The Problem With Loudspeakers

I have been listening critically to audio systems, especially to amps driving loudspeakers, since the late 1960s; since my friend Bill Brier forced me to compare a tubed Dynaco Stereo 70 amplifier to a solid-state Dynaco Stereo 120. We played saxophone jazz, piano concertos, and Bach organ music through Brier’s Dynaco A50 loudspeakers… and listened for, what Brier called, “A proper ten octaves.”

I was 17, Brier was five years older. I helped him build drag-race cars and nitro-burning motorbikes. He played acoustic bass in a jazz band and tuned pianos for a living. On top of all that, he had taught himself to play every instrument in the classical orchestra and memorized the complete keyboard works of J.S. Bach.

His favorite thing was to drive his 1958 MGA (powered by a Chevy mouse motor) to places like Boston or Philadelphia; where he would break into churches and play Bach fugues on real pipe organs. Every time they arrested him he protested, “I have a right to play this organ! Because I can!” Brier believed his talent and intelligence placed him above the law. So did I.

The beauty of my first critical listening day was how my older friend guided my attention by explaining the meanings of words like timbre, frequency, octave, and harmonics. Brier explained how a classical organ concerto contains ten full octaves of sound that are exponentially related and how I should be aware of these intervals while listening to recordings. During the comparison, he explained counterpoint, then played one E. Power Biggs recording after another – with rapid-fire pontificating and gesturing towards the vibrating air between the large Dynacos.

Bill Brier taught me how to tune car engines with a tach & dwell meter, but more importantly, he explained how to listen to “all eight cylinders” through the carburetor and exhaust manifolds. He taught me to recognize the sounds of camshafts and tappets and the low-pressure air screaming through a carburetor’s Venturi. He taught me to “feel and smell the garage” while listening with my entire body; he said I must “analyze everything at once.”

What I learned from all this was to listen mindfully; giving special attention to the inner nature of the sounds.

Brier was an amp builder too, and during our solid-state vs.tubed Dynaco comparison, he explained, using words I didn’t understand, how the junction between the amplifier and a loudspeaker was like the combustion chamber of a car engine. “Imagine you can see inside a combustion chamber while the engine is running. Notice the valves smacking against their seats. The intake valve is the amplifier’s power supply. The exhaust valve is the speaker’s voice coil. Notice what happens during those short moments, called overlap, when both valves are open at the same time. That is when the intake and exhaust are completely connected: where the tuning of the intake is enhanced by the tuning of the exhaust; by allowing the exhaust headers to suck extra fuel-air mixture into the chamber.” In loudspeakers, he said, “…that is where the amplifier’s output connects most directly with the speaker’s dynamic impedance.” According to Bill, “That is why the Stereo 120 is superior to the Stereo 70”

On the day of the Dynaco comparisons, I got my first lessons about tight bass, and loudspeaker-Q, which Brier described, not as damping, but more like the compression ratio in an automotive engine. He said car pistons and cone drivers were the same.

He explained Ohm’s law and how because of my Stereo 70’s high output impedance it could not connect as effectively to the Dynaco A50’s paralleled woofers as the lower impedance solid state Stereo 120 could.

During the decades that followed, I practiced listening for this amp-speaker junction. I have tried, but mostly failed, to understand the complex dynamic interactions that take place at this freakishly reactive meeting place where a complex music signal meets a pool of stored energy, earth ground, complex-series and parallel-filters; plus the reactive forces generated by a rapidly-moving voice coil. This is where voltage and current do their wicked dance. This black hole hell-region absorbs 99 per cent of the energy that enters it. The one per cent that survives to actually shake the cones has been twisted and battered in both the time and frequency domains.

Is this really the best way to generate natural sound from a loudspeaker?

Because modern hi-fi systems rely mainly on voltage amplifiers, the necessary voltage-to-current conversion is accomplished in the above mentioned crossover/driver hell region. This means that nonlinearities of the driver, caused mainly by driver motion and the resultant back-electro-motive-force (EMF) are preserved as an integral part of the dynamic system.

Some of these non-linearities can be ameliorated by flattening dynamic impedances with zobel or “trap” filters but these strategies absorb even more power and further reduce speaker efficiency.

Unfortunately, anything that adds time delay or energy storage between the amplifier’s output and the speaker’s input changes the total “Q” (or “quality factor”) of the box, magnet, voice coil and suspension system; and that change is always audible. Changing loudspeaker Q makes the sound (subjectively speaking) softer or harder, thicker or thinner, slower or faster.

In my oversimplified imagination, these are the most obvious amp-speaker interactions that affect our perceptions of amplifier or loudspeaker quality.

That is why last month I suggested that “The Problem With Tubes” was really a problem with speakers.

My point was: if you design a giant speaker with a variety of drivers, with a variety of magnet-voice-coil-suspension systems, in a variety of boxes, all connected to a complex series-parallel crossover fitted with a variety of error-correcting filter networks, you have created an insane Frankenstein monster that costs insane money and needs to be driven by a ridiculously complex, excessively heavy, overly-powerful, and unnecessarily-expensive amplifier. And to what end? Impressive measured frequency response? Ten per cent better imaging than some other giant robot?

When a loudspeaker’s main selling point is how its dazzling array of drivers achieve an unparalleled degree of time alignment between the frequency ranges and a time-coherent presentation that clarifies information smudged by lessor speakers… I pray that you can see how impractical and unmanageable this makes the amp-speaker interface. I hope you can also see how this 250-pound, million-dollar, multi-driver fully-error-corrected speaker requires a 250-pound, million-dollar, error-corrected amplifier to drive it. Throwing parts, fixit-filters, and digital processing, at audio reproduction problems seems lazy and defective as a strategy.

You be the judge: are these sane intelligent cost-effective ways to create natural-sounding low-distortion music in our homes? Or is it possibly a conspiracy to relieve unsuspecting oligarchs of their offshore holdings?

What if a speaker had only one full-range driver? And no crossover? What if that single driver had a ruler-flat impedance and phase characteristic? What if that loudspeaker had tight deep bass and extended treble? And it could pass a decent 300Hz square wave? What if it needed only 1.3mW to play at 90dB SPLs?

What if it was a headphone?

If you compare InnerFidelity’s Audeze LCD-4 and JPS Labs AB-1266 headphone measurements (1) to any of Stereophile’s measurements of multi-driver floor-standing loudspeakers (2) you will understand every point I’ve tried to make.


grantray's picture

1. Thank god you answered your run of what-ifs with a headphone and not the WE755A.
2. What load were you hooking the phones up when going direct to DH-SET amp? 4, 8, or 16?

Herb Reichert's picture

I have a tendency to prefer the 4-ohm tap - but I always try both

grantray's picture

How about a shot of those cables you made?

PeterInVan's picture

As much as I enjoy my LCD-2F headphones, my personal preference is listening to the KEF LS50W powered speakers.
KEF has eliminated the need for amps, preamps, DACs, interconnects ... all external devices.
They have dedicated DACS and Amplifiers for each speaker, as well as DSP processing for room tuning.
The LS50W is a worthy alternative to high end headphones in terms of power, sound quality, sound stage and imaging.

geoffreyvanhouwaert's picture

straight into the speakerouputs, and here i was about to buy a pass hpa-1

Herb Reichert's picture

the Pass Labs HPA-1 is a very fine headphone amplifier, but is is (to my taste) a little shy on gain for headphones like the Abyss AB-1266 Phi and HiFiMan Susvara which, when connected to the First Watt J2 or SIT3 amplifiers become stunningly clear and untethered dynamically.

I like gain and energy in my hifi chain - it works like caffeine

I am certain both headphone manufacturers would STRONGLY discourage you from trying my experiments and void your warranty if they found out.

just saying :)

Ortofan's picture

... from those integrated amps and receivers whose headphone outputs are connected to the power amp output through a single resistor?
Of course, that would increase the output impedance of the driving circuit and could result in frequency response variations that are a function of the impedance variations of the headphones.

pbarach's picture

I don't have headphones wired for balanced output. So to duplicate what you did, Herb, I need a headphone cable that terminates in banana plugs. My DAC source feeds into a line-level preamp, which is connected to a basic power amp, and then the banana plugs are connected to the speaker outputs of the amp. Is that correct?

Herb Reichert's picture

Just like speakers - but you must be extremely careful to avoid turn-on spikes or signals that could damage the headphone drivers.

Everclear's picture

Can a headphone amp which can put out more watts, say like the Sim Audio headphone amp for example, do the job? .........

Everclear's picture

Another example is Cary Audio CAD-300 SEI integrated amp ........ It is supposed to put out 15 tube watts via the headphone output ...........

Herb Reichert's picture

implies that these headphones don't just need more watts. They need more gain and higher voltage and definitely no series resistance.

ednaz's picture

...of five foot tall left and right headphone cups sitting where my speakers sit. Not pretty.

DH's picture

Use a system like DEQX, which is one box that both time aligns your drivers and does DRC; or buy one of the new speakers like the Kii Three or Dutch&Dutch - both of which do much of the same on their own, and include amps matched to the drivers in the speakers.

None of the above solutions is a million dollars of weighs 250 lbs.

Many of us don't enjoy either the sound of headphones or wearing them. That's the LAST solution I'd go to for enjoyable listening.

Everclear's picture

A planar magnetic loudspeaker like the Magnapans may come close to the type of loudspeaker HR is describing ....... similar to the planar magnetic headphones .........

Everclear's picture

Sorry ......... Correct spelling Magnepan ...........

Everclear's picture

Also, electrostatic headphones and loudspeakers have similar properties .............

Everclear's picture

Electrostatic transducers do not have uniform resistive load, but for headphones and some loudspeakers, they are crossover less designs .............

Herb Reichert's picture

What I am lobbying for are speakers with high ruler-flat impedance, high sensitivity, minimum phase, and minimum excursional non-linearities ----- like good headphones.

To me, speakers with reactive impedances below 2-ohms are broken and non-functional -- no matter how good their frequency responce measures

Duck851's picture

*Edit – (I'm assuming you mean 'Grado' nor 'Grade.' If I'm wrong, I'll change it back!

Didn't Grado whip up a pair of speakers that used their headphone drivers?

Duck851's picture

Yep, Grado, not Grade - thanks! Do you recall them?

canuckmgh's picture

Cube Audio, Nenuphar floorstander, single 10" driver with three whizzers, sensitivity 92dB, impedance 6 Ohms, claimed freq. resp. 30Hz - 18kHz (6dB). Reviewed elsewhere and given its top award.

Alex Halberstadt's picture

genius, Herb!

Everclear's picture

Probably the appropriate title of the article is "Loudspeakers, the PITA (not the bread) devices" :-) ..........

kenyanrandrews's picture

I am a noobie to this game so forgive my ignorance of the measurements and science - much of this article was not understandable for me.....but.... I have just bought a Musical Fidelity A1 amp to take the place of the Onkyo TX8270. Powering small bookshelf speakers. What I hear is less resistance, higher energy and less digital trickery leading to a WAY nicer sound from the simple A1 amp. Is this the gist of your experiment?

rt66indierock's picture

I can think of three off the top of my head frequency range, intermodulation distortion and frequency mixing.

And to shoot the messenger you supported MQA I have to question your judgement.

Genez's picture

That is why I refer to my Audience 1+1 V2's as my 'off head' headphones.
I simply adjust the bottom with some linear dynamic EQ to room acoustics and enjoy. To achieve this effect these speakers need pinpoint, precise adjustments, in height,width,and angle. Then the magic begins.

MixerRog's picture

Having build & designed both Tube Amplifiers & Solid State amplifiers for years. I much prefer Solid State sound with a solid non boomy bass due to the very low output impedance of the Solid State Amp versus the much higher output impedance of a Tube Amp. A low output impedance causes a much higher damping factor as you likely know which in turn causes much less cone overshoot in the bass region. Tube Amplifiers in general have much higher harmonic distortion also.

I have told many people if they want to have a tube sound with lower harmonic distortion also, just put resistance is series with the speaker such as small gauge speaker wire or high power resistors.

Damping Factor is defined as speaker impedance divided by amplifier impedance so as an example, a solid state amp with a .1 ohm output impedance & an 8 ohm speaker would have a damping factor of 80 versus a tube amp with output of 8 ohms & 8 ohm speaker having a damping factor of 1! Tube Amplifier feedback can cause this the damping factor to rise as high as about 20 but that is even rare that I have seen while very good Solid State Amps can even approach a damping Factor of 1000!

For me after over 50 years of Audio work, the single biggest sound difference of Tube versus Solid State sound is due to the Damping Factor difference.