Amplifier brand/designer with the most unique and consistent “house sound?”

Different resistors, capacitors, wires endow amplifiers with unique and often predictable coloration/qualities.  Do these components vary in their distortion/phase attributes in predictable/measurable ways?  

Yes. Its also predictable that if a part such as a capacitor or resistor does not measure as well as another part of the same value (assuming that the values are identical) that when the inferior part is installed the circuit will not perform as well. Sometimes the differences are very slight, sometimes not, depending on where in the circuit the part is installed, since certain parts of the circuit are more critical than others. 

For example I've seen capacitors of different types but the same value have an easily measured difference in high frequency bandwidth. Resistors of course can have lower noise- and this is no secret or weird snake oil; Ampex used low noise resistors in critical points of their tape machines back in the 1950s.

If you get rid of distortion and phase shift across all frequencies, will all amps sound the same?  Should they?  Even with different topologies, parts?  No more house sounds?

If their output impedance is also the same then at that point, yes, they will all sound the same assuming their power supplies are up to the task.

I would not characterize doing this as a mathematical extreme. Its simply what you do if you want to get it right.


The key is getting enough feedback and at that point it won't matter if its tube, solid state or class D. This is almost impossible with tubes as there are usually too many frequency poles resulting in a low 'Phase Margin'. If the phase margin of any amp is exceeded when feedback is applied, the amp will go into oscillation. This is one reason why tube amps tend to have less feedback than solid state, the other being that they are also inherently more linear and don't need that much feedback to have acceptable distortion figures.


Designers who have been up to the task of installing enough feedback in traditional solid state designs are rare. But it appears that the Benchmark and the Soulution are two amps where this is exactly what has been done.


In class D this is possible though a different means which is to add so much feedback that the phase margin of the amp is exceeded, causing it to oscillate. The oscillation is then used as the switching frequency. Such amps can employ quite a bit of feedback and the sound of them is really all about how well the power supply is built.

That being said, are there other factors aside from harmonic and intermodular distortion that might account for different house sounds? Are there other factors which make different amps sound different, even with negligible distortion of any kind?

If the amp is not using enough feedback then bandwidth and related phase shift will also play a role. For example a rolloff at 50KHz can make the amp or preamp sound a bit dark. This is because the phase shift can go down to 1/10th the cutoff frequency. In the same way, a cutoff above 2Hz will affect how the circuit presents bass. If the bass has phase shift it manifests as less impact, up to 10x the cutoff frequency.

This is why wide bandwidth has been such a big deal. But if you have enough feedback, then you don't need so much bandwidth because the circuit can correct phase shift as well as distortion.

What a genius and I know that he voiced the Silver 7s 900 on Apogee Scintillas and the 350s were voiced on that ( I have a pair of Crimson, the only color choice they should have IMHO)) and now voiced the 275s on those. Great clean bass and extended highs. I believe but can't prove he voiced his amps based on his ear, not on so-called industry or acoustic specs, etc.

Its unlikely that the amp was 'voiced'as you describe. As an amp manufacturer I hear about this sort of thing a lot. But 'voicing' really isn't a thing; if you compensate an amp to work a certain way with a certain speaker, you've created a synergy. Another way of putting this is you have one distortion put in place to offset another distortion (and probably one of an entirely different type), which results in compromised performance. So I suspect what is true is that Carver simply used the Apogees to listen to the results of his work. You do the math to design the circuit, but you'd be an idiot if you didn't also put the finished circuit on the bench and see if it needs any tweaking, since the math and the actual parts often aren't the same thing. After it performs well on the bench then you play it and see what you think. But this process isn't voicing because you aren't making changes to make it work on that speaker. You're making changes so that the measurements show you its worth a listen.


Now there is another issue, that of the spec sheet as opposed to what we hear. If you are pragmatic as an engineer, you know that the rules of human hearing are not entirely taken into account on the spec sheets. This is why some amps can measure poorly but seem to sound just fine. Robert Harley commented about this in an article (about amplifier types) for Stereophile a few years back but made the comment (about SETs) that 'no-one knows why' they can sound so good and measure so bad. If you've been reading what I wrote here and in my post above, you can clearly see he only made that statement out of an incorrect assumption that designers didn't know why that can be so. Its certainly possible and advisable if you want to make a listenable amplifier to keep the human hearing rules in mind and design for them rather than spec sheets that ignore how we hear.


(One way spec sheets ignore human hearing is showing a low THD number which looks great on paper. But if that low THD is entirely higher ordered harmonic content the amp is going to sound pretty bad - it will be bright and harsh. If you want the amp to sound right, you have to either keep the higher orders way way down or have enough 2nd and 3rd harmonic to mask their presence. You also **at the same time** must have the same distortion percentage at 100Hz, 1000Hz and 10,000Hz. Many solid state amps have excellent distortion figures at 100Hz, which is why they are usually measured at that frequency. But if you measure at higher frequencies, like 10KHz, you find that the distortion is higher. Amps like this are brighter and harsher on this account, even though they have 'low THD' on paper. This phenomena is caused by insufficient Gain Bandwidth Product to support enough feedback, so as frequency goes up the feedback amount goes down.) 


Now if you're going to say 'that's voicing' when the idea is perhaps to design to keep the higher ordered harmonics inaudible then I'm going to agree, although I simply see it as measuring properly. But that still isn't done for a particular loudspeaker.

So some like more 2nd harmonics and says that the amp sounds great, when others will even almost not want to turn it on to listening to the same amp, when it has according to measurements not and them to high distortion (as harmonics are).

So House sound is the amount of distortion/harmonics. Or maybe it can be both deviating from flat line AND have more or less harmonics?

The ear treats all forms of distortion as a tonality. The lower orders (2nd, 3rd, 4th) bring richness, bloom, lushness, that sort of thing. The higher orders bring brightness and harshness. The ear is keenly sensitive to the higher orders and really not so much to the lower orders. The ear is more sensitive to higher ordered harmonics than anything else since it uses them to sense sound pressure. For this reason a good amplifier/good system will be low in this and will not sound 'loud' even when it is. The ear is also sensitive to intermodulation distortion; the ear assigns brightness and harshness to IMD.


The 2nd harmonic brings richness, if in sufficient amount to mask the higher ordered harmonics. Amps that express the 2nd as their primary distortion product mathematically have a 'quadratic non-linearity'.

The purest form of this in amplifiers is an SET.


The 3rd harmonic is treated by the ear the same as the 2nd. Amps that express the 3rd as the primary distortion component have a 'cubic non-linearity'. Amps like this tend to have less distortion overall, since as the order of the harmonic is increased, their amplitude falls off at a quicker rate. So you don't need as much 3rd to mask the higher orders, and amps of this type tend to have 10th the amplitude of the primary distortion as opposed to amps with a quadratic non-linearity. So they sound more neutral.

Amps of this type will be fully differential from input to output.


Amps can have both non-linearities if single-ended and push-pull circuits are combined. Algebraic summing occurs, so amps like this tend to also have a prominent 5th harmonic.

Amps with a single-ended input and push-pull output usually fall into this last category, although it is possible to build an amp that has a single-ended input but is fully differential from input to output.

The use of feedback adds complexity to this picture. Traditionally, feedback suppresses distortion but adds some of its own, all of which tend to be higher ordered harmonics (plus some IMD), so it can be quite audible as brightness. If insufficient feedback is used, essentially it will reduce frequency response errors when the amp is driving a speaker since the output impedance will seem lower. But it will also be brighter due to the added distortion. Apparently the ear favors tonality brought on by distortion over actual frequency response. 


If enough feedback is used (over 35-40dB) then the amp can compensate for the distortion caused by the feedback (this also allows it to correct other things, like phase shift). Because this is a bit of a trick, and because solid state amps in particular have needed feedback to operate correctly, we've seen about 60 years worth of harshness and brightness in solid state amps, which is why tubes are still around.

But recently there have been advances in semiconductors and also (more importantly) the will with designers to create **in fact for real actually** neutral amps that have almost no distortion signature (house sound) at all. IOW there are now amps that in fact have enough feedback, something new on the audiophile scene.

In a nutshell, a 'house sound', IOW a distortion signature, should be seen as a bad thing, whether its the ever-loving 2nd harmonic of SETs and many other tube amps, or the brightness and harshness of traditional solid state.


One way to minimize these signatures is to allow the amplifier to drive a load that allows it to make less distortion. In all cases this will be a loudspeaker of higher impedance. This is easily seen in the specs of all amplifiers! The speaker should also be a reasonably good efficiency, so the amp isn't being asked to make full power. SETs are a special case of this; if you want to get the most out of them you need a speaker that will never ask the amp to make more than 20-25% of full power (higher ordered harmonics show up if this rule is broken). Solid state amps don't have the graceful clipping character of tube amps (solid state amps made with Static Induction Transistors being the one exception, but SITs are long out of production) so they should always have more power than would be asked of them as well, but 10% more is often enough.


So IF sound quality is your goal, to get the most out of *any* amplifier a speaker that is higher impedance and higher efficiency should be used. If you really like the 'house sound' anyway, keep in mind its not neutral and by definition less musical, even if the amp has a euphonic character.