John Dunlavy's posts to rec.audio.* during 2000: New messages will be added here as I see them on Usenet; hopefully in less than a week. Thanks to Dunlavy for this information. Since July 1999, Mr. Dunlavy has been posting with much greater frequency and in a more interactive manner than previously; I am archiving primarily the posts where he explains something I believe will be of general, enduring interest. Last revised 22 September 2000. -------------------------------------------------------------------------------- From daljhd@cs.com Wed Jan 26 11:19:57 2000 Newsgroups: rec.audio.high-end Subject: Re: the effect of coil warm up? From: daljhd@cs.com (DALJHD) Date: 26 Jan 2000 17:19:57 GMT Subject: the effect of coil warm up? From: John Dunlavy Date: 1/26/00 8:00 AM MST In his post of 25 Jan., Espen Braathen wrote: "As the loudspeaker gets used I guess the speaker coil will be heated and thus the resistance will increase. Could anyone explain what effect this has (significant or insignificant)?" - - - - - - - - - - - Hi Espen, Yes! The heating of the voice coil has two effects. First, there is a brief "break-in period" for loudspeaker drivers during which their voice coil and supporting cylinder (called a former) heat up to a relatively high temperature. This elevated temperature produces a "curing action" that hardens the insulation of the wire and more firmly attaches the wire to the former. This hardening improves the transfer of high frequency energy from the coil to the cylindrical former and into the apex of the cone attached to the former. This, of course, usually improves the performance of the driver with respect to its impulse-response, step-response, and its amplitude Vs frequency performance at high-frequencies. Many manufacturers of high-quality, "high-end" loudspeakers submit the drivers they use to an appropriate "break-in period" before measuring and selecting them for use. Once drivers are selected for use and the "initial" crossover boards are assembled, the loudspeaker is rolled into a large anechoic chamber, free of reflections that could alter the measured response of the loudspeaker. At this time, the crossover is physically located outside the chamber on a table adjacent to an array of accurate lab equipment used the measure the properties of the loudspeaker. Large guage insulated wires (typically 4 AWG copper), with a length shorter than about 10 feet, are used to connect the crossover circuits on the bench to the loudspeaker (and its drivers) within the anechoic chamber. (The use of such large guage wire ensures that the measured response of the loudspeaker will not change when the crossover is finally placed inside the loudspeaker.) Then, an engineer or well-trained senior technician measures the initial or "start-up" response properties of the loudspeaker using an array of equipment capable of accurately assessing its impulse-response, step-response, modulus of amplitude Vs frequency (frequency response), cumulative spectral decay (waterfall) and input impedance. (During the initial design phase of a new loudspeaker model, measurements also include non-linear distortion Vs SPL, radiation patterns in both vertical and horizontal planes Vs frequency, etc.) During the Q-C of each "production loudspeaker" in the anechoic chamber, the values of capacitors, inductors and resistors comprising the crossover network (located on the test bench outside the chamber) are "tweaked or altered" until all of the advertised performance specifications are met with respect to frequency response, impulse response, step response, waterfall decay, phase response, impedance Vs freq., etc. All of these measurements are recorded and filed for each loudspeaker. Yup, I believe that such full and precise measurements of a loudspeaker are essential for assessing the performance of loudspeakers intended to be called "accurate". Sure, some designers will comment that such measurements mean little regarding how a loudspeaker will sound when heard within a typical listening room where reflections will alter the sound quality. They are wrong! A full set of accurate measurements (including radiation patterns), competently assessed by a knowledgeable person, can reveal a great deal about a loudspeaker's potential to yield audibly accurate reproduction within a listening room of typical size possessing average acoustical properties. However, it is only prudent to submit a new loudspeaker design to a "live Vs loudspeaker" session, where the performance of the loudspeaker is compared to the sounds of live musical instruments - in real-time. We perform this comparison within a special "listening evaluation room" in which there is a Samic grand piano, a guitar, flute, etc. We record the various instruments, individually, within an anechoic environment, using a professional DAT recorder and a pair of ACO Pacific or B&K instrumentation microphones. Our design work is not considered complete until there is no audible difference between the live mucical instruments and their reproduction by the loudspeakers being evaluated. We have demonstrated this "feat" to numerous visitors who believed they would be able to discern an audible difference but discovered they could not. Anyway, I thought some readers might like a brief view of how the accuracy of loudspeakers might be determined - both by a full set of accurate anechoic measurements and properly controlled comparisons with live musical instruments. Comments and questions are welcome. John D. From daljhd@cs.com Mon Feb 07 14:40:45 2000 Newsgroups: rec.audio.high-end Subject: Re: Loudspeaker Shootout was: Re: Metal Drivers, What's up with these? From: daljhd@cs.com (DALJHD) Date: 7 Feb 2000 20:40:45 GMT It appeals to me that neither measurements nor listening, used in isolation, provide a valid or useful yardstick for determining the "true accuracy" of any audiophile component or system. For example, a complete set of truly accurate anechoic measurements of a loudspeaker, made at a distance of 10 ft., can "predict the potential" of the loudspeaker to accurately reproduce complex musical sounds, if competently interpreted by someone skilled to do so. However, the best, most complete and most accurate set of measurements cannot guarantee "audible accuracy" with absolute certainty. By this, I mean that they cannot unconditionally guarantee that a given loudspeaker will reproduce musical instruments with an accuracy that prevents a competent listener from discerning any difference between the live music and its reproduction during "blind", live-vs-reproduced listening sessions. Thus, the "true accuracy" of loudspeakers must be determined by an interactive use of both a complete set of accurate measurements and carefully-controlled, blind listening comparisons with live music (recorded anechoically using very accurate equipment and instrumentation quality microphones). This is not the first time that the above comments have been posted. Perhaps, one of these times they will be seriously considered and discussed. Best of "accurate" listening, John D. From daljhd@cs.com Fri Feb 18 10:54:36 2000 Newsgroups: rec.audio.high-end Subject: Re: Aftermarket cable recommendations? From: daljhd@cs.com (DALJHD) Date: 18 Feb 2000 16:54:36 GMT Subject: Re: Aftermarket cable recommendations? From: daljhd@cs.com Date: 2/17/00 17:30 PM Mountain Standard Time mtrycraft@aol.com, replying to my earlier post, asked: >John, how would the power cord alter the audio signal in the >first place? I thing you missed my point - unless a power cord longer than about 10 ft. and made of wire smaller in diameter than about 16 AWG is used with a 300 w. per ch. amp, it is unlikely to make any audible difference in performance compared to cords using a much larger guage. The whole issue of "super power cords", etc. is a regretable sham that distracts many innocent audiophiles lacking appropriate technical backgrounds from enjoying the meaningful aspects of our delightful and rewarding hobby. Indeed, some of the claims being made for A.C. power cords, loudspeaker cables and interconnect cables are nothing short of outright lies based upon misrepresentations or distortions of the wealth of technical knowledge well known and understood by competent engineers and physicists. Spreading such false information often cheats innocent audiophiles of money that could be better spent for the purchase of items or upgrades capable of yielding a true audible improvement in system performance. If technically meaningful, accurate, guaranteed performance specifications are not available for a cable, etc. being considered for purchase - it is likely that the product claims are based upon floobydust, buzzard salve, snake-oil, etc. and can only lighten your wallet and diminish your faith in what the audiophile hobby has to offer for those who wait for meaningful and well-documented advances in technology. John D. From daljhd@cs.com Sun Feb 27 15:57:38 2000 Newsgroups: rec.audio.tech Subject: Re: Duelund tells another lie. From: daljhd@cs.com (DALJHD) Date: 27 Feb 2000 21:57:38 GMT Lets start by assuming a listener with reasonably good hearing (both ears possessing essentially "matched properties"), seated equidistant and equi-angularly from two "accurately pair-matched" loudspeakers, at all frequencies above about 300 Hz) within a room possessing reasonably decent acoustical properties. When listening to an accurately recorded performance (using a proper stereo mic set-up, etc.) the following perceived reconstructions of the original soundstage will be heard: 1) If the loudspeakers are positioned with an included angle of less than about 30 degrees and equidistant from the listener, the perceived width of the soundstage will be limited and the angular separation between individual sounds blurred. (2) If, in the above example, one of the loudspeakers is slightly closer than the other to the listener, the apparent location of a "center sound" will be skewed toward the direction of the closer loudspeaker. (3) If the angular separation between the loudspeakers is increased, while retaining their equal distances from the listener, the perceived width of the stereo soundstage will be widened and the angular separation between individual sounds will be increased. (Angular separations of from 90 to 120 degrees can produce incredibly life-like imaging with loudspeakers "precision-matched" as a stereo-pair.) (4) If, within a very large room, the angular separation is increased to 180 degrees (the loudspeakers being to the left and right of the listener and symmetrically located with respect to all walls), the perceived width and location of instruments will approximate that heard though a pair of stereo headphones. (However, if even the slightest difference exists between the path distances from each loudspeaker to the center of the listeners head, the stereo soundstage will quickly dissolve.) (5) The dimensions of the listening room, the symmetry of the loudspeakers (and the listener) with respect to room boundaries (and some very large objects), reflections from walls, etc., can audibly alter the perceived width of the soundstage, the separation between instruments, instrument "size, width and height", etc. Hopefully, this info (gained from many years of experimentation, etc.) will help some readers to better understand the influence of room boundaries (and the location of loudspeakers within a room) on the perceived imaging and soundstage available from their loudspeakers and system. John D. From daljhd@cs.com Sun Feb 27 17:00:08 2000 Newsgroups: rec.audio.tech Subject: Re: Looking for an inexpensive loudspeaker measurement tool From: daljhd@cs.com (DALJHD) Date: 27 Feb 2000 23:00:08 GMT Sorry, but the example of a loudspeaker's frequency response (modulus of ampl. vs. freq.) shown on the subject website is totally unrealistically "flat". Using even the most accurate available tweeter, mid and woofer drivers, mounted in a "time and phase aligned configuration", a multi-element minimum-phas x-over network, etc., an "anechoic response" no flatter than about plus/minus 3-5 dB (without "smoothing") will be achieved "on-axis" at a distance of 10 feet. To do better than this requires accurate physical "time-alignment" of all drivers, the use of absorbing materal between drivers and the edges of the enclosure, numerous "compensation networks" within the crossover network to flatten out the many peaks and valleys in the response curve, proper amounts and kinds of "acoustical absorbing material" within the mid and bass enclosures, etc., etc.. And what about achieving accurate impulse, step and waterfall responses - all important for achieving truly accurate audible reproduction. The same applies to radiation patterns and "off-axis" responses of a loudspeaker. A well designed, accurate loudspeaker must also possess a smooth curve of impedance vs frequency, without any high values of reactance. The list goes on and on - representing a task that is unlikely to result in loudspeaker that is truly accurate. However, if a reasonably competent audio amateur merely wants to design a "good sounding loudspeaker", chances are that the subject measurement gear might prove helpful. It all depends upon the goal: true audible and measurable accuracy or a loudspeaker that delivers subjectively "good sound" quality. The bottom line is simple: if "true well-documented loudspeaker accuracy" is the goal - there is no substitute for technical competency, lots of relevant experience, and access to a professional-level measurement facility. Having said all this, I should add that I started designing loudspeakers back in the mid-1940's while still at university, using minimal test equipment, no anechoic chamber, etc. My speakers sounded good until I compared them to some professionally designed speakers at the broadcast station where I worked - and where the Chief Engineer knew more than a little about loudspeakers. So - if you want to design "reasonably good sounding loudspeakers", but not truly accurate ones, begin by acquiring a minimal amount of test equipment (an audio oscillator, good mic, oscilloscope, etc.), depend upon pleasing your ears and A-B comparisons with professionally designed loudspeakers (using equal volume levels, etc.), and so forth. And don't be dissappointed if your first endeavors don't match your expectations. (Neither did mine!) Best of luck, John D. From daljhd@cs.com Wed Mar 08 17:43:31 2000 Newsgroups: rec.audio.tech Subject: Re: Martin Logan From: daljhd@cs.com (DALJHD) Date: 08 Mar 2000 23:43:31 GMT Subject: Re: Martin Logan From: John Dunlavy @DALJHD@cs.com Date: 3/8/00 4:P.M. Mountain Standard Time Eric rmtlaser@uswest.net in Msg.-id.:<38C6C732.BEB89ED7@uswest.net> wrote: >Anyone who listens to Horn or the better panel loudspeakers >knows that cone speakers are compressors at any level. I >have even seen them used in a studio as compressors! >Cheers >Eric Hmmm! Perhaps Eric has some actual measured data to prove his comments regarding cone driver compression, etc? I and my technical staff have measured the linearity of numerous "audiophile" cone and dome drivers and have never encountered the "compression" Eric seems to believe exists - even at levels well above those normally encountered in audiophile listening, e.g., 105 dB SPL at 1 meter on-axis. . However, measurements of the impulse-response, step- response, waterfall, etc. of horn-loaded drivers and membrane loudspeakers with large-surface areas, etc., reveal very poor accuracy. Those with a decent background in Physics can easily appreciate why this is so. To begin, horn radiators exhibit "dispersive" properties in both time and frequency domains that inherently leads to inaccurate reproduction of impulses, steps, etc. Likewise, loudspeakers using large, planar, radiating surfaces encounter "beaming" problems due to their narrow, "far-field" radiation patterns. Their reproduction of transients, etc. is also poor because their low efficiency leads to the storage of energy. Another reason for their poor transient response has to do with reflections from the exterior rigid frame used to support the diaphragm, causing the appearance of "standing waves" across their surface. Is this mere guessing or uninformed speculation? No - ask any competent physicist for their "considered opinion" regarding the impulse response, radiation properties, etc., of a membrane surface activated by a loosely-coupled field, etc. - with little of no "restortion force" present to return the membrane to its starting position after the signal voltage is removed. Further, I have accurate anechoic chamber measurements of a variety of planar loudspeakers to prove the veracity of the above comments. Perhaps, it is time to let the truth be known regarding the "TRUE" reproduction accuracy of planar surface radiators. John D. From daljhd@cs.com Sun Mar 12 18:44:36 2000 Newsgroups: rec.audio.tech Subject: Re: Martin Logan From: daljhd@cs.com (DALJHD) Date: 13 Mar 2000 00:44:36 GMT Subject: Re: Martin Logan From: DALJHD@cs.com Date: 3/12/00 In his post of 3/8/00, Re: Martin Logan, Murray Peterson commented, "daljhd@cs.com (DALJHD) wrote: >>Their reproduction of transients, etc. is also >>poor because their low efficiency leads to the storage of energy." Murray replied: >I would think that the energy storage of a cone drive would be much >larger due to its higher mass, and not really have anything to do >with the efficiency of the driver (at least directly). An >electrostatic with a well designed direct-drive amplifier is very >efficient, which would seem to invalidate your assertion. First, the membrane loudspeakers with which I am familiar are not "very efficient", as Murray claims. Indeed, the reason that full-range electrostatic loudspeakers require large radiating surfaces (diaphragms) is because their "efficiency per unit of surface area" is very low compared to that of conventional cone and dome drivers with much smaller radiating surfaces. (The same applies to loudspeakers using membrane diaphragms with large surface areas activated by magnetic fields.) Second, because of their relatively low efficiency per unit of radiating area, membrane diaphragms "store" the energy not radiated during each cycle at each frequency. Some of this stored energy is dissipated as heat - but much of it is also radiated as "ringing" following impulsive signals, etc. Proof of the existence of this stored energy is readily seen by examining the impulse, step and waterfall responses of any large surface membrane loudspeaker. Indeed, all of the loudspeakers using membrane diaphragms that we have measured (quite a few), including both magnetic and electrostatic types, exhibited considerable impulse-response "ringing" - many, many times that of a well-designed loudspeaker using "time-aligned" cone and dome drivers fed by a well-designed 1st-order crossover network. Such stored energy is also usually evident by the large number of "peaks and valleys" in the modulus of amplitude Vs frequency exhibited by membrane loudspeakers. (Well designed loudspeakers using cone and dome drivers do not exhibit these rapid excursions in amplitude response peaks, etc. Nor do they exhibit radiation patterns with the narrow beamwidths and high side-lobe levels exhibited by loudspeakers with large area membrane diaphragms.) Murray further questioned whether accurate, meaningful measurements exist to prove what I have said about loudspeakers using membrane radiating surfaces. They certainly do and I will soon FAX or e-mail a full set of accurate anechoic measurements of typical membrane loudspeakers and properly-designed loudspeakers using cone and dome drivers.ould Best of listening, John D. From daljhd@cs.com Thu Mar 30 09:25:49 2000 Newsgroups: rec.audio.tech Subject: Re: Phasew coherence From: daljhd@cs.com (DALJHD) Date: 30 Mar 2000 15:25:49 GMT As I said, the differences between the 1st and 4th order crossovers were clearly audible when listening to "impulsive type sounds and tone bursts" within one of our large anechoic chambers. I agree that the greater audibile differences we heard between the two crossover slopes within one of our audio listening rooms would best be explained by their different directivity patterns. After conducting a lot of listening comparisons between loudspeakers using different crossover slopes and time-aligned and non time-aligned drivers, I have come to believe that the combination of a first-order crossover feeding time-aligned drivers offers the most accurate reproduction of complex musical transients, etc. This conclusion has been repeatedly proven true by numerous "blind" comparisons with live musical sounds, transients, etc., conducted within both reverberent and anechoic listening environments. We have also verified this audible superiority by comparing loudspeaker reproduction with that heard through SONY MDR-7506 headphones - a standard for accuracy among top recording and mastering engineers Best of listening, John D. From daljhd@cs.com Wed Apr 05 11:34:07 2000 Newsgroups: rec.audio.high-end Subject: Re: Listen & decide for yourself From: daljhd@cs.com (DALJHD) Date: 5 Apr 2000 16:34:07 GMT Nonsense! Today's well-equipped audio labs can measure all of the audible (and other) properties of cables at levels very, very much lower than the threshold levels of human perception. The problem lies not with measurement technology and measurements but with the "imaginative powers" of those who (for whatever reasons) want to believe that audible differences exist between cables within a properly operating audiophile system. During the past several years, I have hosted many visiting audiophiles, mag editors, etc. who wanted to prove during their visit that they could discern audible differences between #12 AWG Zip Cord and several brands of expensive, large diameter cables. Using time/phase coherent loudspeakers (with a plus/minus 1 dB "freq. resp.") fed by state-of-the-art power-amps, cables were periodically switched with the listener knowing which cable was being heard. The listener almost always claimed to hear differences between cables that they described in considerable "detail". However, after a few cycles of switching, without the listener being aware, the cables were no longer switched - although the cycle of switching appeared to continue. Amazingly, the expert listeners almost always continued to "hear" the same differences as before - although listening to the same cable without any switching taking place. Since the listeners were made to feel comfortable and were never under any stress, believing that the cables were actually being switched, I am only able to conclude that the "expectation effect", as it is often called, played a key role in them honestly believing that audible differences existed and were being heard. Of course,.there may be cases where the loudspeakers being used are poorly designed and exhibit a wide range of input impedances - especially a high series inductance, high parallel capacitance, or a very wide range of input impedances vs frequency. Thus, though highly doubtful, it is theoretically possible that some poorly designed loudspeakers fed by poorly designed amps (with an unusual sensitivity to small differences load impedance) could result in differences being heard between loudspeaker cables. Certainly, the same "perceived differences" might also hold true for interconnect cables used to connect poorly designed pre-amps to power-amps, etc. Indeed, exploring possible audible differences between interconnect cables, we have conducted A-B comparisons with high-quality, low-capacitance interconnect cables, pitting some of the most expensive ones against those available from Radio Shack. We obtained the same results as those with loudspeaker cables - the "expectation effect" played a major role in the "perception of audible differences". And, as with loudspeaker cables, accurate measurements of the electrical properties of several popular brands of interconnect cables, e.g., capacitance, loss vs. freq. into various load impedances with different source impedances, V.P. factor vs. freq., impulse response into different loads, etc., etc. failed to reveal any properties that might explain the audible differences being claimed by some listeners. Of course, very long lengths of interconnect cables possessing a very high capacitance might result in a "roll-off" of high frequencies if used with some tube equipment exhibiting relatively high input and output impedances. But with today's generally well-designed amps, pre-amps, signal sources, etc., the existance of genuine audible differences between interconnect cables of reasonable length is not very likely. But if you prefer to ignore the findings of competent engineering and science and sincerely believe that audible differences might exist, go ahead and purchase expensive, exotic looking cables and enjoy the fantasy. John D. From daljhd@cs.com Sun Apr 30 18:48:41 2000 Newsgroups: rec.audio.high-end Subject: Re: speaker design software From: daljhd@cs.com (DALJHD) Date: 30 Apr 2000 23:48:41 GMT Loudspeaker design "by the numbers", i.e.,following "cookbook instructions" will only yield a speaker with an "un-smoothed" ampl. vs. freq. of about Plus/minus 3-5 dB - if the builder is lucky. Then there is the question of the loudspeaker's impulse response, step response, waterfall response, non-linear distortion vs. SPL, input impedance, etc. - Gulp! It is unfortunate that "kit builder" mags, suppliers, etc. seldom (if ever) inform people that the freq. resp. of even the best and most expensive loudspeaker drivers (tweeter, mid and bass) barely meet plus/minus 3-5 dB limits (on axis) out-of-the-box. (The advertised measurements for drivers typically portray a "smoothed response", measured at a distance of only 1 meter, or less, rather than the "real response", measured anechoically, without smoothing, at a distance of about 3 meters.) Access to an "anechoic (or quasi-anechoic) measurement capability" allows a designer to add resistors, caps, and inductors to the crossover circuit (or alter their value), while "interactively" measuring the results obtained, until the desired level of performance is achieved. The design of "truly accurate" loudspeakers is hardly a text-book exercise or one to be undertaken without appropriate experience and proper measurement capabilities/facilities. However, having said this, everyone has to start somewhere and using a "cookbook design" as a learning experience can offer some very tangible rewards. But having a background in E.E. or physics will prove to be very helpful - even if it is at a "basic level". Thus, I would encourage do-it-yourself speaker designers to find and read some books on loudspeaker performance and design, check-out drivers and their specs from competent sources like Vifa, Scan-Speak, Seas, etc., buy an inexpensive but reasonably accurate microphone, an oscilloscope, a good multimeter with a dB scale, etc. for obtaining basic-level measurements. At the same time, make friends with a competent person (preferably an audiophile) from the E.E. or physics dept. of a local university and go forward - "slowly". And don't forget: loudspeakers may look simple - but accurate ones are probably the most difficult to design of any component comprising an audiophile system. Best of listening, John D. From daljhd@cs.com Mon May 01 10:56:36 2000 Newsgroups: rec.audio.high-end Subject: Re: speaker design software From: daljhd@cs.com (DALJHD) Date: 1 May 2000 15:56:36 GMT In commenting on my post, Steen Dueland wrote: "You are so right in this view, but why then do you stick to the first order solution, there are others with the same virtues and easier to realize?" Well, lets see! Linear systems theory teaches that the least accurate component within a system of series-connected components determines the "accuracy limits' of the entire system. If this is true, and I believe it is, why are we so concerned with the square-wave reproduction, impulse-response, etc. of audiophile amps, pre-amps, CD/DVD and SACD players, loudspeaker cables, etc - but not loudspeakers.? Perhaps, there is something about this "equation" that I do not fully understand? Have some of the subjective listening evaluations, usually performed under questionable, poorly-controlled circumstances by persons often lacking appropriate credentials, fully proved that loudspeakers are not required to possess the same level of" waveform accuracy" as other system components? And, yes, I have read the many papers published in AES Journals, etc., regarding the audibility of various loudspeaker properties and believe that many of their conclusions resulted from listening sessions being conducted under less than rigorous circumstances, etc. My own (DAL's) carefully-controlled listening sessions that have pitted loudspeaker reproduction against live musical instruments and voices (accurately recorded anechoically) have pretty clearly validated the need for loudspeakers to be truly accurate with respect to modulus of ampl. vs. freq., impulse-response, step-response, low levels of non-linear distortion at typical listening levels, controlled symmetrical radiation patterns, etc. If such properties are obtainable without adding significantly to the price of loudspeakers intended to be truly accurate, why not achieve them? I am prepared to be enlightened. John D. From daljhd@cs.com Mon May 08 10:25:05 2000 Newsgroups: rec.audio.high-end Subject: Re: need advice on position tolerant speakers ? From: daljhd@cs.com (DALJHD) Date: 8 May 2000 15:25:05 GMT I believe that the location of loudspeakers within a room requires attention to two separate listener requirements: 1) best "spectral balance" (combination of direct-arrival sound with reflected sound produces "standing waves" and considerably alters spectral balance with different loudspeaker placements relative to room boundaries, etc.). 2) the most accurate imaging (most precise soundstage reproduction) requires that both loudspeakers be located equidistant from and equi-angular relative to the desired primary listening seat. Imaging is usually also best when the loudspeakers and the "primary" listening location are symmetrically located with respect to the side walls and the wall behind the loudspeakers. The wall behind the listening seat can also affect the accuracy of the perceived soundstage - but usually to a lesser extent than the side and front walls. John D. (P.S. Peruse a paper entitled "Listening Room Considerations" which can be found at www.dunlavyaudio.com. It provides a lot of useful information relative to getting the most from the proper location of high-accuracy loudspeakers within various listening rooms, etc.) From daljhd@cs.com Wed Jun 14 10:31:16 2000 Newsgroups: rec.audio.high-end Subject: Re: Hales Reference III From: daljhd@cs.com (DALJHD) Date: 14 Jun 2000 15:31:16 GMT Hi! For a "woofer port" to significantly alter the directivity of a loudspeaker, three main conditions would have to exist: 1) The radiation from the port would have to be significant in amplitude compared to the direct radiation from the woofer . 2) The port would have to be located a significant distance from the woofer, relative to a wavelength at the frequencies of concern, e.g., at least about one-eighth of a wavelength. 3) The "phase" of the sound energy radiated by the port relative to that of the woofer would have to be such that the two combined in a manner capable of shifting the axis or width of the radiation pattern, creating partial-nulling in some directions, etc. But, it is not likely that a typical port will measurably or audibly alter the radiation patterns of a loudspeaker John D. From daljhd@cs.com Sun Jul 09 10:49:32 2000 Newsgroups: rec.audio.high-end Subject: Re: Stereophile Recommendations: Reliable? From: daljhd@cs.com (DALJHD) Date: 9 Jul 2000 15:49:32 GMT It is written in stone: "Never argue with anyone who buys ink by the barrel and paper by the roll." So true, as I have now discovered! J.A., in forwarding me a copy of his "EQUIPMENT REPORT FOLLOW-UP" entitled, "Dunlavy SC-IV/A Loudspeaker", clearly stated that DAL would not be allowed to provide any "Comments", as are usually accepted for the "Manufacturers Comments" section of the magazine. Hmmm! This is unfortunate because J.A.'s "Report" contains many significant errors associated with his attempt to measure the important properties of a large size loudspeaker, like the SC-IV/A, at a relativey short, on-axis distance, within a relatively small room (with lots of reflections from nearby walls, the ceiling and the floor). The use of the MLSSA system does permit the taking of measurements within relatively small rooms with reflective surfaces, but only with a restriction being placed on the lowest frequency at which " reasonable accuracy" can be obtained, etc. But measurements taken at a distance of only 8 ft., within a room the size of J.A.'s (only 16' X 19' X 8'?), with reflections from nearby floor, ceiling and wall boundaries arriving only a very few milliseconds after the direct-arrival signal, cannot be considered to accurately represent the properties of any loudspeaker the size of the SC-IV/A. By contrast, DAL has invested in two, very costly, large, "world-class" anechoic chambers (24 ft. long X 20 ft. wide X 16 ft. high) with all internal surfaces covered by alternating, deep wedges of high-density, open-cell, acoustical foam. Believe me, these chambers are really "acoustically dead" and reflection-free at frequencies above about 130 Hz. (I lost a $10.00 bet with an employee that I could spend a full ten minutes within one, shortly after they were constructed.) A third, smaller chamber is used to perform "incoming-Q.C. measurements" of every driver selected for use within a production loudspeaker. And, equally important, both of the two large chambers are accompanied by a wide range of lab-quality equipment, including MLSSA systems (operating on "fast H-P computers"), Time-Domain Spectrometry equipment, Low dist. HP sig, gens, HP Spectrum analyzers, White 1/3 octave real-time analyzers, etc. etc. Thus, we believe we have the capabilities and experience to make measurements of loudspeakers with "guaranteed" levels of accuracy. I wonder if Stereophile would entertain such a guarantee for the accuracy of their measurements. So! I will let readers judge whose measurements they wish to believe - DAL's or Stereophile's? John D. From daljhd@cs.com Wed Sep 20 14:33:10 2000 Newsgroups: rec.audio.tech Subject: Re: Speaker impedence question From: daljhd@cs.com (DALJHD) Date: 20 Sep 2000 19:33:10 GMT Whoh! Attempting to design an "accurate" loudspeaker, one with a plus/minus 1 dB ampl. vs freq., excellent impulse, step and waterfall responses, etc., is an exercise in folly if one uses only PSPICE. First, even the most accurate mid-range and woofer drivers exhibit no better than a plus/minus 4-5 dB modulus of ampl vs. freq., an impulse response with copious "ringing" and a step response with king-size glitches. And the best tweeter drivers are not much better. Without a good anechoic chamber, a reference-standard microphone, and a powerful measurement program (such as MLSSA) running on at least a 486 (or Pentium) computer, designing a truly accurate loudspeaker would task the resources of a "Divine Being" to the limit. Of course, designing a loudspeaker that sounds subjectively "good, pleasing, musical, etc." is an entirely different matter - as many hobbyists have probably discovered. But most critical audiophile listeners eventually grow tired of the sound of such loudspeakers and become attracted to loudspeakers that exhibit a more accurate level of performance with respect to ampl vs freq., impulse response, step response, waterfall, well-controlled vertical and horizontal plane radiation patterns, relatively flat input impedance (amplifier friendly), low levels of harmonic, IM and doppler distortion, etc. But many audiophiles are fully content with loudspeakers that merely sound "musical", etc. Indeed, in many instances, "accuracy" is a purely subjective attribute that resides only in the mind of each individual listener. But listening to well-recorded music through a high-quality audiophile system feeding a pair of truly accurate loudspeakers can be an experience that inspires many audiophiles to reconsider their definition and appreciation of "accuracy". John D. End of archive. New posts will be added as I become aware of them. -------------------------------------------------------------------------------- I hope you find this information as interesting and useful as I have. The other contributions to these threads (and many others) can be found by searching at www.dejanews.com (deja.com) and/or in the rec.audio.high-end archives. These and other topics are also discussed in a series of 'white papers' available from Dunlavy Audio Labs, as well as in the August 1996 'Stereophile' interview with John Dunlavy. I have no connection with DAL except customer (and enthusiast...:-).