| Columns Retired Columns & Blogs |
August 7, 2008 - 5:02am
#1
Loudspeaker distortions
| Loudspeakers Amplification Digital Sources | |
Analog Sources Accessories Featured | |
Music Columns Retired Columns | |
Show Reports | |
Features Latest News Community | |
Resources Subscriptions |
© 2024 StereophileAVTech Media Americas Inc., USA
All rights reserved
I too would like to know the reason for no speaker distortion measurments. I realize and have no problem with the fact that speaker distortion levels will be in the 5% to 10% range and make the debate over amplifer distortion levels seem insignificant. There I go again, looking at the overall system "big picture".
The definition of "distortion" is always a moving target. Electronically measurable distortion is more abstract than speaker distortion, since the speaker depends so much on what is driving it, both in terms of source components and in terms of amplification. Electronics can "drive" test instruments without speakers. But speakers have to be driven. There are no electronically isolated abstract measurements of the physical vibrations that appear, finally, to your ears. The speaker has to be driven, or it is the sound of one hand clapping. The speaker is at the end of a long chain. Finally, it either sounds musical, in your room, with your software, or it doesn't.
Distortion percentages don't matter, with speakers. There is a reason why JA doesn't attempt to concoct such a measurement. He understands that there is a built-in futility for some measurement attempts. Why waste your time. Most speaker "measurements" are based on how they handle an input signal. They have to be driven.
Paul Klipsche, an audio pioneer to whom we all ought to bow down in homage, for his contributions to speaker design over the years, once made the mistake of pursuing this red herring. He claimed "distortion" of less than 2% in his K-horns. But he had to drive them first, and, even more difficult, he had to concoct a truly baroque scheme for setting the beasts up and "measuring" them.
And for all his trouble? A speaker that played loud and clean, but created true, physical pain in the listener's head. I mean, PAIN. I know. I owned 'em. It was love/hate all the way. Massed violins sounded like shattering glass at a car wreck. Bass was divine. Tight, clean, and deep. They only measured down to 40 Hertz, but they absolutely wiped out the competition that was claiming "flat to 25." The only thing that was worse than listening to K-Horns was a trip to the dentist, come root-canal time. And not by much.
So much for "low distortion" in speakers. That is like claiming you drive a golf ball with your hands. Your hands are at the end of the chain, just as the speakers are at the end of theirs. We have seemingly mastered the art of measurement. But not with speakers. Not by a long shot.
The speaker has to be connected to an amplifier - that is the essential fact of the matter. Which amplifier? One with a good degree of negative feedback? That would assure a good degree of damping factor and the lowest overall distortion figures.
How many of us use high NFB amplifiers?
Yep, I thought so. So which amp becomes the "reference" that qualifies as standing in for all the various amplifiers someone might use? I use tubes. Not much feedback and not much damping factor. Not much point in using my amplifier to test distortion from the speaker. But then the numbers with a solid state amplifier won't mean much to me either. A high current design or a high voltage design? It matters when dealing with how the speaker reacts because the speaker is reactive.
What sort of signal are we going to measure? Steady state or impulse? Well, what does the music you prefer rely on most? Unless you only listen to sustained organ tones there is no good answer for that question but a speaker can display considerably more distortion of certain types with either type of signal. And how many signals do we inject into the speaker to take the measurement? Music is seldom one or even three frequencies. Which frequencies shallw e emasure? A monitor isn't going to measure the same as a full range system.
How many locations shall we plot? Off axis distortion is higher on some types of drivers than on others. How relevant is off axis distortion if you have your speakers toed in?
How loud will we go before the measurement is taken? At 100dB a K'Horn isn't even warmed up while an 82dB monitor is straining to it's limits and probably beyond. If the speaker system includes a passive crossover, warming up is very important since the temperature of the components within the speaker, including the voice coil(s) of the driver(s), will change the operation of the system.
What sort of distortion should we measure? Most drivers should produce relatively low levels of high order harmonic distortion but can produce large quantities of IM distortion when reproducing a music signal. If the measurement is of benign harmonic distortion, is it that important since all speakers will produce harmonic distortion? Is harmonic distortion by itself of any real value unless we know how it breaks down in order of additive information? According to JGH's logic of measurements, no.
Shall we measure out of bandwidth distortions? A multi-way system would have a different consideration for this measurement than a single driver system. The single driver though would have other problems not seen in the measurements which would reveal the inherent issues of a multi-way system.
Jan, your questions are both valid and important. All quantative tests should be performed using parameters that fit the application of the device being tested.
Easy to say, but not so easy to define.
While, as you point out, there are many factors that influence a distortion measurement and by itself is not a significant specification, it can be used as one more piece of information along with all of the other data.
Sure it can, if you think these sorts of measurements are really that important. I don't. Is the difference between two speakers - one with 10% THD at 30Hz@90dB and another with 12% THD at similar specs when driven with impulse of less than 30 microsecond - more important than one that has off axis response that excites the room reflections of my room? Or significant diffraction from the cabinet edges? Or constricted vertical dispersion?
Ethan has one point that is worth considering, the amount of distortion or variation from flat response in your amplifier is so much less important than how your speaker reacts with your room. Do you choose your amplifier by looking at THD measurements?
A rising frequency response in a speaker with limited off axis response is comparable to a speaker with flat response with broad off axis response how? Concentrate on what's important. Is the speaker going to drive your amplifier to premature suicide? That's what I need to know, the rest is pretty much secondary. If it changes with what I play through it, it doesn't much matter to me.
I didn't include it in my post but what about a speaker system that is spiked vs. not spiked? Anyone who has experimented with spikes knows they lower the distortion when properly installed by minimizing the random motion of the cabinet relative to the driver. So are the measurements to be made with spikes if they are not included with the speakers? That would rule out a lot of standmount speakers. And panel speakers. Now are we going to review the spikes or the speaker?
This all gets to the point of madness where you just want to shout, "Just go listen for chrissake!!!"
They all "know" that spikes lower speaker distortion based on what? Subjective analysis or by subjective measurement?
No I don't select my amplifier based on its distortion.
Why should I care when I know that my speakers distortion level is of a magnitude that completely swamps out the tiny insignificant level of distortion from any available amplifier.
Your comments about distortion measurements can also be said for all other measurements. There are many ways to measure any parameter and each method has limitations. Does this mean that measurements are useless and have no purpose? If you believe that, then you would be better served by reading Absolute Sound. Measurements are very usefull when used with the knowledge of each measurements limitation and its relationship with other measurements.
Hell yea!
No, but IMO it means most measurements are not that relevant on the consumer end of the process. They are relevant to a designer but I assume the product I buy has been designed by someone with competence in the process and I'm aware of those components that are designed with end results dissimilar to my own desires.
A mass market receiver, as we all should know, is not designed for ultimate sound quality but the measurements are excellent by virtually any technical standard - if you don't look too closely.
Ask anyone who believes all amplifiers sound alike when they measure alike just which measurements tell them that. It won't take long before you've reached the point of amplifiers not measuring the same. You merely need to get beyond watts and gross THD and SN measurements in most cases. Ask about weighting systems that allow a component to look quite good while sounding quite bad or not performing in the real world at all. Find the rare budget product that does sound quite good and measurements typically don't stack up to a logical conclusion. Ask just what that new "ASAK" circuit is which replaced last years "SAKA" circuit.
We should all know the drill by now.
A tube amplifier doesn't measure like a transistor amplifier. So which should I pick? A single tube, single ended amplifier doesn't measure like a paralled, push pull amp will. And what are we to do with those oddball OTL's?
I read Stereophile and I see AD gush over a product that JA can hardly find the ability to say anything good about after measurements are taken. Which portion of the review should I trust? Why am I reading Stereophile in the first place? Only partially because TAS has hit another downslide into two page reviews that say nothing and a magazine full of no information other than the bizarre systems HP puts together and promises to review in future issues and then never does because he gets more new toys to play with. TAS has had it's moments in history, now is not one of them.
I always read the measurements section of the Stereophile reviews. How much useful information do I find there? Not as much as I do in ST's or AD's columns without any measurements. I view the measurements as providing an insight into how the designer went about their job. To give a full view of what a product can accomplish is generally not possible due to the limiations of space and variations in system matching. Every measurement, if you're interested in measurements, begs to have another measurement taken and then another and then another ... Without those others, you are many times only guessing what the preceeding measurements truly mean.
If measurements are your thing, there are magazines that do those things and tell you nothing about how the product sounds - but you'll get a great description of how many buttons are on the remote. Read an objectivist magazine and two sentences about sound quality are "copy/pasted" into every review. Just like Hirsch used to say, "It sounds good and it represents good value for the money. We recommend you give this component an audition."
What about the music?!
Jan is wrong once more of course. Read this from a master at speaker design. http://legacyaudio.com/index.php?option=com_content&task=view&id=63&Itemid=192
http://legacyaudio.com/index.php?option=com_content&task=view&id=60&Itemid=189
Once more upd reads only what he wants to read and understands only a small portion of that. I'll assume this is what you wanted me to read despite the fact you were too lazy to point to anything specific in the linked article.
Clearly, we see that there is more to the story than what on paper measurements would lead you to think.
That's what I've said, upd! If you would have read and put some effort into comprehending my post, you might have seen that. Instead you just want to inject another personal insult in order to make you feel better about yourself and your pitiful amount of comprehension skills. I-IV-V on that ol' eight bar Blues standard, "The Elliott Papers Scandal"!
The problem with what you've swallowed and can't hack up is the Thiel and the B&W are both recommended components in Stereophile. Do you seriously believe a speaker with 66% THD could be a class A Recommended Component if that were the only measurement that effected it's performance? Do you seriously believe that 66% THD is relevant to how a consumer might use the product and no one at Stereophile noticed? Why don't you explain how either of those situations could happen, upd?
Do you seriously believe that THD alone is the relevant number - without any other information - when it comes to audio? The K'Horn does have incredibly low THD. It still sounds like pounded puppy poop.
To any thinking human being what Legacy has posted as "truth" sounds like another case of, "Let me see a few more numbers, please." One hundred watt input for the B&W's? 100dB output for the Thiels? Why aren't those two measurements the same? To a casual reader - the worst kind - the numbers would easily appear to be similar. But they obviously are not and yet they appear side by side as if to make you, the unthinking reader, put them together in your head as identical. Wouldn't you want measurements that reflected similar conditions if you were going to draw rational conclusions from the information? Wouldn't you want to know more about how a respected speaker company can produce a respected product that would appear - from the numbers provided by a competitor - to be dogshit?
Not if you're going to be dishonest in your advertising. Why do that and what does it mean about how you run your company? Could it be because Legacy wants to play mass market manufacturer games and hopes no one will notice? There's absolutely no mention of input or output level when they exclaim the superiority of their product, is there? What conclusion might we draw from that omission?
Of course, if you just want to read what confirms your own uninformed, biased opinion, I guess you don't want to get too close to anything that might contradict your "opinion". And what upd has shown us is an example of not thinking at its best.
Jan,
Of course, you're right about incomplete specifications or test parameters misleading the reader. I skimmed the article and the major take-away that I could see that matches acoustic theory is that the diaphragm should move very little to minimize distortion.
As far as I can tell, speaker designers balance the distortion and excursion (loudness level) among many other things, as we all know. It's the speaker that can play loud, show low distortion and maintain linearity (e.g. frequency response and spl output, throughout its range of operation) that should both measure well and sound well, if there is at least, correspondingly very low non-linear distortions. This is not a catch-all description, but I think we all can appreciate the challenges the designer faces.
(Edited for spelling)
cyclebrain,
Yes, speakers can have relatively high distortion measurements compared to amplifiers. But I've read some time ago that there are some speakers playing at a relatively normal level that have quite low distortion, perhaps in the midrange frequencies. Either way, speaker distortion is not readily measured or widely published. I also seem to remember seeing some distortion figures published for subwoofers as a marketing tool - lower distortion mens better bass, thereby differentiating that manufacturer from another.
But as a tradeoff in trying to get low distortion by reduced excursion and maintaining high SPL a large surface area is used. If using one large driver then response is sacrificed because of mass. If using multiple drivers then frequency/phase interaction between the drivers surface.
Damn, it's always something.
For some people, the speaker distortions measurement seem to be a problem.
Are this people manufacturers? As I know a lot of measurements were done at BBC in the 60ties at the development of the BBC monitor range.
Think LS 3/5 !
Of course Jan you miss the basic point bill D. is making, he is merely writing a synopsis of teh concepts of moving air at low distortions, at realistic levels, something you don't understadn since mini speakers do nothing of the sort. His concept is to move air realistically it takes the drivers of size and drivers to match the area they are working in. Not one driver to do it all. Once you realize the stupidity of trying to move all freq with a single mid range, you will understand what hi fi really is. All these companies that claim their miniature speaker system will fill the room etc are basically full of crap. BLOSE has done it for decades, and many others have copied his route, it just doesn't work, never will. Jeeeeeez, iI knew that in teh early 70's when I was a kid just getting into real sound, which is why I had speakers stacked to teh cieling in a room. One day you will have an awakening, and discover true hi fi, realism. Bill D. makes it happen, he understands it all, and works it into a system that does it masterfully. And priced for mortals, living in a real world. This synopsis on his website is pointing out, u cannot move low freq at real levels without drivers to do it, if you want low distortion, so it sounds like a real bass instrument. Man are you thick or what?
Well, actually, the driver should move in a linear fashion over its entire range of motion. Limiting the movement only decreases the frequency extention and the total SPL.
If you change the surface material of the driver, you'll change the distortion product. It might change at the frequency extremes of the driver or it might change at the limits of the drivers SPL range or the attack of the driver might be compromised along with the decay. But it will change since one component of the system has changed.
If you make the voice coil gap smaller or larger, you will change the distortion product by expotential numbers - think Lowther. But you will also change the ability of the driver to operate at high temperatures - high SPL over a period of time will heat the coil which makes it expand and constrict movement in an excessively narrow gap - unless you do some very fancy engineering - think Lowther.
There are numerous driver manufacturers using long throw gaps for their low frequency drivers. This design cuts down overall distortion numbers while typically adding other problems in other areas of performance.
As with everything I've ever encountered in audio, if the designer gives you one benefit, they are likely to compromise two others in the process. It is always a trade off of benefits vs. drawbacks. The designer chooses the mix that is preferred and you choose the designer's choices.
If lowering system distortion were all that simple and all that important, everyone would be doing the same thing and there would be no variation in speakers. The difference in cost and performance between an iron core inductor and an air gap inductor wouldn't matter. Cheap electrolytic capacitors would still be in use in the crossover.
Look at the similarities in designs and you'll see where the generally accepted methods of loudspeaker design exist. Then look at where the "outside the box" types such as Lowther go to find that litle bit extra. The kick with Lowther is they have been doing the same thing for the last 70 years. Quad has been doing more or less the same thing for almost fifty years.
You pays your money and you makes your choices. That's how hifi works as far as I can tell.
Of course, upd, you don't want to see anything that would not be in keeping with the decisions you've already made and which are unshakable in your opinion. So I won't go into the rest of the errors you made in that post, you don't want to get close to the truth any way.
Close to teh trutch, close to teh music....maybe you need an air gap, cus your thinking is constricted and sounds lacking in logic. Single drivers, yeah, I'm sure they are worked within limits of the drivers operation, no speaker maker of any merit would try to delude anyone that a single driver is full range, none. the ones who want to sell oddities to the unknowing, that is the market. Henry Kloss talked about single driver back in teh 50's, while at AR, KLH, then Advent further along, never could get anything that was close to hi fi, he thus went 2 way at least. Single drivers are not full range, don't matter how mcuh magic components, magic speak, magic materials in the driver, ain't happening. Your listening acuity lacks logic and any realm of wanting to hear LIVE music, cus' all the conversations with you are based on your listening to a single driver driven by a highly distorted amplifier. You needs to start getting into reality
The truth is the truth, upd. You don't own it, you don't even come close to having a good grasp on it and you won't read it because you don't want to get too close to it. So you stay stupid by choice.
I-IV-V is all you got. If you can only play every song in the key of Bb minor, you aren't very interesting and you can't play with a group. Yep, upd is a Bb minor kind o'guy.
Hi Dr. S,
Back in 1994, Martin Colloms did some distortion measurements of a Wilson Grand SLAMM in Stereophile here. These are the best distortion measurements I've ever seen for a speaker. I think most speaker manufacturers don't publish distortion measurements because they don't want you to know how bad they are. Publishing these measurements seems a bit more common in the pro audio world. For example, some distortion measurements of a JBL pro monitor, the LSR6332, can be found in this PDF file. Note that they scale the distortion up by 20 dB, I guess to make the graphs more readable. So the actual distortion is 20 dB less than what's shown.
Hi Andy C.,
thanks for catching the point. You
Can't have any standrds in hi fi audio, magic is first. The FTC did come out with power rating standrds years ago about pre conditioning 1/3 power etc, somehow though in high end magic ckts that is negated, since pre conditioning, testing to a standard is not what makes it sound better.......but when JA uses that test standard, it showed what a piece of crap some over priced Chord amp was when it was missing about 500 WATTS. But they decided what's 500 watts in a super pricey amp, there still are some more.....that's called mis representing the product, but in a pro application that would probably get the product dumped since a design calls for specs to be met....home hi fi, a world where what matters, is mostly made up BS. Or taht $27K obsolete Cd player with being miswired out of teh factory, who said QC matters? for a $27K Cd only player ...so so many episodes of hi fi nonsense..MF loved how out of phase sounds, which is why he might think a $100,000 TT is needed, but then he records onto a 10 cent CD and everyone says, wow, that musta come from that $100K TT which means digital is great if on a 10 cent medium you can capture all this supposed superiority of a $100,000 spinner.....
The general trend in "audiophile" speakers has for the last four decades been to lower the efficiency of the playback system. "Watts are cheap" has been the prevailing motto among all but the well informed listener who can esily hear the flaws of that concept.
Lower system efficiency has produced sound that literally struggles to get out of the box and arrives as a jangled mess when it finally makes it to the listening position. Not the sensitivity necessarily but the efficiency of the system including the amplifiers. In the Elliott article that upd so miserably misinterpreted, the point was made that raising the system efficiency of only the speakers will vastly improve the performance of the amplifier/speaker system.
Raising only the voltage sensivity of the system is not paying attention to the music. Why does a K'Horn have such low measured THD? In part because it operates at an incredibly high 10% system efficiency. Unfortunately, the K'Horn has other drawbacks to it's overall performance and obvious limitations to its wide spread use. But the Klipsch LaScala is a Stereophile recommended item and it uses the same drivers in a similar configuration to the K'Horn.
Most modern audiophile speaker systems operate at less than 2% efficiency. Even when the stated voltage sensitivity is close to 89dB@2.83volts, due to the manner in which sensitivity is measured the actual efficiency of the system can still be a low 1-2% if the designer has loaded the crossover with mulitple notch filters to produce a smoother on paper frequency response. The same problem occurs when the designer opts for multiway speaker systems with each division requiring numerous capacitors and inductors to make the roll out/in work properly. Along with padding resistors on the drivers, multiway speakers can represent multiple problems for an amplifier.
Simple is always better in this regard but it's not how the consumer audio industry has headed over the last few decades.
For all the things the LS3/5a does well, one of its most outstanding qualities is its amplifier friendly impedance/electrical phase angle swing. Short of a high impedance, efficient and well designed single driver, full range system, and despite its low voltage sensitivity, the 3/5a is one of the most friendly speakers an amplifier can ever cozy up to. It is exceptionally easy for any amplifier to drive. One of the ways the BBC designers went about this low sensitivity but amplifier friendly speaker was to insist on high quality drivers that were well matched and which didn't stray from the original spec coming from the factory. It made for a fair number of rejected drivers due to the inability of what became a mass market supplier to maintain spec, but it worked for the 3/5a.
If there were a spec that should be increasingly pointed out, I would say it is system efficiency. As a rule, when system efficiency rises, distortion at any SPL lowers. It doesn't matter if the driver being measured is 4" or 18", paper or magnesium, 84dB or 104dB, when the efficiency goes up, overall distortion - which is too broad to be relevant - goes down.
The efficiency measurement would point out what problems exist in many highly regarded and high priced speakers on today's market. Slamming high watts into such a speaker only to result in power compression is counterproductive as the Elliott article made clear (though upd never did figure that one out even when he quoted the article as saying so). IMO any speaker that requires an amplifier to struggle producing more than a few watts to sound good probably isn't worth the box it's shipped in. There are exceptions as there always will be but they are few and far between.
Did I read this correct? Jan believes in a measurement?
System efficiency? This rating would give a higher rating to a class D amplifier or a horn loaded speaker. This spec is like any other. Usefull when taken in context. Misleading when used alone.
BTW don't both speaker manufacturers and Stereophile list speaker efficiency?
Not that I can find. JA measures the voltage sensitivity of the system and reports the variance from the manufacturer's spec. Sensitivity and efficiency are not interchangeable.
You are somewhat misreading what I posted above. System efficiency rises when the baseline efficiency of the speaker system rises. Less power is required to drive the speaker to any desired SPL. At the present time that would somewhat argue against class A operation but class A is not that much of a factor in modern power amplifier design and is easily manageable in a pre amp. Class D and T would be high on the "let's explore this direction" next step. But simply moving away from speakers which require high current in order to hang on the driver would, IMO, be a step in the right direction. Speakers have not moved above that 10-12% efficiency rating in decades. Why?
I find it interesting that it has been stated by some authorities that distortion does not matter in speakers. This is nonsense in my opinion. An electrical or acoustical network is defined by it's small signal transfer function, and the large signal distortions. It is that simple. Speakers have the added characteristic of the transfer function being a function of observation point, but this is simply another dimension to the question.
I have to wonder if some of these "experts" are referring to the missing fundamental effect where distortion can help provide an "illusion" of the fundamental:
http://www.nationmaster.com/encyclopedia/Missing-fundamental
They may also be referring to the fact that speaker motor distortions tend to be more of a soft clipping or compression effect rather than hard limiting. We are much more tolerant of soft nonlinearities than hard. Low distortion is a significant goal in my speaker designs and it is easy to hear the difference when a clean reference is available. True fundamentals also tend to shake floors and walls, and hit you in the chest.
High Fidelity measured bass distortion vs. SPL; something I noted as a child in the early 1970s. IMF transmission lines had remarkably low distortion which sparked my interest to study them.
Don Keele did some distortion measurements in his reviews for Audio. I liked his pulsed tests where he measured max SPL output versus frequency. This primarily tests the displacement limited output of the system but is sometimes swamped by a saturating inductor, or other effects. Wilson speakers often did very well in these tests and this is probably one reason why they're liked by some mastering engineers.
There was an interesting AES paper where Gander of JBL measured a number of popular pro drivers and discussed how motor design impacts the distortion performance. He makes a point of how it is important to measure at constant SPL and not constant voltage input. Clearly a more sensitive driver producing more fundamental would be penalized with constant voltage input.
There are some tricks involved in optimizing a driver. The Advent woofer is an excellent design employing some tricks, while AR woofers tend to measure as having lower distortion.
Tell Klippel that distortion doesn't matter, LOL:
http://www.klippel.de/pubs/default.asp
Pete B.
Please explain. Yes the two have different units that are not interchangeable, but I believe that both specs are equally usefull.
Welcome Pete. How long have you been viewing this nonsense before you felt compelled to reply?
Just for fun I have to question the link to your engineering experience. Do you realize that you provide no analog background?
I don't think so. From all I know the major distortion prduct of most drivers will be harmonic in nature and low order in content with very little above third order. This would be considered fairly benign in most designer's viewpoint even when the numbers would suggest a terrible result. Therefore the Legacy advertisement can selectively pull a number for the B&W which doesn't inform the reader of the true nature of the speaker's performance. Is the measurement of the driver or from the output of the port?
IMD is low in most drivers, is it not? I would guess that designers whose speakers do not measure exceptionally low distortion consider speaker distortion to be of less importance than the designer who strives for low distortion. The problem with the latter type is the ability to close your ears to every other problem because you're searching for one item.
Who cares, other than the designer, if the distortion comes from the driver or the port? Both are outputs that contribute to the total output from the speaker.
I don't know because for some reason distortion measurments for speakers aren't well defined or always measured.
I would say the sensitivity spec is about as useful as the "zoom out" map on GoogleMaps. Knowing the voltage sensitivity of a speaker system, particularly a multiway speaker system, doesn't really tell you how difficult the speaker will be to drive. Two systems rated at similar sensitivity can play at substantially different levels with the same amplifier.
"Zoom in" to find the impedance curve and the electrical phase angle plot and then you have a better idea of what sort of amplifier will be required even without the sensitivity spec. My "reference" for this sort of speaker is the Thiel CS3.5. Remember the ads Thiel ran in Stereophile and TAS which showed a close up of their 22 element crossover? They were exceptionally proud of the number of caps and inductors in that crossover and suggested other manufacturers were cheating you out of parts you should be paying for if the other manufacturer didn't also include a board full of caps, resistors and inductors.
Well, of course, each and every one of those caps, resistors and inductors sucked the life-juice from the amplifier and made for a speaker with a not all that bad sensitivity spec but one that required a hefty two hundred watt minimum amplifier just to begin to get the speaker sounding decent. From a system efficiency viewpoint, the speaker was terrible. Not just in regard to the amplifier required but merely in terms of speaker efficiency vs. sensitivity. B&W has designed speakers similar to this for the last fifteen years.
It revolves around the "watts are cheap" attitude. A friend recently had to wait for a tube to come for his PrimaLuna amp and replaced the amp temporarily with his old 200 watt Sony. He's much happier now that the tube has arrived with the 40(?) watt PL driving his Quad 63's than with the Sony and for numerous reasons. Now the Quad happens to be one of my favorite speakers and it is not that easy to drive but a well designed amp of modest wattage can do quite well with that speaker. I wouldn't want to hear the PL driving those old Thiels or any of the recent B&W's. My LS3/5a's are rated at 82dB sensitivity but can do well with a 40 watt tube amplifier. I've heard a pretty good 120 watt amp on the old Thiels and it was pitiful, the speaker dipped down to a very low impedance point and the phase angle at that frequency was quite high. An amplifier without lots of available current simply couldn't make the speaker work well. Good amplifiers with lots of current on tap don't come cheap.
No, the voltage sensitivity does not, IMO, tell me much of anything about the speaker's ability to play well with an amplifier of modest cost and power output. If you wish to substitute the impedance/phase angle measurement for efficiency, that would be helpful but not well understood. And it will not really point up the fact that in far too many consumer speakers most of the input signal to the speaker gets lost as heat in the crossover and voice coil(s).
True, but depending on the tuning of the port, you might be hearing far higher output from the driver itself and much lower input from the port. Since the Legacy ad doesn't specify how the measurement was taken you don't know how the speaker produces 66% distortion but still is well regarded. I'll ask you the same question I asked upd but he ignored. How do you explain the fact that, if the Legacy ad is correct, a speaker that produces 66% distortion at low "E" would not be noticed by a decent reviewer or JA when he takes measurements?
Thanks for the welcome. I read on and off here and realize that it is just casual conversation for some.
There is simply more demand for digital design. I went to school for analog design. Senior project involved FM discriminator design. Additional senior project involved the development of a Thiele and Small type model for transmission line speakers. My graduate project involved the design of an analog adaptive filter and frequency shifter circuitry - all analog. I've designed reconstruction filters for video from simple 3rd order to 7th order delay equalized that provide broadcast quality video as validated by a VM-700. Been doing crossover design for more than 20 years. Designed tracking OP Amp based filters.
Built audio power amps as a kid, started with tubes as a child - those didn't work, then transistor amps from scratch that did work. Germanium transistor based amps in between that only worked for a few minutes.
Lots of engineers on the job know digital design but fear analog. I was usually given the analog design tasks in most of our design groups. I was also given the "hard" problems, LOL. I designed the pipeline control logic for an IBM 370 compatible mini computer. The experts said it could not be done without an expensive ($50K to $100K) ASIC but I proved them wrong. Larger companies have RF and power supply specialists who handled those analog designs. The difficult problems in digital design often involve analog issues - transmission line effects, cross talk, ground bounce, etc. I often taught digital designers how to debug and address these issues that are often a mystery to inexperienced people.
Pete B.
Speakers exhibit all forms of distortion, more than most amplifiers. The difference is that the early onset is soft, then as the voice coil bottoms it is hard and similar to clipping. But there are many other sources, saturation in the pole piece, thermal compression, DC offset. Speakers are a mess, luckily our ears are somewhat forgiving.
Check out Zaph's distortion tests, there's plenty of 3rd-5th with most drivers and he's not pushing them very hard:
http://www.zaphaudio.com/6.5test/
Pete B.
Conjugate load matching circuits in speakers add significantly to the number of components and they make the system easier to drive. Your logic is flawed.
Pete B.
Niether spec will provide the information as to a speakers complex load.
Actually two speakers with similar sensitivities will have similar output levels becuase sensitivity ratings are based sensitivity specs are based on voltage inputs. Audio amps are pretty close to being a constant voltage source.
An efficiency rating is based on power in vs power out.
Using a constant voltage source driving different impedence loads will provide an unfair advantage to the speaker with the highest impedence when using efficiency as a measure.
"Zoom in" to find the impedance curve and the electrical phase angle plot and then you have a better idea of what sort of amplifier will be required even without the sensitivity spec. My "reference" for this sort of speaker is the Thiel CS3.5. Remember the ads Thiel ran in Stereophile and TAS which showed a close up of their 22 element crossover? They were exceptionally proud of the number of caps and inductors in that crossover and suggested other manufacturers were cheating you out of parts you should be paying for if the other manufacturer didn't also include a board full of caps, resistors and inductors.
Well, of course, each and every one of those caps, resistors and inductors sucked the life-juice from the amplifier and made for a speaker with a not all that bad sensitivity spec but one that required a hefty two hundred watt minimum amplifier just to begin to get the speaker sounding decent. From a system efficiency viewpoint, the speaker was terrible. Not just in regard to the amplifier required but merely in terms of speaker efficiency vs. sensitivity. B&W has designed speakers similar to this for the last fifteen years.
It revolves around the "watts are cheap" attitude. A friend recently had to wait for a tube to come for his PrimaLuna amp and replaced the amp temporarily with his old 200 watt Sony. He's much happier now that the tube has arrived with the 40(?) watt PL driving his Quad 63's than with the Sony and for numerous reasons. Now the Quad happens to be one of my favorite speakers and it is not that easy to drive but a well designed amp of modest wattage can do quite well with that speaker. I wouldn't want to hear the PL driving those old Thiels or any of the recent B&W's. My LS3/5a's are rated at 82dB sensitivity but can do well with a 40 watt tube amplifier. I've heard a pretty good 120 watt amp on the old Thiels and it was pitiful, the speaker dipped down to a very low impedance point and the phase angle at that frequency was quite high. An amplifier without lots of available current simply couldn't make the speaker work well. Good amplifiers with lots of current on tap don't come cheap.
No, the voltage sensitivity does not, IMO, tell me much of anything about the speaker's ability to play well with an amplifier of modest cost and power output. If you wish to substitute the impedance/phase angle measurement for efficiency, that would be helpful but not well understood. And it will not really point up the fact that in far too many consumer speakers most of the input signal to the speaker gets lost as heat in the crossover and voice coil(s).
Here is a page where surplus NHT drivers are for sale and the factory Klippel driver data is provided:
http://home.comcast.net/~jhidley/
There is extensive nonlinear data there because designers *do* take distortion into consideration. The fact is that they don't want consumers to worry or know about it.
In the same way that they want them to believe in burn in to get over buyers remorse. This was posted by Ken Kantor:
http://www.classicspeakerpages.net/IP.Bo...ost&p=52320
IIRC, Velodyne tried a marketing approach regarding low distortion mesurements when they came out with their full range speaker.
It got relatively poor reviews, and did not appear to linger long in the marketplace.
I'll go see if I can find any of their literature or some old reviews...
Couldn't find a Velodyne review in the Stereophile way-backs. Perhaps I'm thinking of TAS.
Anyway, in 1998, JA published a multi-part article about speaker measurement that I haven't had time to go back and re-read yet, but he did mention "non-linear distortions."
Perhaps he will be able to comment.
Jan, Legacy mentioned one aspect of the speaker's characteristics, relating to one tiny teeney ittsy bittsy part of the complete system, involving lotsa details. If you want the entire specs and design parameters, read more on his website, and ask him questions, in his question section, you will become an educated consumer once Legacy's Bill D. sets you straight about loudspeakers. Why not mention that you think a single driver 5" midrange doing full spectrum things is better than his overly capacitated designs, using all those drivers to move air with low distortion. He will set you straight.....you will soon relaize you don't know what you think you do. And I would never say you don't belong here, since eventually when you get your head out of your confused state and get edumacated in speakers. Loudspeakers without crossovers and drivers are not a system at all, they are speakers in a box with no function. What are the SPECS on that single driver tha is trying to reproduce a 49Hz note and a 15600 one...I bet the spec sheets on that driver will make you wonder...wait a minute....it only goes to 8K Hz........
Measuring Loudspeakers
RG
I recall a Velodyne full range review in Audio but that was many years ago. As I recall they sacrificed too many other parameters in order to obtain such low distortion, but I don't recall the details. Let me see if I can find the review in my library.
Pete B.
Why not leave what I use out of this for once and forever?
My reservations are with the manner in which the Legacy advertisement flirts with misinformation. 100 watts/100dB/no information on their own speaker. I expect that sort of advertising from consumer level Technics or some no name company selling at a flea market. If a manufactuer wants my business, they have to earn it. The type of ad you showed us from Legacy does not make me want to talk to the Legacy designer.
You seem not to be bothered by companies that talk down their competition. I have a problem with that. Either you can sell me your product on its own merits without telling me the other guy's stuff stinks, or you won't sell me your product at all. Either your product stands on its own merits or it doesn't stand at all. That's how I conducted myself on the salesfloor and its what I'm looking for from a company I can respect just as much as they approach the sale with respect for my intelligence as a client.
So, from here on out, upd, what I use in my system has no bearing on any of your remarks and I would appreciate you not mentioning my system unless it is somehow relevant to the discussion. Your constant "5 inch driver/can't be cutting it" and "Radio Shack table radio" comments cannot be relevant to every post you make.
We are not discussing my speakers. Got it?
You bang your head against the wall demeaning what others use in every post you place on this forum. That alone is reason enough for you to not exist on a disussion forum. You are the epitomy of rudeness. If you cannot respect the other members of the forum enough to not attack their choice of components - as you have done with my system, remember when you started the thread on my amplifier that resulted in your first banishment? - then you do not belong on a discussion forum.
If, as Elk has pointed out on the "capacitors" thread, you continue to drive the discussion into the ground for your own amusement while willfully ignoring facts and deliberately avoiding questions and discussion of the topic, then you do not belong on a discussion forum. The forum does not need a class clown who is intent on keeping others as uninformed as he is.
That's not that difficult to understand, upd.
When you want to actually become a part of the discussion you will be welcomed as a contributing member by everyone on this forum and not just a small group of sychophants who find your disruptions amusing "when you look deeper into their meaning". If all you can do is say no to everything that doesn't meet your approval and call other members fools and disparage their choice in equipment, then you have no place on a discussion forum.
It's your choice, upd. No one can make you discuss anything, you'll have to want to do so rationally. If you choose not to do so, then, please, allow the rest of us to have a discussion without a loud child disrupting our thoughts.
How many times has this same message been posted for you, upd? This is at least the third time you and I have reached this point.
Then please educate me. I am not a designer nor an engineer so what I know is what I believe I've been taught.
Here is what I believe to be true.
The issue at hand with any passive crossover is the conjugate reactance of the circuit, a reactance that is largely determined by the capacitors and inductors within the crossover circuit and it's connection type.
Given a single driver, the reactance of the load is fairly benign and easily predictable with frequency when compared to a multi-element passive crossover connected in parallel with the drivers (the most common connection type in consumer speakers).
The difference to the amplifier between a parallel connected and a series connected passive crossover is the degree to which the impedance swing and the concurrent phase angle of the circuit will require either higher voltage or higher current draw from the amplifier's power supply.
With the exception of a few of the Pass amplifiers, all consumer amplifiers operate essentially as voltage sources and therefore are more well suited to a load which requires voltage delivery rather than high current delivery. Such a load is high in nominal impedance and benign in its phase angle. This is not the trend in many contemporary speakers sold to the audiophile market.
Any speaker system with reactive components (discounting the driver's voice coil itself) added to the signal path will present a widely varying load to the amplifier due in part to the reactance of the driver's vc. If the impedance of the load is high, the amplifier can easily remain stable and operate efficiently even if the degree of phase angle shift is relatively high at the same frequency. The same is true of a low impedance load if the degree of phase shift is relatively minimal at the same frequency. It is when the impedance drops and the phase angle shift is large at the same frequency that an amplifier will be asked to deliver maximum current. This is the trend in many contemporary speakers sold to the audiophile community.
Therefore, a speaker with a high component count in its passive crossover is in danger of treading close to the edge of being a "difficult load". I know just having numerous elements in the crossover does not ensure a difficult load, the LS3/5a employed a rather high component count crossover but it is regarded as an "amplifier friendly" speaker since it's nominal impedance remained high and its phase angle was never in opposition to its impedance. However, as employed by Thiel in the CS3.5, with its numerous notch filters to correct frequency response, the crossover is what contributed to the speaker system's low impedance and high phase angle. It was not the three drivers alone which pushed the speaker into the difficult to drive category. If that were true, wouldn't any three or four way speaker with a passive crossover be an equally difficult load? That's not the case in my experience.
As a side note, Thiel utilizes first order filters exclusively in their speakers. This would imply a fairly low parts count when compared to a speaker with a higher order filter. My understanding of the matter is it is the inclusion of numerous notch filters in the Thiel crossover which contribute to its high component count.
Since no consumer speaker today reaches above 10-12% system efficiency and the average is 1-2%, isn't it fair to say each element added to the passive crossover represents one more point where signal voltage is converted and lost to heat dissipation? What else, other than the inefficiency of the driver itself, accounts for the astonishingly low efficiency of today's speaker market compared to the days when tubes, with their inherently low current delivery, worked into high impedance speaker loads with an average efficiency no lower than what we see today as our highest efficiency systems? With that in mind, wouldn't higher system efficiencies be desirable and why haven't they been sought out?
I am in firmly the camp of the fewest parts required to do the task sucessfully is always better than more parts that only solve problems created by more parts.
If, then, the higher parts count results in higher heat losses and a more difficult load for the amplifier to drive, please explain how this would make for an easier to drive system.
I would disagree based on what I have heard and read. First, not all speakers live up to their sensitivity spec as JA points out in his measurements. Some can miss by a margin large enough to indicate a doubling of suggested amplifier power in order to reach a desired SPL.
Second, they may play relatively close in output level when fed a test signal of restriced frequency range but they might not sound even relatively close to one another with a music source.
Finally, the sensitivity spec/measurement does not indicate how difficult the speaker will be to drive and therefore can be misleading when a lower priced speaker requires a significantly more expensive amplifier to perform up to expectations. I assume from JA's "How we measure loudspeakers" this is why JA and the old Audio magazine routinely measure(d) for impedance against electrical phase angle. This indicates overall system efficiency.
I fail to see how a less difficult to drive load can be considered an "unfair" advantage. Aren't we after all the advantages we can find? Why put ourself with one and one half feet on the slippery slope to a deep hole before we even begin? It wasn't all that long ago that 16 Ohm speakers were common. Today, the average would probably be less than 6 Ohms nominal for the audiophile market. That drift toward more difficult loads alone would make for a dramatic shift in the amplifier required to drive such a load which would result in lower overall system efficiency. The "watts are cheap" marketing at work.
Maybe I'm not at all understanding how to interpret these graphs; http://www.zaphaudio.com/6.5test/compare.html. As I click on F2/F3/F4/F5, the information changes and I am not certain what I should be taking away from that change. However, what I tend to see in virtually all of the graphs is a F4 & F5 which exist sometimes as much a -50dB beneath F2 which itself is relatively low compared to F1. Not inconsequential but also not at the point where anyone would expect to see a driver alone producing 66% THD. Ignoring the occasional spike in response, the nominal level of F5 tends to reside at -100db or lower with F1 at +20dB.
Is that correct?
If so, then would that not support my thinking that higher order harmonic components are fairly well suppressed in most modern drivers - at least for those represented in this series of tests?
Sorry to say I didn't find anything useful to this discussion on these pages; http://home.comcast.net/~jhidley/Klippel_reports/11-047-1_LSI_Detailed.htm
What's there that I'm supposed to notice?
Cus that's YOUR reference, ain't it. You keep insisting how a single driver crossoverless thing is superior to a full range system with enough drivers to move air with no distortion. your point of reference is therefore extremely low, and answer teh question what live music is your point of view coming from, at least once a week I get some LIVE music, not relying on long term memory, cus' the entire reason is LIVE at home, not knowing what live is, you thus think 5" works...5" never works, can never ever deliver full range dynamics, at live levels, impossible, and not having a crossover is hardly a benefit, ain't it brilliant to send a signal to a driver it can't handle, now that brilliant design, what kind of nudnick puts a single driver and feeds it teh entire freq range, crossovers where developed so individual drivers operate in their optimum range, cus' true designers know no one driver can do it all....none, ever. clock radioa are not hi fi Have you heard a live drum or guitar recently, you telling me a complete drum set coming out of a 5" midrange, sounds like a drum set? Where on planet midget? Where everything is small lifeless and tiny? Don't you also use skinny wire cus' that sounds better than wire that can handle the current flow? Bizzaro world for sure.....
Pages