Soft underbelly

January 22, 2019
 by Paul McGowan

Life and stereo systems are compromises—a fact that causes us to make less than perfect choices. It’s helpful to understand the consequences of each choice.

Take for example amplifier power. In a perfect world, we’d have unlimited power and the best sound quality imaginable. In the real world of limited budgets, we’re likely to have to make some tough decisions: Power vs. ultimate sound quality.

If you find yourself on an economic leash that narrows your choice between high power and the sound quality benefits of small high-bias amplifiers, the first thing you want to do is determine where your soft underbelly is most vulnerable—dynamics or see-through transparency.

If your listening habits are mostly small ensemble with the occasional burst of dynamics you’re likely better off focusing your amplifier choices to those you believe offer the sweetest presentation of detail and nuance at reasonable listening levels.

If instead, you’re anticipating dynamic and spectacular, then high power trumps inner detail. There’s nothing more jarring than being jerked out of sonic bliss by the strain of struggling dynamics.

Of course, as manufacturers, we try and offer the best of both worlds as best we can but getting there isn’t always easy nor inexpensive.

When compromise strikes, do your best to figure out what’s important so you can let go of what you can’t have.

Subscribe to Paul's Posts

23 comments on “Soft underbelly”

  1. If every detail concerning component matching, room placement, room treatment vibration control, grounding, component interference, power supply etc., etc. matters when optimizing a stereo set-up then just simply exchanging a single component without further optimizing the new set-up is not a good idea. Stereo is a highly complex matter and not as simple as buying a new mono radio or Walkman or Discman. Concerning the limited budget and space available today for most music lovers I wouldn’t start today buying a set-up of separate components but a modern sophisticated active speaker (as from Devialet, KEF, Genelec, Kii Audio, etc.). Surpassing these minimalistic concepts significantly concerning sound quality will require insane budgets and a lot of time for optimizing. However if one is lucky as you, Paul, starting from scratch and building a dedicated music room we are telling a different story far from reality for most music lovers. And isn’t your statement about inherent compromising rather frustrating and discouraging?

    1. Nice thought, but most active speakers out there are not what I would consider high fidelity but rather fall into the “OK for the cost” category.

      It’s rather like buying a receiver to drive your system: not that a receiver can’t sound good but it’s most often the source of significant (bringing this topic full circle)… compromise.

      1. Interesting thought, but “most” isn’t a useful concept.

        I have a pair of Vanatoo Transparent Ones (about $500) for our holiday home in Normandy. These are extremely good speakers if for whatever reason one simply has no space for anything bigger than a bookshelf unit. But, no, they’re not as good as Paul’s giant Infinity’s, nor my old B&W 800Ds.

        I also have a pair of Kii THREEs for the home here. Which are about the size of the traditional 2-way stand-mount speaker, and which are noticeably better than my old B&W 800Ds (driven by Stellar S300s bi-amped). But these are $12K plus stands, so you expect a good speaker…

        Having just had a listen to Paul’s system (thanks, PS Audio!) I agree that the Kii’s don’t have the ability to drive so much bass that the room explodes, which can be a risk with the IRS’s. But they go deep and clean and can provide a clean higher listening level in my 13x14x10 high room above 100dB-C.

  2. A pretty good maxim would seem to be to get the most appropriate and affordable speakers and an amplifier capable of driving them. I presume that is why the Sprout and similar devices usually have a sub-out: the assumption that they will be partnered with speakers incapable of low bass as it is dramatically cheaper to use a subwoofer than full range speakers and big amplifiers. Paul makes a good point, in that I sometimes despair when I read about people getting great speakers and pairing them with expensive under-powered amplifiers.

    1. I don’t think that small speakers with a sub are *quite* as effective as large full range ones (which often benefit from a sub as well to give added definition to percussive elements in music) but they come pretty darned close. I have been very impressed by my new small+sub combination.

      1. Indeed, in one of my sound rooms I have a pair of Epos ELS 3 together with a pair of Legacy: Foundation passive subwoofers, since the amplified subwoofers I’ve heard have been one note bass, which I detest.

        Each Foundation contains 2 subwoofers per channel, which in turn are fed by one amplifier for each subwoofer, so I use 3 amplifiers per channel. This is to keep the Xmax to a minimum minimorum.

        This system is optimized to listen to videos of classical music that do exist and very good, in addition to informal listening.

        However, I have the firm conviction that if one pays a huge amount of money for a full-range speaker system, one should not use subwoofers with them, because if so, the manufacturer would be charging too much for a full range speaker , when in reality, it is NOT, or in its absence, something is wrong with the acoustics of the room

        What was said in the previous paragraph, is not new to me, because when I bought my Infinity QLS, whose frequency response according to Nudell (RIP) is 18 Hz to 32 KHz @ + – 2 dB would have been a fool to think of buying a pair of subwoofers to use with them.

        Subwoofers to go lower than 18 Hz?

         No thanks!!

        1. I do not think there is any fundamental disagreement between us. I would ideally have full range speakers which were flat between 20hz and well over 20khz. The use of a sub (ideally a pair of subs) is a compromise to compensate for loss of bass response in speakers which start to roll off at a higher frequency than 20hz, and I have found it enhances more than it detracts, but it is a plaster. The question of extreme subsonics is an interesting one. When I look at spectral plots of orchestral pieces I normally see then dropping virtually to zero around 20hz, but then rising again below 10hz. I suspect this rise is due to hall resonances, and if one had an amp and speakers which could reproduce them you might get a better feeling for the experience of listening live. Whether this verisimilitude would be desirable from the point of view of the music is another matter entirely. It is something I would want to try before buying into it.

          1. Honestly, I do not understand your point well.

            Among the systems of speakers considered Hi-Fi, the QLS presents one of the most flat response, thanks to the novel (for the time) use of a woofer of 2 voice coils dissimilar, aligned so that one of them comes into operation in the moment and in the frequency, in which the cabinet necessarily increases the acoustic impedance due to its resonance, at this moment this second coil, enters in parallel to force the amplifier to deliver more current to the speaker, thus achieving a flatter response than other conventional speaker systems.

            This is because in other types of speakers, increasing the impedance of the system decreases the current that the amplifier is capable of delivering, therefore the response curve is far from being flat

            Thus, this speaker is able to deliver a response not only more flat, but more widespread especially in the lower part of the spectrum, which has never been very easy for most manufacturers, including the best known.

            My point is that, speakers like the one mentioned here, do not need imperiously subwoofers, which on the other hand if they are not properly installed, cause more difficulties than benefits.

            Regarding: …. “and if one had an amp and speakers which could reproduce them you might get a better feeling for the experience of listening live.” ….

            I think that what you have said, even if it was hypothetical, will not be possible even in the distant future, not even with the progressive advancement of technology, unless a listening room has the dimensions and characteristics of a concert hall. .

  3. Nelson Pass’s axiom that ‘the first watt is the most important’ is very relevant. Most of us listen at levels where the average power per channel from the amp is less than a watt. If you look at the broadcasting program standards which say that the average program level should be -23db down from the maximum, then you would not need an amp with more than 200Wpc. At my normal 70db listening level that means that my amp should never need to deliver more than 6Wpc!

    1. Dynamic range plus speaker efficiently are the bugaboos there. Anything with even reasonable dynamic range is going to require power; whether that requires more than 200 watts is largely dependent on your room and whether your speakers have an efficiency of 85 dB/watt or 100 dB/watt (at 1m.)

      Do you want to listen to a rock concert at live volumes or a solo cellist at background levels? That’s a factor too.

      It is however true that if an amplifier can’t produce five amazing sounding watts per channel, it’s not going to magically produce a mind-blowing 300.

      1. Chris and I use the same brand of speaker (Harbeth), PSA also has a pair, and they are a known quantity. Mine, subject to a few upgrades, are in their 42nd year of continuous production, in considerable numbers. The range is small and there are really only 3 drivers in terms of power considerations, 5 and 8 inch mid bass and a 12 inch bass for the full range speaker. Moreover, as they are mostly BBC licensed, there are other companies making very similar products, such as Spendor. Certainly in the UK, and also in Europe, customers are well aware of the amplifiers that they should be paired with, basically Rega, Quad/Audiolab and Croft. They are matched in performance and price, Rega probably being the most popular as there is a wider choice of about 3 models at different cost and power levels. Besides the value for money, a long-established known quantity makes life very easy for the consumer and you can’t really go wrong.

        Moreover, within the limitations of the speakers, the amplifiers do not really present a need to compromise. These brands operate in the hugely popular 70w-140w into 8 ohms market, which is not PSA’s zone, probably a good thing as there are a lot of manufacturers making excellent products offering great value for money.

  4. Careful selection of components is always a primary consideration. As I have said before, I utilize both old and new school design for my needs, however, that doesn’t mean you cannot obtain very high performance, even if on a budget. Case in point, I utilize an amp that was produced 35 years ago, purchased at a very reasonable cost, and refurbished to replace dated components. It brings with it 0.003% THD at rated power, >120dB of dynamic range, >3dB headroom, 200V/µs slew rate, 160,000µF of power supply tank circuit, 290 watts continuous. All for much less than a grand. You don’t have to sacrifice sound quality if you take the time to understand the goal, and select carefully.

    1. Barsley,
      Out of interest to me and I’m sure others, would you please ‘name that amp’. I appreciate this is a PS Audio forum but Paul seems very open with regard to other manufacturers receiving a mention and I don’t believe he would have any objections. Do correct me if I’m wrong Paul.

      1. That is an amazing specification and better that a pair of 300w Quad 909 mono blocks that I had, 20 years old and cost me £1,000.
        The above specification is pretty much equalled by the Audiolab 8300MB, a mono amplifier designed in the same office as Quad and made in the same factory. A new pair would cost £1,800. There are three typical audio industry problems: the brand is much better known for smaller integrated amps and DACs, not mono blocks, at that price, no one needs that much power and finally the current Quad QMP monoblocks are not much more expensive. They are almost too powerful for their market positioning and I’m not aware they were very popular.

        1. From HiFi Classic’s review; “The M-80’s power performance into 8-ohm loads was quite impressive, but the really special qualities of the amplifier became more apparent when we drove lower impedance loads. For example, its dynamic headroom into 4 ohms was no less than 3.78 dB, with short-term power getting up to 787 watts per channel.

          Finally, we came to the 2-ohm measurement, the nemesis of so many amplifiers. It wasn’t even practical to drive both channels simultaneously into 2-ohm loads, since this overloaded our 15-ampere test-bench power system! Driving one channel, the clipping power was 612 watts, and the dynamic power was a staggering 1045 watts. The amplifier was stable when driving reactive simulated speaker loads and the standard IHF reactive load.

          Given Yamaha’s emphasis on the low-distortion circuitry of the M-80, we should not have been too surprised at the results of our distortion measurements. They were indeed difficult and time-consuming to make, since the 1,000-Hz THD of this amplifier (driving 8-ohm loads) was less than 0.0025 percent up to the clipping point, and typically well under 0.001 percent.”

  5. “Take for example amplifier power. In a perfect world, we’d have unlimited power”

    For the second time in two days I find myself disagreeing with Paul. As an engineer my job is to solve a problem by first defining it in all of its aspects as completely and precisely as possible. So say you had an amplifier and speaker that could play as loud as 250 db. Would that be ideal? No, far from it. 210 db is lethal to human beings. What would be the point? It’s more than just a waste of money, it creates an unnecessarily dangerous machine that could have dire unintended consequences. What’s more that alone would not address many other aspects of the requirement for an ideal sound system. We specify required performance in all aspects and with tolerances and then we try to build things that achieve those goals. Many times we overdesign by a reasonable margin to account for unanticipated problems we hadn’t foreseen or changing requirements. Engineering is an art that requires among other things knowledge of science.

    Paul has receptacles in his offices that are fed most likely from a 75 kva transformer but possibly 112.5 or even 150. Would it be ideal if it were 7500 instead? No, that would be far more than enough for his entire building. It’s not just a waste of money, it’s bad engineering. It would create other problems that would have to be solved. If something you engineer doesn’t perform to your satisfaction, the thing to do is go back to your original assumptions and try to figure out what you missed, not just do the same thing on a larger scale. That may be part of the solution but it is not necessarily all of the solution so by itself it might not be satisfactory.

  6. One thing many seem to forget is that passive crossovers are the source of a great deal of inefficiency in loudspeakers. That’s why so many pro systems use active crossovers. When you multi-amplify and an amp channel doesn’t have to deal with a crossover, but just power a single loudspeaker driver, you need much, MUCH less power, so that “First Watt” has even more impact.

    Much of what you pay for in a high cost power amp is the ability to manage the highly complex loads that modern passive crossovers have. They don’t just act as high & low pass filters. They also do some amount of EQ, and Engineers Only Know What Else. You need a megabuck amp to cope with all that. But when you use active crossovers, you can get away with much less powerful and costly amps and do just fine.

    The reason you see so little of this is that manufacturers have a much more difficult time selling a package that has to have active crossover, multiple amp channels and loudspeakers. The common mindset is to pair reputable amps and loudspeakers together, leaving a sizable budget for speaker cables capable of being used as a murder weapon. You see active crossovers in a few loudspeaker products that have the crossover and class D amp channels built into the box, like the small KEF wireless LS50, with line level and bluetooth wireless inputs. Most of the time you see active crossovers is in DIY setups or kit systems like the Seigfried Linkwitz designs.

  7. I’m conducting my own “less is more” experiment that comes down firmly on the side of naturalness over loudness. (My speakers in my listening room are Magnepan 0.7s, so that tells you what I like.) I recently built a Pass amp camp amp (single gain stage, 8 Watts/channel class A) from diyAudio, and was so pleased with the sound I built another and run them as monoblocs (15 Watts each). They are fed in balanced mode by a Music Hall 25.3 DAC and driving 89db bookshelf speakers to comfortable levels. Now I’m building a Pass diy Zen Balanced Line Stage preamp, probably configured in a 10db gain mode. If all is well I’ll finish up with higher efficiency speakers, preferably full-range drivers to avoid crossover issues. “The first watt is the most important” indeed.

  8. Hi for me I have a modest two channel system that I started to put together in 1985 it consists of a Pair of Acoustic Research 93s modified with scanspeak midrange and vifa tweeters and bravox woofers a 80s AR Turntable with a sumiko FT3 arm and a new Nagaoki cartridge and a wife friendly Onkyo receiver I think it is important to put your money in the source and speakers

  9. IMO the major effort for improving his best systems right now for Paul is what I call the sledge hammer approach. If you can’t knock down the wall with the sledge hammer you have, get a bigger one. Eventually you will get one big enough to knock it down… if you can lift it. 🙂

    Building a loudspeaker that can play sound undistorted at 125 db or 135 db at all frequencies will not solve the problem of the dynamics of a symphony orchestra playing as loud as it can reproduced from a pair of speakers. The problem is that 90 percent or more of sound you hear live comes from reflections. But the 2 channel sound system doesn’t design for that and the recording doesn’t have them as part of the signal. The reflections come from everywhere, all directions in rapid sequence. If you could block out the direct sound and hear just the reflections the loudness would drop by less than 1db. For all practical intents and purposes that is inaudible. This first arriving sound tells you where the source is. It also sets the stage for what is to come by way of reflections. While the loudness of the reflections from any one direction is inaudible, in aggregate that are almost all of the sound. So making the source and what little reflections are in it coming from two directions and playing it much louder is a bigger sledge hammer but it still won’t knock down the wall between you and the music.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

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