Solid states more accurate than tubes?

What are some of the speakers that sound especially accurate and good with Atma-sphere OTL amps? Some examples would be helpful to illustrate you're point.

Quad ESLs, Sound Lab ESLs, Classic Audio Loudspeakers (I have T-3s at home), Audiokinesis, Harbeth (some models work better than others, check with us on that one), ProAc, Merlin,  Spendor,  JM Labs,  Most single-driver full range loudspeakers like Lowther, PHY and so on, Sonus Faber (as long as the speaker employs a single woofer), JBL, Altec and many other classic horn systems, Avant Garde, Vandersteen (usually the older ones- some of the newer ones are harder to drive), DeVore Fidelity, Voxativ, Spatial Audio, PureAudio Project, Salk... I can go on for quite a bit! Especially with our smaller OTLs, what you want to look for is that the woofer impedance curve is benign and nominally 8 ohms. So many 2-way speakers work quite well.


What we came to conclude is that flat frequency response is meaningless if the amplifier imparts audible coloration. A good example of that is harshness and brightness caused by low level high ordered harmonic distortion in solid state amps. Its hard to call that 'neutral' when everything you play through it is bright... The reason for this is the ear converts all forms of distortion into tonality (assigning 'bright' and 'harsh' to the higher ordered harmonics) and will favor that over actual frequency response, to such a degree that getting the speaker to be perfectly flat (which appears to be a Sisyphean task anyway) is pointless if the amp causes colorations due to distortion.

If you can keep the harmonic distortion to the lower orders the distortion will be less audible (as the ear is far less sensitive to the lower orders), and there are ways to do that without feedback (the latter of which is often rightly blamed for brightness and harshness in most amplifiers- the reason that is the case is a topic for another thread). If you promote linearity in the amplifier you can achieve fairly low distortion without feedback. Our amps at full power can be as low as 0.5% THD with about 0.05% or less IMD, which is better than a lot of tubes amps running feedback.  If you can get the IM and higher ordered harmonics down, that is how you get to 'neutral'.


That is what we've been doing for the last 45 years, but that has also meant that we have to vet the speaker to work with our amps. That's really not all that hard which is why we've been around for so long.

but then those easy to drive/efficient speakers themselves are not the most accurate things to listen to.

This is highly debatable! Many of the Wilson speakers made in the last 20 years are very easy to drive. An easy load to drive has a serious advantage- less thermal compression. A horn can be properly designed so that it has smooth response, and has controlled directivity so that early reflections from side walls won't cause harshness. The idea that only hard to drive speakers can be accurate is a topic for another thread, but also obviously false.

Many people take the brightness of traditional solid state as detail but in fact most of the time its simply distortion. If the amp is truly low distortion at all frequencies then it will be more detailed but also relaxed.

If we are talking about preferences, that is another matter entirely unrelated to what is 'more accurate'!

The problem here is what solid state amp vs what tube amp.

And this has everything to do with Gain Bandwidth Product, which is a factor that can limit how much feedback an amplifier has. For this reason, IMO traditional solid state amps are not more accurate, as nearly all of them in the last 60 years have been brighter and harsher than real life.

But tube amps also have insufficient Gain Bandwidth Product; what has saved them and why they are still around so many decades after being declared ’obsolete’ is that they tend to make lower ordered harmonic distortion, which is fairly inaudible to the human ear, as opposed to the higher ordered harmonic distortion of solid state.

OTOH if you can get enough gain bandwidth product, then it is possible to build a solid state amp that finally has enough feedback, such that at higher frequencies distortion is still controlled. In terms of traditional solid state amps you can count on one hand the amps that have succeeded in this way- the Benchmark amp appears to one of them. There are also class D amps that do too- because its relatively easy to get gain with class D and you don’t have to worry if the amp goes into oscillation- in fact in a class D amp, that would be something that you want!


Its this oscillation issue that is why prior amp designs have had insufficient feedback: add too much and the amp will go into oscillation at some frequency, owing to the phase margins of the amp being exceeded. Because that generally isn’t a good behavior on the part of an amp, lessor amounts of feedback are used to promote stability. But that means the amp won’t have enough feedback at higher frequencies and so it will sound harsh. But nevermind, many manufacturers have a simple solution for that called ’lying’.


This is why feedback has gotten a bad rap (I’ve not yet mentioned that its application results in distortion as well; see the writings of Norman Crowhurst from 50 years ago...). A pragmatic designer might instead build an amp with no feedback at all, generally designing so it makes the distortion that is relatively inaudible and doesn’t make the distortion that is. At least that way you avoid the gain bandwidth issue! And there are ways to minimize distortion without feedback; some zero feedback amps have some pretty impressive distortion figures.

So it really does make a difference of which amps you’re talking about; ignoring all this simply means that the blanket statement

Solid states more accurate than tubes?

is simply false and nothing for it.