Amen, Clueless. Why is it again you say you generally don't desire to post at length much anymore? :-) |
Clueless and Twl, I have printed and read through the De Lima paper. Read through it a few times, in fact. You are right, there are actual measurement graphs in there, similar to the types I called for above, that would seem to demonstrate the phenomenon he advocates and that you are talking about.
The paper itself, while stipulating many worthwhile points that I could agree with, did not impress me overall. It had a few inconsistencies that made me question De Lima's rigorousness. I listed these on a piece of paper so I could post about them, but I couldn't get those graphs out of my head. I was reasonably sure that my arguments above were sound, yet here was some empirical evidence to the contrary staring me in the face.
This bothered me all night, and I did not go to sleep, instead turning it over in my head. Finally in the wee hours I quit and decided to take a shower. I should have known - I always do my best thinking (and sometimes songwriting) in the shower. :-)
There I had an epiphany, and I now believe that I can debunk De Lima's paper (or at least seriously call it into question), graphs included. I am not going to post my analysis and critique right now, both because it needs to be a detailed post and I am tired and want to go to sleep, and because I want a little more time to recheck myself in the light of day and make sure I'm not mistaken or being rash - not only because I could always be wrong - but also because if I am right, what I say might carry a bit of reputational consequences.
In the meantime, I encourage anybody reading to use Clueless' link to the article, and consider these questions for themselves. What I am thinking of is not so advanced or anything, and it's quite possible (if I am correct in my assessment) that someone with more technical knowledge than myself would see the flaw I believe I have found in a much quicker time than I did. I will give this hint: the fatal problem area of his argument as I see it (one which I have already generally suggested in my posts above) is given away by a careful logical analysis of what is contained and implied in his graphs and their associated explanatory paragraphs (further hint: you must focus on what he doesn't say as well as what he does). I shall return. |
I've emailed De Lima with a couple of basic questions about his published graphs, asking for a little information which is not provided in his article or graph labels (but should be), so that I don't have to work from any assumptions if I don't have to. I'll continue with my intended post after I (hopefully) get a reply, or amend my intended critique if the answers I get fundamentally change my assumptions about what he purports to have shown. Just want to be on as firm a footing as I can here. |
For those of us who enjoy listening at high volumes, but want a relatively compact, easy to place speaker, I'm not sure there is any choice other than to go dynamic drivers/higher power, and ditto for those who prefer panel speakers. In fact, that's basically my whole concern about low power: it's not the simple, flea-powered amps that bother me in theory, it's how you're limited in the speaker department. I've only ever heard some SET systems playing in a store in passing, and haven't ever really auditioned one. But I have a hard time believing that - outside of large horns (which may have other problems, but what doesn't?) - high-SPL, wide-bandwidth, uncolored, non-resonant performance will be available from an easy-to-drive and easy-to-control ultra-high-efficiency speaker. The laws of physics are against it happening, so I assume that folks happy with their SET systems must be giving up some combination of volume, headroom, extension, neutrality, or clean decay behavior to get their special kicks (some of which are always subject to the suspicion of increased, but pleasing, low-order distortion-induced euphony). I've not properly auditioned Avant-Gardes or the like, and don't have anything against increased efficiency where possible, but it otherwise seems to me that the world's most competent speakers still demand something in the way of clean power to sound most like live music, and I for one am not ready to simply view this as a wattage 'rat-race'. I would love to be proved wrong, and have someone play me a bandwidth- and volume-unlimited SET system that sounds accurate while kicking out the live jams at a handful of watts, but I don't know that I'll ever have the opportunity (or that the speakers needed would fit my house/budget), so I'll stay with what I know works reasonably well. |
Twl, I don't consider the word "euphonic" to signify a bad thing qualitatively; only if such a quality is actually masking higher levels of distortion so as to sway a listener into preferring a version of reality which is farther from the truth does it raise my eyebrows. Otherwise, inevetible distortions should be both low in level and low in order, ideally speaking.
Yes, I've heard about the theory of canceling distortions in SET/single driver systems, but I don't buy it. There is just no way that I can see where, even if you stipulate identical-order and -level distortions (which I strongly doubt can ever be the case, but as you say the theory doesn't depend upon 'identical' to work to some degree), the temporal relationships and the inherent non-linear components will never match up - in whatever phase - to yield much cancellation.
In fact, I don't know why there should really be any cancellation at all: this reversing the leads business is flawed thinking to my mind. The speaker's distortions are *caused* (for the most part) by its response to the input signal. Reversing the phase of that signal will merely do the same for those distortions. (And not even considering for the time being an amp's ability to exert its control over the driver's movement, or whether the reverse occurs.)
Obviously, this isn't a unique indictment of single-driver speakers; any speaker, fed by any type of amp, will receive the amp's output signal - including its embedded distortions - and then further distort that whole signal, amp distortions included. The cancellation argument, were it true, should apply in some measure to any amp/speaker combination; the position that it will only work with low-order distortions makes no sense to me. Besides, SET amps have higher-order distortion products too. They just don't internally cancel the even orders, so the distortion profile tends to be characteristically different than push-pull amps. But it is not exclusively low-order - it is predominantly low-order.
Don't misunderstand me - the combination of class-A operation, simple and few parts and circuits, low or no feedback, and no crossover notch artifacts or device mis-matching or -tracking makes SET's appealing in theory even despite low power and higher total THD levels. Getting a speaker to give of its best under those restrictions can be a different matter, as I'm sure you'd agree.
But rationalizations like the cancellation argument just make me more skeptical than I might be otherwise. Simply the fact that apparently only the SET crowd feels the need to come up with something like this rhetoric points the finger of doubt, to me. The whole case is too facile and unsupported, as far as I can see.
Look at the contention that single-driver speakers will have lower-order distortion, for instance. How so? I have always thought that the wider the bandwidth any one driver is called upon to cover, the higher the THD over its bandwidth that will result. This makes sense, and so does the limited total bandwidth of a speaker which depends on just one driver (dynamic driver, that is). True, other distortions, such as phase distortions and crossover-induced distortions, can be lower. But if you restrict any single driver to operating only within its prime comfort zone, as in a multi-way speaker, THD (as well as non-linearities) must be lower at all orders.
I am not suggesting that higher-power, push-pull amps, of either class-A or -A/B operation, don't have their own characteristic sonic flaws, some of which are unique to their operation and shouldn't be overlooked in the quest for higher power, others of which are merely unmasked by the same (and not unmasked in SET's, presumably due to other, higher-level 'euphonic' distortions). But I don't know that anybody can claim that either type of amp, alone, sounds 'better' than the other, because the partnered speakers tend to be so different.
In that sense, SET owners will tend to have the advantage concerning comparisions, since they could reasonably experiment with higher-powered push-pull amps on their efficient speakers and draw some conclusions, whereas conventional low-impedance, low-efficiency speaker owners (like me) cannot really do the reverse comparision. But the results of such comparisions will not speak directly to the question of whether an SET/single-driver or horn combo will 'outperform' a higher-power/low-to-moderate-efficiency pairing. It almost certainly would depend in large part to the listener's priorities, so all we can definitively say is that all various amp/speaker pairings will sound different from one another.
I am also not saying that the very idea of single-driver speakers doesn't have powerful attractions. Consistent dispersion at all angles, uniform resonance properties, phase coherence, and lack of crossover-induced complications and limitations are all desirable. The trade-offs we already know. If I lived near you, I'd ask to come over and have a listen. I'd even bring my amps.
As things are, all I can do is use my mind the best I can to consider the arguments. I find the cancellation argument implausible, and more than a little ingenuous in its seemingly willful selectivity and dubious premises. I'm not on this thread to make anyone defend their system to me, just to give my own view that power can be used for good as well as evil, particularly in freeing up speaker design possibilities. I have no agenda or animus, just intellectual curiousity and a fondness for playing devil's advocate. |
Yes, Thomas is very passionate, or at least much more so than me when it comes to gear and audio. As to his less humble rhetoric about who among us is a real music enthusiast, most of the real music lovers I've known weren't even audiophiles at all, and had low-fi to mid-fi systems (which isn't to say they're somehow better than we can be - just that there's no absolute connection between the two). I am much more circumspect about not only the possible virtues of my system (it never fools me), but about the possible pros or cons of any particular system approach. But then again, Twl has more experience than me.
The rant about specsmanship is not wholly unjustified, but it's also increasingly not necessary, I think (and especially not in the context of what's been written in this thread, but as Twl says, he wasn't directing it at my arguments per se). At least within the friendly confines of this forum, I think it has somewhat the quality of a straw man. His system is very unusual anyway: not only doesn't he have any output transformer on his tubes, but the whole thing is battery-powered, a virtue which is probably its own best argument for not being able to dump power like a turbine.
I am not in agreement about the term "euphonic" as applied to THD products being a misnomer. I do believe that moderate amounts of low-order harmonic distortion (probably up to a few percent at least, particularly in the lower frequency ranges) can sound pleasing. I also believe this is not true of higher-order products (at least as it applies to stereo systems, as opposed to electronic music-making), but that is the conventional wisdom. I am a tube user myself, and I have often wondered if my preference for this technology is in large part a reflection of a preference for the harmonic 'tube signature' above the harmonic 'solid-state signature' - or maybe even above no detectable signature at all, given the much lower measured distortion levels SS gear can provide.
Oops, there I went talking about measurements. I agree in many ways with Twl that measurements have often obscured due consideration of what our ears hear. I don't think this is because all measurements are worthless, but just because we apparently can't find ways to measure possible unidentified mechanisms that seem to affect what we hear. In the past, it's also been because many folks who fetishize measurements simply haven't concerned themselves with what listeners hear, a fatal mistake from a purely scientific point of view.
But I have a problem with the internal inconsistency of both refuting the supremacy of technically measurable qualities on the one hand, combined with the advancement of this theory of mutual THD cancellation occurring between a single-driver speaker and an SET amp on the other hand. First of all, let me acknowledge that while Twl advances this argument here, he is not responsible for having created it, just promulgating it. I also want to state that Thomas is, in my opinion, more technically competent than myself. But of course those things aren't enough to prevent me from attempting to take him to task a bit on the matter. :-)
Aside from this theory's (to me unmistakable) specious, lift-yourself-up-by-your-own-bootstraps quality, my problem with the argument is mainly one of selectivity in the service of rationalization. I have no qualms with someone taking the position that they will choose to trust their ears first, and if some apparent conflict arises between what they hear and what can be meausured, they will disregard the supposed implications of the measurement as being faulty, incomplete, or irrelevant.
But then you see something like the distortion-cancellation proposal, which reminds me of nothing so much as 'creation science', wherein a belief (creationism) that presumably doesn't depend on observation of the real world for its support gets shaky knees in light of what can be scientifically supported, and cobbles together its own psuedo-scientific 'theory' as a counteractant, despite the predictable fact that it cannot be confirmed by the evidence. That analogy is not meant to imply that preferring the sound of SET's is comparable to a belief system or denying the real world, just that I find it telling when any movement first contests the methods of its opposition based on their having reached different conclusions, but then attempts to put forth a spurious bastardization of those methods as an additional prop for their position.
To me, it's got to be one or the other: either the measurements don't ultimately matter (for potentially valid reasons) and one does not require a plausible (as in confirmable) technical explanation for what cannot be understood in technical terms, or we have to affirm that there will be some connection between what is discoverable by our ears and what is discoverable through testable technical hypothoses.
In other words, if one is going to so eloquently take exception to the applicability of those technical arguments which are testable (as Twl has), then don't try to simultaneously propone one which has not been confirmed in support of your preference. I think a back-to-back rereading of Twl's first argument in favor of the alleged phenomenon of distortion-cancellation and his second argument in disfavor of reliance on technically known and observable phenomena makes the contradiction clear. Simply proposing a technical mechanism which is heretofore unknown does not reconcile the two positions.
A prime reason the distortion-cancellation argument is quite likely just too beautiful to be true is its very quality of not having been tested, because it could have been by now. This theory might well be easy enough to partially confirm or deny through normal measurements taken of the output vs. input from an SET/single-driver combo, comparing it to the same speaker driven by an amp with, say, push-pull SS topology, and also doing the same comparsions using an appropriate multi-way speaker. Somebody please email me when this is done by an SET-supporter showing that the theory proves to be correct (somehow I doubt they'll be rushing to test it).
But while I'm not trusting this theory or the motives behind offering it, neither am I saying that the appeal of SET/high-sensitivity speaker combos is going to be due to their distortion signatures. I am not adverse to giving preliminary credibility to Twl's contention that distortion (or at least some kinds of distortion) might in reality be lower for such combos - at least within limited frequency and level parameters.
My bringing up the matter of extension was not, as assumed by Twl, primarily about the low end of single-driver speakers. Even such conventional audiophiles as John Atkinson have demonstrated long-standing biases toward basically dispensing with the pursuit of non-rolled response in the bottom octave (maybe octave-and-a-half) in their personal systems - in spite of their high-powered amps and multi-way speakers - simply because of the difficulties and compromises involved in trying to do it right. My point was also not focused on any possible extension limitations of SET amps themselves, which can in theory be overcome if the speaker doesn't demand too much in the way of low-end current.
My point had more to do with the HF limitations of a single-driver design (at least with a pistonic dynamic driver; traveling wave designs [Ohm Walsh] or electrostat's are different). Obviously, single-driver operation is not a prerequisite for SET use, but I brought it up partly because of the distortion-cancellation theory's stipulation of an un-crossed-over single-driver design. Twl of course must not be bothered by it in his system, but I can't see any way for a relatively large and massive single driver to equal the HF response and dispersion of tweeters in multi-way designs.
I was also not refering (again as assumed) to intermodulation distortion when I said the THD of a single driver must be higher. I actually remain unconvinced about the legitimacy of supposed deleterious effects from Doppler distortions occurring in wide-bandwidth drivers, but wasn't talking about that anyway. I said what I meant - that THD must go up as a driver is called upon to handle a very wide bandwidth. If the driver is naturally best suited for the lows, THD must be greater in the mids and highs. If optimized for the mids, then in the lows and highs. If the driver could pristinely handle the highs, then distortion would rise rapidly for the mids and lows. I believe that similar tradeoffs must exist for macro- vs. micro-dynamic capabilities as well. The very same concerns, only to a quite lesser degree, apply in first-order multi-way designs like my own Thiels. Again, all these limitations can be looked at as trade-offs, with other compensating inherent virtues perhaps present.
My stressing of volume capability and dynamic headroom is due to my belief that on the whole, audiophiles place too little emphasis on the quality of amplitudinal fidelity. To me, a lack of purity is no more deadly when it comes to destroying the illusion of live than is too-quiet playback (or compressed peaks). Unfortunately, our rooms tend to be the most limiting factor in being able to achieve higher levels with comfort, and I will never consider myself to be a genuine pursuer of the audiophile truth until the time when (or if) I make a comprehensively acoustically designed and treated listening room. Next to the room factor, all this hoo-ha about what type of amp pales in comparision anyway.
I say this because I have had the experience of listening to my own band's playing and singing reproduced for me in such a scientifically designed and implemented space, and over innocuous little self-powered mini-monitors costing less than the average high end MC carts of today, with the audio signal being routed through a mixing board and hundreds of feet of utilitarian cable, I heard a tangible illusion of live reality that put to shame anything I've ever heard through a high end system in a normal (or even a treated normal) listening room or showroom. I can only imagine what a carefully set up high end system would sound like in that room. Talk about not knowing what can be achieved with an audio system - my head kept jerking around in involuntary surprise when the music started, and it was all I could do to keep from pointing 'at' the sound in disbelief like some kind of delerious fool. I could only try to hold myself in my seat thoughout the whole process and not smile too conspicuously (and we didn't even employ purist recording techniques, although it's true that what I'm talking about was the sound of the original master tape). When I play back these recordings on my system, some tonal qualities and such are more complete, but I am no longer in the band's presence, musically or emotionally. Just wanted to put a little perspective on things before I go. :-) |
Twl, I appreciate your position regardless of whether or not I personally buy into De Lima's theory. Thanks for taking the time to explicate your thoughts in depth. As you probably know from our past discussions, I have always been curious to get a good audition of a system along the lines of what you promote, but haven't done it so far, and admit to not having made any great efforts to do so (then again, the same thing applies to me with any audio gear - I stay away from the shows completely and the shops as much as I can, and none of my friends are audiophiles, so there you have it...).
As far as the one technical point you raise in the above responses subsequent to my last post: Although I am willing for the time being to take your word about the HF capabilities of the Lowther - or at least its perceived performance (being that un-rolled, wide-dispersion response above a certain point is of debatable necessity) - I don't really believe that this driver could be so unique in terms of its moving mass. I mean, designers and manufacturers of dynamic cone drivers have been expending great effort on reducing driver mass while increasing stiffness (something the Lowther must possess if it is to cover the frequency range it is called upon to do) through applied research and technology for many decades now, the whole world over. I will do some investigation online of this driver, because the technology presumably required for its intended application intrigues me, but will say for the time being that I find it difficult to accept the proposition that it could be so very much lower in mass than many other manufacturers' designs (even if those designs aren't intended to be used full-range).
As far as the electrostat reference we have both made goes, yes, these are much lower in moving mass per area of driver surface than any dynamic design, even a dome tweeter. This is possible because the diaphragm is uniformly driven over its entire surface, so rigidity is not needed, and HF response is not impeded by high moving mass. (I also want to mention that my reservations about the validity of the Doppler-intermodulation theoretical critique of wide-range drivers has as much or more to do with the fact that the eardrum itself is a single-membrane transducer, as it does with microphone physics - I'm just not convinced that the ear/brain is in fact sensitive to this supposed problem as it applies to driver operation.)
Clueless, I am sorry if any of my comments may have offended you, and thanks for the link. They were not intended as an attack, so much as that I personally hate to see what is quite likely a very valid movement in terms of not only providing people with another listening option, but also in serving as a useful check on - and critique of - the status quo, feeling that it must engage in what I perceive to be a black-magic, slight-of-hand pseudo-technical argument in order to bolster its case.
I of course realize that disortion cancellation is an accepted phenomenon in certain circuit designs, but I do not feel there is a direct analogy between balanced/opposing-phase circuits or lines and the SET/single-driver theory. I will read the article you have linked (although I may have read it before, I'm not certain - I know I have read about this in the past, but I can't recall whether it was De Lima himself or another writer[s] citing his [or the same] arguments). In general, I feel it is very easy for a technically literate writer who is so inclined to baffle the masses with speculative theories if they choose. My response is, show me some evidence. I may not be an electrical or acoustical engineer, but I find I can usually rely on my own BS detector to weed out the spurious arguments. In audiophilia, technical competence has never been a guarantee against quackery, even among nice guys. :-)
Again, I'm not questioning anyone's beliefs, to which they are entitled. To me, it's not about belief - if something is plausible, and its effect demonstrable, then the mechanism should be verifiable. Since no one seems to have verified this mechanistic belief, I will go with my instinct that it is not plausible. The proposition that SET/single-driver systems may be the most 'pure', on the other hand, I do not reject on its face at all; I just feel that other, and fairly obvious and well-known, factors could account for that being so, if true. If something in his paper gives me a different perspective, I will post about it here.
At the end of the day, everything we have been talking about here just comes back to the hard reality of the 'audiophile condition', if you will - the fact that there is no, and can never be any, 'perfect' speaker or amplifier design. In the real world, everything is a design trade-off dictated by the laws of physics, and I think we all acknowledge that. It's simply left to us try and pick our poisons based on our experiences and preferences. My own inclinations are pretty conventional: low coloration (even frequency response), wide bandwidth, wide and even dispersion, comfortable headroom, low distortion, accurate transient response with low overhang, etc. I do not lightly dismiss Twl's, or any other careful listener's, positions in favor of simpler systems. I do not know if his or their preferences are a good match for my own, but I do know that they are sincere, and wouldn't be listening to plainly deficient sound reproduction by anyone's standards. I regard myself as still of an open mind, not only because their advocacy leads me to believe that there must be something to it, but mostly because I am cognizant of the reality of the paradigm of trade-offs inherent in any particular engineering approach stipulated above, and that must include the approaches my own system employs. |
Patrick, is there a URL with a picture posted of your baby (the car, I mean)? And have you ever seen the admittedly obscure 1974 documentary film "Funny Car Summer" (if not, you can look it up on imdb.com)?
Macrojack, everyone prefers gas over electric, but I have to say that I do have a 'soft spot' for microwaves. Maybe that comes from standing too close to one for many years. ;^)
FWIW, no response yet from De Lima. And happy birthday Django! |
Chevy II, '66 or '67 I'm guessing, 496 I'm told, plain white and very stealthy-looking, and if that's your garage interior in the photo, one hell of a neat'n'clean well-outfitted shop...
Still no De Lima, I may have to proceed on some assumptions without him, but will give it through tomorrow... |
Yeah Patrick, I was wondering about that number, whether you'd bored it or what, but I guess either you typed a 9 instead of a 0 or my eyes are going - don't know, 'cause the email's been deleted now! :-)
Twl, I'm actually starting to become glad I've delayed now - some more time to try and digest the implications of the technical appendix at the bottom of the article (which is frankly over my head) has got me beginning to think that maybe the particular inconsistency or incompleteness I thought I detected in the body of the article is answered or addressed in the appendix. I don't know yet, I'm trying to do further research to educate myself some more, but I'm holding off on any accusations of a "smoking gun" for the moment. I haven't gotten a reply from De Lima yet; he's the one I'd really like to chat with right about now, and I don't think I'm gonna turn myself into an EE overnight. :-) But I will say that further consideration has me thinking that one of my assumptions is probably incorrect, and I'm modifying my approach. When I've got some things settled in my mind, I will certainly be happy to discuss both my thought process and my ultimate conclusions with you, but am begging off at present - still thinking, still researching. Regardless, there are still some aspects and implications of both the theory and some statements in the article that I will post about later, independent of whether I decide the theory is likely to be valid or not. Sorry, that's the best I can do right now. :-( But like the song says, we've got all the time in the world, and I'm taking some of it. |
Yeah...I see sort of an 'Audiophile Cannonball Run' concept shaping up here...we blitz around the country in stages (trying not to get pulled), with members' houses as individual leg destinations, where we all stop for competitive auditioning of our systems' sound for bonus points, before going out to eat every night...
Talk about a 'power race'... :-) |
Well, we'll see about the way I think - I've finally got my initial reply from Mr. De Lima, and am proposing to engage in a limited dialog with him about his theory. More later... |
Now *that* sounds like the automotive equivalent of Thomas' audio system! I just hope the system is a little more comfortable and doesn't leave spots on the carpet...(I used to own a stock Spitfire :-) |
