when done properly the results are killer.

  5. Rob Hughes

    December 23, 2011 at 8:15 am

    Hi Paul,

    I think it would be helpful to qualify your statement that “there are only 2 or 3 speakers systems that don’t require a subwoofer”. The lowest note on a piano has a fundamental of ~27 Hz, so any speaker that reaches that deep is quite fully able to reproduce nearly all music. About the only place such a system would need help would be with pipe organ stuff. I personally run a set of vintage Polk SRS 2.3 TLs, and listen to a wide range of music, including electronic. The low bass notes shake the room, and I have no sub, nor is there a need for one. Even in my HT, my mains will reach to 18 hz, but of course, there’s a sub there.

    So to the main point of my question/statement – are you referring to home theater, or 2 channel reproduction? While I don’t disagree that many systems would benefit from a sub (or better, 2), I think blanket statements should be avoided.

    That said, most of the great systems I’ve heard, and I’m speaking strictly of 2 channel reproduction here, do not have subs, nor do they need them. Of course, they also feature monolith-sized mains.

    Really, it comes down to the room, personal tastes, and the type of music being listened to. But where I would agree is that every HT system does need at least 1 sub, and probably more.

    • December 23, 2011 at 8:32 am

      I am referring to two channels and I stick by my blanket statement. The problem is most speaker systems can’t reproduce 20Hz notes (or even 27Hz) at your listening chair. Most manufacturers measure loudspeaker bass with a closely placed microphone (1 meter) and while that’s interesting it’s not meaningful.

      The problem is two fold: having enough woofer area to couple the air in the room and placement.

      Without enough square inches of woofer area you cannot couple the air – which is a reactive load. The next big problem is that even if you have enough woofer area to couple the air in the room, your loudspeaker probably isn’t in the right place in the room to get the bass to your listening position. Separate subs that can be placed in the optimum position for bass is pretty much a necessity in almost every situation.

      • Gordon Meltzer

        December 24, 2011 at 12:08 am

        My Wilson Maxx 3′s have big well designed woofers and they’re spec’d at 3db down at 20Hz. Shoiuld be all set, right? And the bass is stunning……15 feet from the listening position. In the (otherwise) sweet spot, there is NO BASS and subs are needed.

        • December 24, 2011 at 11:38 am

          I would think those Wilson’s have plenty of bass – some of the few speakers that would.

  6. Frank LaFond

    December 23, 2011 at 8:18 am

    The average room contributes considerably more to the lack of bass fidelity than does the average subwoofer. Flatness better than 3dB for a small room certainly would be magnificent. Many years ago Nelson Pass designed the Phantom Acoustics Shadow, a kind of servo for the entire room. Sophisticated active room correction addresses both room and subwoofer issues. So while I certainly agree that a servo has the potential for great improvement in a sub, feedback mechanisms that include the room as well as the sub hold more promise in my mind. We’ve talked about it before, but some of the high-end Velodyne subs feature both servos and room correction, which should really be something!

  7. Dan Schwartz

    December 23, 2011 at 8:40 am

    I don’t think you ever had a chance to hear my Snell Type B prototypes. They were dubbed “the Refrigerators” by Kevin V. I used them in 2 different rooms – still own them but have loaned them out foe the last 16 years – and they hit 20 Hz, easily, from a pair of 10s. 16 was also pretty easy for them. No servo, but seamless. The integration across the bandwidth was nice, using a simple crossover. They just take a lot of space. The production B was quite a bit smaller, but still used the forward and rear-firing woofer to help with room modes. REG did an extensive piece on them back then (the first production pair), and when he was done with them we took them over to Bill Bottrell’s studio and he still owns them.

  8. emfoods

    December 23, 2011 at 9:16 am

    Do the servo units wear out? Since these are doing much more than a conventional speaker, are the servo subs going to slow down or get less efficient over time (like an electric motor)? I’ve got a wonderful Velodyne 15″ Servo Sub (ULD 15 II) and am debating rebuilding the amp. Should the sub get rebuilt as well?

  9. hahax

    December 23, 2011 at 9:36 am

    I’ve always thought there were 2 critical factors in a woofer that goes into the last octave to make it worthwhile. It has to be well damped with a tight Q. Lots of bass without definition is boring. And it has to have a properly designed crossover, one designed for the woofer and driver above it. A generic crossover isn’t sufficient just like it isn’t between a mid-range and a tweeter if you want fidelity rather than just to impress.

    By the way I have a set of the Snell woofers mentioned in another reply.(crossed at 85 hz) to s a pair of SEAS Froy 3 speakers and with a custom passive crossover all designed by the same man. They are stupendous woofers although mine weren’t designed to go as low as the Snell boxes but were designed with a tighter Q.

    The best deep bass I ever heard was a few years ago at HPs on a pair of Scaena with 4 18″ drivers. He was playing a cut on ‘The Lost World’ with a real volcano in the background. Things fell off shelves and eyes were bulging in fear of flying woofers. These are an equalized design driven below ressonance so they are very well damped but need huge amounts of power.

  10. Soundminded

    December 23, 2011 at 1:22 pm

    I think if I were going to design a servo subwoofer I’d start out with the best dual voice coil acoustic suspension subwoofer driver I could get my hands on and optimize its performace mechanically as a high quality conventional subwoofer first. It should perform at least as well as AR1W before feedback. I’d use one voice coil to drive the woofer, the other as the velocimeter to generate the feedback voltage. I’d use a digital time delay to synchronize the input voltage with the feedback voltage due to the electromechanical propagation delay. I’d also bandwidth limit the feedback loop reverse gain H to prevent unnecessary phase/gain problems. The completed design would have to prove itself not only as being significantly audibly better than the same system without feedback but better than a multiple driver system of comparable cost. For example, servo feedback AR1W would have to outperfrom non servo feedback Teledyne AR9 appreciably to be worth the trouble and expense. It seems to me like a rather daunting task.

  11. acuvox

    December 23, 2011 at 1:31 pm

    Agreed about the necessity of subwoofers, but not convinced about servo. Let me explain: First, the response of hearing is usually mis-quoted as the Fletcher-Munson headphone experiment. The pinnae roll off starting around 300Hz and by the bottom octaves hearing is mostly sensing pressure on the chest cavity – so headphones have nothing to do with it. Other measurements of hearing using speakers indicate much better bass hearing than the Fletcher-Munson curves, indicating extension below 20Hz.

    Further, we hear TIME especially in rectangular modulation envelopes such that the frequency of the rhythm is part of reproduction, as in PRaT. This can literally go to zero frequency. I hear group delay as TIME DISTORTION. When you hit a woofer with an envelope transient, it will ring at any frequencies where there is an impedance peak and distort the waveform, both at the beginning and end of the note. This time distortion is present in all electro-dynamic speakers (diaphragm and voice coil, Kellog & Rice patent 1925). This signal is common in music, corresponding to all percussion and plectrum instruments, but also playable by skilled wind and bow practitioners.

    This distortion is so universal that 99.9% of recordings have the envelope transients attenuated by dynamic processing and/or bass roll off. Even when not explicitly rolled, traditional gradient microphones (cardioid, figure 8, etc), phantom powered microphones and preamps roll at 30-40Hz. A subwoofer will take this resonance low enough to escape the recording limitations. Note that this resonance typically can’t be eliminated by servo control – the system will go chaotic if the Qms is over 2 or so, and there will be insufficient jerk in lieu of a very low inductance achieved with a solid Faraday.

    Another reason for a subwoofer is Doppler distortion. This is inter-modulation from the finite cone motion. 10% second harmonic ranges from pleasant to inaudible, but inharmonic spectral contamination is annoying at .01%. By the time you get to accurate acoustic reproduction levels (115dB for orchestra), you need 10″ midranges, 15″ woofers and three or more subs to control this factor. Again, this something that is not corrected by any servo system I have heard.

  12. Soundminded

    December 23, 2011 at 2:12 pm

    acuvox I have to disagree with you. Tonality of low frequencies is audible to very low frequencies. This can be demonstrated by switching filters or equalizers in and out to sharply attenuate them and observing audible results. This is especially evident in the lowest octaves of pianos, tubas, bowed and plucked basses, and even bassoons. There is undoubtedly a tactile component caused by pressure sensation as well. Exascerbating what is almost invariably inadequate bass response of speakers measured in anechoic chambers and short dimension home listening room cutoff frequencies, much energy is absorbed by the walls, typically 1/2″ sheetrock nailed 16″ on center on 2″ x 4 ” pine studs in most US frame homes. It’s even worse in basements where thin wood paneling is substituted for sheetrock. According to Beranek, even in concert halls, wood absorbs low frequencies unless it is at least 2 1/2″ thick. To make up for these losses, loudspeakers must produce more, not less low bass than flat.

    Fourier theory correctly demonstrates that steady state sinusoidal response of amplitude and phase versus frequency is directly interchangeable with amplitude as a function of time. Transient response to square waves is yet another way of presenting the same data. In fact it is a classical textbook problem to be given one and determine either of the other two. Even in full range drivers for all their shortcomings I’ve yet to hear dopplar distortion myself. Other non-linear distortions are likely much greater. Bose claims they are not audible until they reach 6%.

    My experience is that the best properly designed and functioning acoustic suspension woofers carefully integrated into full range loudspeaker systems can produce bass notes that sound just like bass from musical instruments to my ears. The successful implimentation of servo feedback control can further improve performance, the questoin is whether or not it is worth the cost and effort and can comparable results be achieved by other means that are more cost effective and less complex.

    • December 25, 2011 at 3:19 pm

      Many engineers have attempted to make servo bass systems using a dual voice coil and while they indeed work they are typically not very good – for a few reasons – but chief among them is it’s very difficult to get a true measure of velocity and positional accuracy with a voice coil which gets increasing non-linear at it’s extremes.

      A better way is what we did – which is to use an accelerometer. A really inexpensive accelerometer is a piezo electric microphone or buzzer disc. These are chap and if you mount it properly, the bending piezo element gives a near perfect representation of velocity and position. If you then integrate that signal, you get acceleration.

      Then all you need to do is use differential amplification to compare the incoming signal with the acceleration signal and put the differences in a controller loop.

      Aside from stability of the loop (you’re right there) another problem is power. For every octave below speaker resonance within a defined enclosure (the enclosure must be sealed) you need 4X power to achieve a flat response. You can imagine that flat response to 12Hz or so with a box resonance of 60Hz or higher takes a lot of power – and designers always want smaller enclosures raising the resonance even higher.

      I addressed this problem in one of my posts with a power supply monitoring system that activated an audio limiter.

  13. acuvox

    December 23, 2011 at 4:14 pm

    1) I did not say bass was not tonal – I said that it does not enter the body through the pinnae as indicated by the difference in loudness curves measured via headphones vs. loudspeakers.

    2) Acoustic bass instruments short of pipe organ have attenuated fundamentals. This means that the timbre is made up disproportionately by overtones and explains your filter experiments.

    3) Sheetrock walls do not absorb enough bass to prevent typical 10dB peaks and nodes. The rule of thumb in the pro world is 4lb/ft2 reflects to 50Hz. I find that the only rooms I can tune for flat bass over a large proportion of the floor area have large bass leaks in the form of windows and openings into larger spaces.

    4) The Fourier transform is mathematically perfect – but you have to integrate from zero frequency and include phase, which most calculations conveniently ignore. I am not talking about square waves, I am talking about a rectangular modulation envelope. This is modeled as a sine wave multiplied by a rectangular wave at a frequency sub-multiple. There are many instruments where the first half cycle is at full intensity, and the modulation is effectively a frequency represented by the length of the note. This frequency is both reproduced (or not as the case may be) and audible – unless you are claiming that the length of notes and rhythm are inaudible.

    The resonance of the speaker is excited by the rectangular wave on both the leading and trailing edges, effectively storing that energy for the duration of the note. Further, both lack of radiated energy at the beginning of the note and the trailing energy at the end of the note are at the characteristic frequency of the woofer which is anharmonic to the music.

    I suggest that since you have been listening predominantly to acoustic suspension woofers rather than to acoustic music with rhythmic bass, your hearing has adapted to this group delay time distortion and you regard it as normal. You are also abetted in this by record producers and engineers who expunge these real acoustic events from the recording.

    If you take a rectangular gated sine wave at the impedance peak of a dynamic loudspeaker and compare it to, for example, a headphone with flat bass the waveform distortion is obvious. Unless you are willing to try this simple experiment, your opinion on this matter is invalid because you are only experiencing the warped audio ecology.

  14. Soundminded

    December 23, 2011 at 5:41 pm

    acuvox, how fortunate I am to have a Steinway grand piano in my music room. I’ve also got a Baldwin Acrosonic in the next room. Both produce outstanding bass notes and make it possible for me to compare the sound of recordings with live at any time I care to. There are also string quartets in my house every week and I get to hear real cellos. This in addition to the concerts I attend more rarely now than I once did.

    For reasons I won’t go into now, headphones are not a valid comparison to speakers. I’ll only mention the necessity of close acoustic coupling with a seal around the ear as one factor. Obviously headphones will not reproduce the experience of a pressure wave on your entire body a woofer or live music will create.

    The leading edge of a square wave or rise time relates to high end of the response of a sound system. The trailing edge or fall time relates to the low end response. When the response is flat from zero to infinity, the waveforms will coincide exactly and the square wave response will be perfect. Reproducing sounds beyond the audible spectrum can affect sound perceived in the audible spectrum though intermodulation distortion. This was the classic flaw in a Japanese electrical engineering professor’s demo to his students. When he used a separate driver for the highest audible frequencies from the one he used for the ultrasonic frequencies the audible difference disappeared. Infrasonic sounds can affect tactile sensations but are not heard as sound. I’ll use my over the ear hedphones Sony MDR V-6 to see what effect a low cut filter and low frequency equalization has on perceived bass tonality. I’ll only report on it if something happens I find surprising.

    • acuvox

      December 25, 2011 at 7:29 am

      soundminded, let me address your points more explicitly:

      1) I also have a grand piano in my living room. I toured the Steinway factory with the AES a couple of years ago, and all of their models have the sound board open on both sides with a width of about 54″, the same as my piano. Adding in the 15″ case height makes the instrument a dipole with a bass roll-off at 100Hz. The size and stiffness of piano soundboards also rolls off the bass so the lowest octave has attenuated fundamentals with little acoustic energy.

      Further, at the lowest frequency the transfer of bass energy from hammer to string to wood is delayed so the note takes a cycle or two to reach full volume and the damper mutes the lowest notes slowly rather than a fast cutoff. These factors make the piano a poor test signal for hearing group delay even though it is technically a percussion instrument.

      2) I am not discussing the reproduction of square waves. Since you are apparently unfamiliar with radio terminology, I will explain time distortion in physics terms.

      In many musical instruments energy is loaded into an inaudible storage and then an energy transfer occurs abruptly to an acoustic radiator. The simplest is a bass drum. The drummer accelerates the drum stick accumulating kinetic energy until it contacts the drum head. This initiates instantaneous movement, reaching peak velocity on contact. The drum head will then resonate according to its tension and mass until the energy is dissipated, which can also occur abruptly if the drummer stops it with his hand.

      The resonance of the woofer will experience a second order impulse function – that is, the acceleration of the signal as it turns the corner from zero output to full velocity approaches a Dirac function, with infinite power for zero time but finite energy. This excitation will result in a spurious output at the woofer characteristic frequency, which is not musically related to the signal, and is large enough in amplitude to be audible.

      In terms of signal generators, this is a rectangular gated sine wave. A sine wave oscillator is followed by a gate synchronized to zero crossing which opens for a integer number of cycles and then closes again, such that at the start the air motion goes from zero to full pressure instantaneously, and at the end it goes from full amplitude to zero. My speaker modeling software, LspCAD, optionally models transient response using this signal and the time distortion matches oscilloscope traces of woofer outputs I made with a DIY generator.

      The woofer resonance typically adds one to three half cycles depending on the Qms of the driver. In my experiments I also did this test with a bass guitar. I am not a skilled player, but I could produce notes that closely resembled this signal with an integral number of cycles.

      I think it is no accident that Leo Fender located his electric bass guitar pickups at 22% or less of the string length, a design that accounts for over 95% of instruments today. This provides adequate fundamental suppression of open strings sufficient to mask the muddiness of time distortion in his sealed cabinets with 80Hz resonance.

      Harry F. Olson called the gated sine wave “the most accurate test of loudspeaker quality” and it was the cover art in his penultimate publication, “Modern Sound reproduction”. Bruel & Kjaer and General Radio made gated sine wave generators on the ’70s. but they didn’t achieve the popularity of Heyser’s methods. I have no argument with Heyser, but most practitioners conveniently ignore phase.

      I have been searching for 15 years for recordings with second order bass transients and found a dozen or so, all on obscure labels. The most mainstream are the “Superbass” albums by Ray Brown, and the effect is subtle. Even audiophile recordings use microphones and preamps which roll off at 30Hz and use distant microphones such that the second order bass transient content is obscured by air travel and room reflections. Understand that all woofers exhibit this distortion, so when the recording engineer monitors the session full bass response will sound muddy on his monitors.

      3) Recent experiments in Japan with separate supertweeter showed that although the subjects reported no conscious audible difference, switching off musical signal from 20-50KHz produced consistent autonomic changes measured via polygraph and affected mood.

      Like you, I maintain my hearing via regular acoustic music. We attend about 50 acoustic concerts a year and my wife plays harpsichord.

      • Soundminded

        December 25, 2011 at 12:28 pm

        Acuvox, I must again disagree with some of your points. The impact of the drumstick on the membrane creates a pressure wave within the membrane. At the point of impact the membrane will be compressed to a degree slowing down the drumstick and the remainder of energy will move the membrane at the point of impact at the same velocity as the stick as it slows. Cohesion of the membrane will result in a pressure wave propagated outward towards the rim which will be reflected back and forth as energy is dissipated. This will cause the membrane to vibrate as a Bessel function, that is it will break up into harmonic modes, the longitudinal displacement at any point governed by the superposition of the modes at the fundimental and harmonic frequencies.

        The waveform analysis will show an initial transient attack of less than infinite rise time and energy which will die out with an exponentially decreasing amplitude envelope of the characteristic harmonic frequencies. This attack is not to be reproduced by the woofer alone but by the other drivers including the tweeter. The group delay displacement distortion arises from the fact that the group delay of the woofer and tweeter (and midrange if there is one) are different from each other. The naive attempt to create phase coherent loudepeakers in the 1970s was based on the fallacy that by matching displacement delays the waves would arrive at the listener at the same time and would be in phase. The flaw is omitting the fact that they are not also coincident in space as well and therefore will create interference patterns in space such that no two points have the same patttern. The irony is that in the 1950s the opposite flaw was argued where coaxial and triaxial designs tried to create a point source of sound coincident in space but no effort was possible to make them coincident in time as well. Even if it can be made to work, phase coherence would only occur on axis because as soon as you move off axis, the distance to the woofer and tweeter are different and you arrive at areas of phase interference again.

        I have listened carefully as I slowly played chromatic scales downward starting one octave below middle C listening with particular attention to the fundimentals. It seems to me that the fundimental just gets lower and lower. While the sounding board may be a bi-polar radiator, usually I have the lid shut so mostly what I hear is eminating downwards and reflected back. I also expect that sound will reverse phase when it is reflected off the floor just the way a bass reflex or open baffle speaker works. Since I listen almost entirely to cds, I think many or most have considerable energy recorded below 40 hz, probably below 30 hz. I know that the timbre of the left hand notes in the lowest octave changes when I cut a high pass rumble or subsonic filter in and out. I also know that speakers which cannot reproduce the lowest octave such as KLH Model 6 unfortunately do not accurately recreate the lowest piano tones the way speakers like AR3 can.

        I don’t know what it’s like to experience exposure to just ultrasonic frequencies, the closest I’ve ever gotten to it are industrial ultrasonic cleaners which also produce an audible screeching sound. I can’t believe it would be a particularly pleasant experience and I see the capability for it in sound systems as not only unnecessary but undesirable. This was another myth advertised in the 1960s, the University Sphericon super tweeter good to 40 khz, Audio Fidelity Records with the “Frey Stereophonic Curtain of Sound” recorded out to 25 khz, and Harman Kardon amplifiers including my first amplifer A500 good to 70 Khz and the A1000 flat from zero to light (well 1 mhz anyway in their advertising claim.)

        Sorry I didn’t get a chance to tour the Steinway factory myself. There was a very fine documentary on PBS called “Note by Note.” I think there was a book out too, this all around the time of the 500,000th Steinway produced maybe about 5 or 10 years ago. Mine is a type M, 5′-6″ made in 1927 and restored in the mid 1980s at the Baldwin factory when Steinway unfortunately was undergoing reorganization. BTW, listening carefully to it last night let me know….it’s time to call the piano tuner.

  15. Gordon Meltzer

    December 24, 2011 at 12:04 am

    Hi Paul,

    Your comments on the wonders of servo bass were amusing in the sense that they are analogous to industry raves from 40 years ago, on the wonders of negative feedback in amplifiers. But in a more important sense, thank you for this forum and for your wonderful products, some of which I am lucky enough to own and which enhance the musical experience.

Leave a Reply