Expensive amps and cheap speakers

November 12, 2017
 by Paul McGowan

Is it overkill to have an expensive high-performance amplifier connected to a less than optimal pair of loudspeakers?

It’s an interesting question. One that harkens back to Linn’s Ivor Tiefenbrun’s statement that the source is everything—a difficult to answer chicken and egg proposition. His original argument went something like this. The source is the most important element in the signal chain because if you can’t retrieve the information in the first place, nothing you do afterward matters much.

In hindsight, we know everything in the audio chain matters, though I have often argued that no matter how good the source is if you can’t pass the information through the chain of electronics you’re in the same pickle.

Here’s the thing. Even a mediocre bookshelf speaker sings to the high heavens when powered by a well-designed audio chain, but the same cannot be said of the opposite. A great pair of speakers driven by poor electronics lays bare the deficiencies of inferior electronics. Thus, I would rather have a less than perfect speaker connected to great electronics than the opposite.

It is not overkill to power less than perfect loudspeakers with perfectionist high-performance amplifiers.

I’ve put together a little video with some more personalized thoughts on the matter, here.

Subscribe to Paul's Posts

30 comments on “Expensive amps and cheap speakers”

  1. I must totally agree, Paul. The question however is what characterizes a good loudspeaker? I could connect various amp designs to a most expensive loudspeaker (4-ways design, external crossover board with some 60 components, linearized impedance, seven time aligned drivers symmetrically arranged around the tweeter) and heard no differences in sound quality with these amps ranging from tiny 5 Watt Triode amps to 1 kW SS amps. In contrast a tiny single driver widebander crossoverless revealed all sonic virtues of the different amps. Concerning the huge and expensive loudspeaker bigger was not better. For the small speaker the result was that a 25 Watt tube amp with output transformers sounded best. The only “advantage” of the big speaker was that it could deliver incredible sound pressure levels.

  2. Modern loudspeakers have crossovers which are much more advanced, and therefore much more complex than the old simple L-C crossovers of days gone by. The load to an amplifier isn’t a simple resistive load. We like to say that a speaker is a 4 ohm or 8 ohm speaker, but not really. Loudspeaker drivers are coils – inductors. The load is reactive, not resistive. Modern crossovers are complex circuits of inductors, capacitors and resistors. There’s NOTHING simple about the impedance load being presented to amplifiers today.

    One important thing that makes one amp “better” than another is how well they manage this insane load.

    And then in steps active crossovers. When you put the crossover between the preamp and your amp channels, then the amp only has to power a single loudspeaker driver. That’s a *MUCH* easier load to drive. An active crossover will also have much less distortion than a passive crossover. It’s beyond obvious that this is a much better route to go, but it doesn’t get done much. Why? Because people want to buy a loudspeaker and an amp to drive it. And maybe they want to change out the loudspeaker. Or the amp. Or…

    But the amp and loudspeaker together are an interactive system. When you multiamplifiy, you can also use much less powerful amps because so much energy is not wasted in heat loss i the passive crossover. And you can be quite successful with much, much less expensive amps because you’ve dramatically simplified the load to them.

    Or you can continue to buy the zillion buck loudspeakers with passive crossovers and Enterprise class boat anchor power amps to drive them. It’s only money.

    1. You can reduce the inductance of the woofer voice coil by a factor of 10 or more, a small effect on loudspeaker system impedance, by the invention of Ragnar Lian, an unsung Scanspeak engineer. Inductance causes a rise on the high end of the driver bandwidth, but is isolated from the amplifier by the crossover, and normal practice is to use an impedance compensation network which linearizes the crossover slope by cancelling the phase shift and impedance rise.

      The main effect of the “Faraday ring” is making the woofer (or midrange) faster, and the side effect is making the crossover simpler. It also makes it possible to fit a much larger motor to the tweeter so it handles more power and reproduces real acoustic transients. Faraday rings are expensive, but well worth it – and yet are extremely rare in contemporary products. I can only surmise that the market hears differently than I do.

      The main woofer impedance bump comes from electro-acoustic interaction because speakers are really MOTORS. This is a mechanical resonance reflected back through the motor winding, a resonance that causes large phase shift and long term energy storage – it shifts signal energy from the front to the tail of each bass note. This is why I have been advocating LOW Qms for 20 years, something other loudspeaker designers argue against.

      Vented designs add a second resonance and impedance bump. Having been raised on a grand piano and “the Boston sound”, I was never partial to vented designs. The double bump and null are three resonant frequencies which smear time and delay energy more (group delay), but this is rarely reported in graphs or words. Aside from the occasional mention of “fast bass” and “PRaT” (which could be other things), temporal distortion is vacant from the audio discourse.

      John Atkinson (former bassist) made one oblique reference to bass transient distortion in his review of the Sony ESS-M9, designed by Dan Anagnos. He wrote about the clarity of short electric bass notes, which made me plug my Fender into an oscilloscope and look at the waveforms. I found that when the note sounded right in headphones, it was an integral number of half cycles – but that was NOT what came out of speakers. After documenting this, I sold my Blond Bassman. It is now a collector’s item worth thousands, but it helped finance my search for a better bass speaker, completed 12 years later.

  3. I have the same bias that Paul expressed in his video; that the speakers are the most important items in determining the sound that you hear. I include their positioning and eventual room conditioning as part of ‘speakers’.

    In the vinyl days I would then have rated the turntable/cartridge combination next in importance. Unless you had an absolute dog of a turntable like an idler wheel Garrard SP25 then I would initially have focused on the cartridge. As you moved up the quality chain the interaction between cartridge and tonearm became increasingly important. Finally you went back to getting a better platter.

    Most modern consumer level amplifiers give surprisingly good performance. Even back in the day (quasi-complimentary class AB with no Baxandall diode) they were seldom the weak point. I did once build an early example of the Lynsley Hood 10w class A amp into a mono system for my sister. Within the limitations of the other components it did sound rather sweet, but it was under-powered for blasting out pop music. I never had any experience of tube amps, which at that time seemed to be relics of the radiogram era.

    I am not really sure about digital sources. They are still too new for an old codger like me. I have not been able to detect any difference between different DACs. However when I switched recently from a class AB to a class G amp (class A at normal listening levels) I could detect an improvement, so perhaps my whole philosophy has been wrong.

  4. I spent the most $ in my system on the DSJ and can honestly say the DSJ will tell you where your weakest link lies.
    I would definately recommend a speaker that has oodles of detail at low volume. I am inclined to try the Magnepan .7’s based on the reviews.

  5. Paul, this is a topic that will elicit a lot of discussion, some very passionate, however, I don’t think this is a fair question. For ages, sages’ advice has been, regardless of the quality of the components, to get the best sound, make sure that everything in the chain works well together. And this includes setting up the speakers and yourself properly in a good listening space.

    From the perspective of the electronics engineer the answer is quite simple: in the digital age the source is already near perfect; the electronics have unmeasurably low distortion; wire is wire; and the speaker is the obvious low performer. So place your funding and your research in bringing speaker distortion in line with even modest electronic components. And who ever needs more than ten watts in a home environment; even on loud passages my output meters rarely exceed one watt.

    I’m reminded of race cars in today’s Brazilian Grand Prix, which will have 1000 horesepower engines and obscenely wide and sticky tires to get that power to the road. Try installing a set of new Goodyear AllSeason tires on champion Hamilton’s F1 race car and I don’t think he will be very competitive. On the other other hand, allow him to keep his race tyres, but saddle the car with a 150 cc scooter motor and the result will be equally frustrating. This is not to imply that the 150 cc motor with the AllSeason tires, though compatible, will make him happy.

  6. I believe that the technical advantages of multi-amplification are beyond doubt.

    It is paradoxical that most audiophiles who have been for a long time, in search of the perfection of their audio installations, have not adopted the multi-amplification route with active Xovers. This is not so easy but it is compared to the acoustic conditioning of a room.

    Perhaps it is due to the home audio industry itself, because it has not offered enough electronic Xovers at reasonable prices, unlike that of professional audio in which passive Xovers have no place.

    On the other hand the bi or multi- amplification that uses passive Xovers IMO does not make sense.

    1. My modest home brew system uses electronic xovers and separate amps because I get more bang for the buck. It would not surprise me if the reasons that professionals do this is because quality xovers are heavy, take up a lot of space, and are expensive. And with electronics they can inexpensively offset some transducer deficiencies. When you have to buy, transport, and store a lot of equipment, those factors become important.

    2. I believe in electronic crossovers. But for me each one needs to be designed for the speaker it’s used for ideally. Look at the crossover of any well designed passive speaker. Just for argument assume it’s a 4th order Linkwitz/reilly acoustically. If you look at the actual crossover it’s probably a combination of 2nd and 3rd order crossovers electrically because a well designed crossover takes into account the electrical and physical characteristics of the drivers and crossovers. And it’s probably correcting driver anomalies in the middle of the driver pass bands. In an analogue world this is a custome design, not just variable slopes and crossover points. Perhaps in the digital world it’s simpler to make the adjustments on a more generic design but then one needs to do difficult measurements as one is adjusting.

    3. “This is not so easy…”
      Active multi amp set up is VERY difficult, and few audiophiles would be able to achive good results. It is like providing amateurs a 31 band equalizer without a reference signal with that they could verify their ‘improvements’. On top of that in a three way system it is already quite a challenge to set the right levels for each speaker (assuming that the filter slopes and delays are set perfectly).
      Another challenge is finding the fitting amps for bass, mid and treble which all have to meet different demands.
      Not everybody is an experienced SE.
      Further there is the problem that you bought an expensive, fine sounding DAC and then route the signal though a more or less mediocre electronic xover with another AD to DA conversion and less refined output stages.
      I have been searching for an ideal 3-way solution for a long time and did not find a suitable piece of gear: a top quality digital pre amp, xover, and triple DAC in one box.
      This unit comes very close
      but is lacking adjustable delays. Mini dsp does not provide a flexible DA conversion according to today’s requirements (AES, DSD/384kHz etc).

  7. I would have thought there are only a few sensible things that can be said on this topic, namely:
    (1) Seeing IvorT was flogging turntables, he would say that, wouldn’t he?
    (2) He was talking in pre-digital days, so largely irrelevant now.
    (3) If your speakers cost about one third to one half your system cost, you are probably on the right track.
    (5) There are enough good loudspeakers out there not to have to buy the rubbish ones.
    (6) Every link is important in that if you leave out the speaker cables you won’t get any sound (basically what Peter Walker of Quad said about cables). I wouldn’t go further than that. At the other extreme, some would consider their system unusable without replacing the stock fuse with a Kryptonite one for $100.
    (7) Everything is relative to your budget

  8. It’s resurfacing a topic covered in past posts. I’m in the speakers are most important camp. Every speaker change to my system has yielded a drastic difference. Far beyond what a component change has ever had. Speakers are the only part of the audio chain that transform what was previously some form of electrical signal to something you can hear.

    To assume that once to a certain level that speakers make less difference, and the electronics make the biggest difference is contradictory to every experience I have had. I have found it’s quite the opposite. I have noticed that once a preamp & amp get past a certain price point, the differences between them become less & less, regardless of the maker. As an example, I have had a Threshold T200, Parasound A21, Emotiva mono blocks, older PASS amp, older Classe amp, Krell, etc. I definitely have my preferences as to which are the best amps, but in reality, they are all very close each other in sound characteristics. However, the same is not true for speakers. At the same relative price point, there are remarkable differences between Magnepan, Elac, B&W, Zu Audio, Monitor, Infinity, etc.

    1. I agree with speaker budgetary priority, but in my classifications all the marks you mention are in the “B” category, along with Bryston, B&K, various TriPath makes etc. I use amps which cost $60 to $160/channel, and are in the same range of small sonic differences at Living Room levels.

      OTOH, there are “magic” amplifiers by designers like Glenn Croft, Tim de Paravacini, Charlie King and Charlie Hanson. I built a highly optimized LM3886 amp, similar to the 47 Laboratory 4706 GainCard amp (but with high current power supply), for a solid B grade sound – and the “A” grade amps blow it away.

  9. This question can’t possibly have a simple answer, because there are so many ways to compromise speakers and so many different permutations of musical waveforms that each illuminate distortions differently in addition to the amp-speaker interactions discussed above. Distortion compounds geometrically, i.e. distortion of distortion sounds worse than distortion of signal, so it adds up exponentially like compound interest rather that summing algebraically.

    The Tiefenbrun principle also has a zero step in the chain – the recording distortions. These are predominantly from the distorted hearing of recording engineers. They are experienced listening to bad speakers – most learned to hear music on bad audio in noisy urban environments – and typically they spend a lot of time worrying how “the mix” will sound on cheap speakers.

    The scientist-engineers at Harman proved that on average, cheap speakers are flat – there are as many with anemic bass as boomy bass, and shrill ones equal muffled ones etc. Attributes nearly all speakers have in common are too small drivers with excursion limited dynamics and Doppler distortion; diffraction lobing and comb filtering; transient bass distortion; and bad room interaction and setup which interferes with articulation, intimacy, envelopment and “imaging”. These lead to the most common recording process distortions: close miking to multi-tracks (spectral and spatial distortion); compression and limiting (dynamic, temporal and spatial distortion), pan pot stereo and artificial reverb (spatial and temporal distortion) and equalization (spectral and temporal distortion).

    Further, all artificiality leads to emotional distortion by affecting the playing. Isolation, headphones, splicing, overdubbing and mixing are all barriers that humans can learn to deal with, but block the natural flow of music from fingers and lips to ears.

    There are also very high selling engineers who use recording chain overloads to sell sound. I confess to using moderate amounts of tape saturation on bass tracks when I was young because it sounded better than dbx compressors and noise reduction, and because consumer systems are down the loudness curve and need bass compression. I also put bass and kick drum on the edge tracks and optimized the alignment adjustment for minimum distortion at the expense of reduced response above 10KHz. Moderate overload was normal in the ’70s and ’80s, but carried to an extreme of pushing all tracks and the mix over 0VU. Now Gen X and Millennials are using digital overloads (clipping) deliberately!

    What is really distressing is that the vast majority of music auditioning is now digitally compressed files, which compounds with all the other synthetic distortions for brain-damaging sound. We are raising a generation with neurologically stunted hearing from streaming, AAC and MP3 compounded with ultra-cheap phone and computer circuitry, earbuds and speakers – particularly wireless speakers which are all dreadful.

    I will say that quality of amplification up to the best chip amps is quite audible on well designed budget-fi speakers like the Martin-Logan Motion 2’s on my desk and the EAW UB12’s* in my bedroom, which are both quality treble satellite speakers needing an external woofer. I have the matching sub from EAW – an SB48 – and I am building a desktop sub for the ML’s, which has to be vibration cancelling.

    *The UB12 has a 1″ dome with a waveguide so it is not suitable for unprocessed music. I use it for TV soundtracks which are reliably compressed and oriented towards speech.

  10. The problem is analogous to one of my tasks as an electrical engineer. In short it relates to providing a power supply to an electrical load. The difference is in the details. One of the first tasks is to study the load even if the load equipment hasn’t been purchased yet or the building not built yet. Undersize or build the wrong type of power delivery system and the equipment won’t work reliably or at all. Oversize it substantially and you’ve wasted other peoples’ money, often by a lot.

    It is good that selecting a loudspeaker is the first order of business, not just because it has the most influence on a sound system but because it defines the minimum requirements of the amplifier and connecting wires. Speakers have widely varying power requirements. It’s not just the number of watts it needs, there are other factors involved. The one size fits all theory is a poor engineering philosophy. Yet when you assemble an audio system you are the engineer. It’s insane to connect a 300 watt per channel amplifier to an Altec Voice of the Theater that can fill a movie theater with a few watts. Connecting a 3 watt per channel SET amplifier to the midrange and tweeter of an Infinity IRS V is also a bad mistake. Some loads present very difficult challenges that have to be handled just right. Some speakers like AR3 present loads at some frequencies below 1 ohm, almost a dead short. 4 ohm speakers do not work well with many home theater receivers and their amplifiers will shut down to protect themselves or blow themselves up. Others like electrostatic speakers present a highly reactive load and need a lot of power. The YG Acoustics Sonja 1.1 presented a nightmare for amplifiers even worse than the 1.3. At low frequencies it was looking into a very large capacitor. At high frequencies a moderately high inductor. This leaves you with a blistering array of choices. Tube or Solid state? Bipolar or MOS? Class A, AB or D? Pentode, Tetrode, or Triode output? Which wires?

    Bottom line, there are no easy answers. Unfortunately for audiophiles the problem often comes down to trial and error or taking other peoples’ advice, some of it having a hidden agenda of selling particular products, others not knowledgeable at all. Will the BHK monoblock amplifier outperform all other amplifiers under all circumstances? Nobody knows. Will it be radical overkill for some installations and a waste of money? Yes. Will it be inadequate for other applications such as where far more power than it can provide is needed? Yes.

    Can Bob Carver make one solid state amplifier sound like any other within its power capabilities as he once claimed? I’m not sure, you’d have to ask him.

  11. Since the signal starts at the front end it stands to reason that it be the best but should be followed by components which can do it justice. so it goes to the very end. Your point that everything before the speaker should be of high quality is absolutely correct. This way one is assured that the weak link is the speaker which can be changed when convenient. A point to remember is that speakers, even not very expensive ones, can be quite good nowadays. Excellent piece of advice. Regards.

    1. In order to do justice to the source the next element in the stereo chain must have better performance data than the source. Otherwise the quality of the source couldn’t be revealed but rather degraded. Following this logic the power amp not only has to match the speaker’s requirements but also has to offer the best parameters being acoustically relevant, isn’t it?

  12. Tiefenbrun makes and sells turntables. Paul makes and sells amplifiers. It stands to reason that each manufacturer will tell you his product is the most important. That is NOT engineering, it is sales and marketing. At least Paul is one of the few in the business who tells the truth even when it is not to his advantage. Since loudspeakers have the most effect on what you will hear, it is natural that this component should be selected first. Which sounds better to your ears, the Elac Unifi UB5 at $500 a pair or the KEF LS50 at over $2000 a pair? What if you paired them with subwoofers, possibly powered with their own amplifiers? This whether you realize it or not is the strategy of so many so called high end speakers like Wilson. Do you need a pair of BHK monoblocks to power either of them? Are there better choices? Cheaper choices that are just as good or almost as good for those speakers?

    There are much better turntables than Linn makes, why is his good enough? He doesn’t like his speakers demonstrated in a room where other speakers are present. What is he afraid of? Direct comparison? Is he saying the presence of other speakers in the same room would affect the performance of his speakers? Maybe you shouldn’t be in the same room with them either? Unfortunately for him, that’s a necessary evil.

    1. Linn doesn’t want his or any other speakers demo’d in a room with other speakers because they cause sympathetic resonance in the other speakers in the room, not to sell more loudspeakers but to hear each speaker at its best and decide accordingly. The single speaker demo makes sense for another reason, most of us only have one pair of speakers in our rooms at a time. Even when we allowed in home demos of non-Linn speakers, we always tried to have one pair in the customer’s room at a time. Hear the speaker as you’ll listen to it for years to come.

  13. I have a pair of discontinued Cambridge Audio S30 speakers in my closet used as my back up speakers. They were “cheap” compared to most equivalent offerings out there. Cheap? Yes… price wise, and possibly because of the rather basic looking cabinet finish. The crossover? Simple first order. Meaning? Its phase coherent. Meaning? Great imaging and impact potential. But, inexpensive…

    Listening with a great front end? They make you wonder what its all about. Its like watching a beautiful model walking down the runway in a pair of Keds sneakers and expensive evening gown. In spite of it. She still remains beautiful.

  14. Years ago I sold a cheap of speakers I had lying around which were originally used with a cheap amp. When I sold them I had them connected to an amp that was 10-15 times the price of the speakers. I was taken back to how good they sounded with a good amp. I told the buyer the speakers will probably not sound that good when you get them home. He still bought them. I too feel more money invested in a good amp is better than spending more on speakers and settling on a mediocre amp. Good speakers under perform so much with poor amplification.

  15. This is quite in line with a source first, or in this case, earlier components are more important, theory. That said I agree that a better amp can bring a less than stellar speaker to life more than a mediocre amp can but its more than that, it’s a synergy thing. Way back when I sold high-end gear we could never get the Quad ESL 63s to sound like we knew they could. One day we hooked them up to, of all things, a PS Audio IIC+, and voila, music. Those speakers had been matched with Audio Research, Conrad-Johnson, Perreaux, Adcom, etc. and the little PS kicked all their asses, synergy.

  16. Yep, I have to agree. Up until last year, I had been using speakers I bought back in 1979 for $280/pair. My system was in sort of holding pattern for many years (due to life events), where a small number of sources came and went. Only in the past few years did I start upgrading…and darned if I didn’t hear each improvement through these same speakers. Granted, I didn’t get the full effect of each upgrade, but I did hear the difference each time. I currently have been living with a pair of Vandersteens until my Martin Logan situation is sorted (I am either refurbishing the project I currently have, or just jumping up to a used set of a more recent product). It only gets better each time I upgrade, and I finally have sources and electronics that are worthy of these speakers.

  17. Hi, I got my hands on a floor model of a Mcintosh MA-5200 for $2000 (it normally sells for $4,500. Unfortunately, I blew my budget. So, i could only afford a pair of Elac Uni-Fi UB5 speakers ($500).

    Does it make sense to flip my amp for $3,500 or so and use the profits to get a cheaper amp and better speakers? Or are the UB5s sufficient (i just ordered them so haven’t heard what they sound like powered by the Mcintosh).

    Thanks in advance!!!

  18. I think a simple way to phrase it would be having an amp and loudspeakers that are well matched.
    One easy example that illustrates this issue is the Sennheiser HD800S headphones. If you read a bad review the first thing I look for is how were they powered? If you hook them up to a decent headphone amp you may think they are not that great and not worth the price, but when you hook them up to a powerful high impedance tube amplifier they come alive and the magic begins. So it not the price or the wattage, its a synergy between the speakers and the amplifier.

Leave a Reply

© 2023 PS Audio, Inc.

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram