Would You Rather Own A Good SET Amp, Or A Great Push Pull Amp?

It surprises me that the THD is still a major parameter for measuring the sound quality. It is quite irrelevant to the quality of the sound as we perceive it. Even the formula by which it is calculated doesn't look relevant. I guess the engineers were unable to develop  a more adequate parameter for measuring the sound quality so far. I suggest this is the only reason why THD still remains a standard. 

So if you listen mainly listen to (an amplified) rock or electronic music, push-pull would be fine. But if you want to fully enjoy an acoustic music, you will do that better with a set amplifier.

@niodari This statement is misleading at best. Electronics really don’t care what kind of music you play. An orchestra is acoustic and it can play as loud as a rock band and can have just as much bass energy.

I don’t regard THD as having much to do with sound quality. It is a metric that is only useful after the distortion spectrum has been analyzed. 

For example the harmonic spectrum of a typical SET is pretty good; imagine what that would be like if it was the same spectrum but a 10th as much at full power. Distortion obscures detail, so an amp of the latter distortion character would be more revealing at any power level and more relaxed. A good PP amp can do exactly that. 

When doing analysis of any amplifier, IME/IMO its a good idea to see what harmonics are showing up at low power, at 6dB below full power and then see how the amp behaves when overdriven. 

Its also a good idea to look at distortion (and its spectrum) not just at 1KHz, but at 20Hz and 5KHz also. Tube amplifiers often fall apart at lower frequencies (SETs in particular); you put in a sine wave but the output can look pretty dreadful. So 20Hz response really can tell a lot about an amp. 

At what frequency does the output fall off? If there is no feedback, phase shift will affect tonality and sound stage presentation to 1/10th or 10x the cutoff frequency, depending on which end of the audio band we’re talking about. 

From those measurements you can get a good idea of how musical the amp can be. 

The thing that isn’t being discussed here is the vast difference in power between the SET vs the PP amp used. If you really want to hear what that is about the SET should be the same power as the PP amp. For example if you have a stock Dynaco ST70 compared to a 35 Watt/channel SET, you’ll find out really quickly how the two actually compare. 

Since high power SETs are not easy to find and really expensive if you do find one, its a bit easier to try to find a PP amp of the same power as the SET. If you can do that again you find the SET has no musical merit over the PP amp, if both are built to good quality construction standards (often PP amps of low power are not given the same quality of parts as a 35 Watt amp might have). 

Alternatively you could compare an SET against a PP amp which uses the same tube complement, in particular the power tubes.

If one amp has feedback and the other does not that too is a tremendous variable!

It is on the account of these three variables that most comparisons between SETs and PP amps don’t hold water; they are instead examples of a logical fallacy known as a limited sample size. 

@atmasphere (Ralph), ​My judgment comes not from a particular knowledge of electronics but from the practice, from my (humble) experience of using and auditioning audio equipment. Hence, it can be subjective. Since your judgment is clearly  based on your extended knowledge of electronics, it should be objective. But, as you have also mentioned in a number of different threads, measurement and theory do not necessarily correctly reflect how the sound is actually reproduced. Much depends on the models the measurement parameters are derived from (they can be misleading). A good example of that is the formula for the THD, as you likely agreed. In the light of your comments in this and the other threats, a more truthful parameter would take into account how an amplifier performs on high harmonics and how low harmonics mask the distortion in high harmonics, and not only at 1khz. It is not impossible to construct such a formula and create a new parameter, call it an Overall Distortion Measure (ODM) or whatever, amean for all tested frequencies or for each frequency range separately. Electronic engineers like you, could surely contribute to this.  

My comment that you call "at least misleading" comes from the earlier cited observations by @larryi  , which completely coincide with my experience using the PP and SET amps, though of different power (I comment just a bit later about my perception regarding amplifiers of different power). Based on your comments, these comparative observations may not be correct because of different power capacities that SET and PP amps typically have. Suggesting that the above observations are correct (see my note below), a simple implication is that a SET amp should manage better with the reproduction of acoustic music (that includes classical music, most of jazz and unplugged rock performances) since it does that in the "purest possible" way at a reasonable volume level. At the same time, for instance, there is no "natural" base in an amplified rock music and a tight and punchy bass can be more pleasing, just as an example. Regarding your note, although the bass energy in a symphony can be about the same as in rock, the base is still natural (acoustic) in any kind of classical music. 

Regarding the power differences in SET and PP amps that you addressed, I personally, very rarely raise the volume of my 5.5W SET amp more than 1/4 and my normal listening level is below that on any kind of music including rock or symphonic or music (and my Thiel CS 6 speakers are not sensible at all, they are at 87). In general, on all amplifiers I have and had in the past, I try to keep a similar volume level, at least below 1/2 even if for rock, independently of the speakers I use. To me, the distortion raises too much above a reasonable volume level,  independently  from how powerful the amp is; that is to say, on e.g., 1/4th volume level, I do not perceive too much louder the music played on my 600 watts SS amp than on the 5.5 watts SET amp on the same volume level. To my perception, independently of the power an amp has, there is more or less the same threshold level above which the distortion is notable, a reason why I don't care too much about the power in watts. 

But, as you have also mentioned in a number of different threads, measurement and theory do not necessarily correctly reflect how the sound is actually reproduced.

@niodari Actually I've said quite the opposite, although decades ago we could hear things we couldn't measure. Those days are gone as measurement technology has vastly improved. Apparently the education of what the measurements mean is still lacking.  

a simple implication is that a SET amp should manage better with the reproduction of acoustic music (that includes classical music, most of jazz and unplugged rock performances) since it does that in the "purest possible" way at a reasonable volume level. At the same time, for instance, there is no "natural" base in an amplified rock music and a tight and punchy bass can be more pleasing, just as an example.

Not sure what "purest possible" way might mean, but if you mean that an SET does that we'll simply have to disagree. 

At any rate there's no genre of music that does not have full spectrum although there are examples in those genres that do not. One has to be careful to avoid committing a logical fallacy due to a limited sample size. I've found that if you really want to get the system to sound right, one important aspect is to set things up so your amps are not working hard. This is very difficult to do with an SET because they only make about 20-25% usable power, above which, as we see in this comment:

on e.g., 1/4th volume level, I do not perceive too much louder the music played on my 600 watts SS amp than on the 5.5 watts SET amp on the same volume level. To my perception, independently of the power an amp has, there is more or less the same threshold level above which the distortion is notable, a reason why I don't care too much about the power in watts. 

-the higher ordered harmonics show up in greater amounts, imparting a sense of loudness due to the fact they are used by the ear to sort out how loud sounds are. So if your amp is generating more of them it will sound 'loud'. A sound pressure meter will show the truth of the matter. 

I've found it the mark of the best systems that they don't sound loud; instead are relaxed and detailed while maintaining authority at any volume. If you read a lot of comments by SET owners, this 'loud' issue is quite common, expressed in comments like '7 Watts is all I need as it plays plenty loud'. If you have clean power you will find yourself naturally and easily playing the system at higher volume levels because the artificial 'loudness cues' generated by the electronics are vastly reduced. 

If you are really intent on using an SET, you would do yourself quite a favor by using a speaker of greater efficiency!

I would look for something about 10dB more efficient- that will allow you to reach the same sound pressure with 1/10th the power and I think you'll find the greater resolution and dynamic character a nice benefit. Finding speakers that are 97dB these days isn't really that hard.

The second most important thing to do (far more important than finding the 'right' tubes for the amp) if you want the most out of the SET is to prevent bass information from entering the amp (I explained earlier why this is so important on this thread). You'll find the audible improvement immediate and obvious. 

Actually I've said quite the opposite, although decades ago we could hear things we couldn't measure. Those days are gone as measurement technology has vastly improved. Apparently the education of what the measurements mean is still lacking.  

@atmasphere ,  My memory could have failed in this issue. Though you agree that THD does not reflect the sound quality, are there other important parameters based on which we can judge about the SQ without the audition? 

If you have clean power you will find yourself naturally and easily playing the system at higher volume levels because the artificial 'loudness cues' generated by the electronics are vastly reduced. 

If you are really intent on using an SET, you would do yourself quite a favor by using a speaker of greater efficiency!

The loudness can be subjective. I was just auditioning a cassette  deck on two different cassettes on my 50 watts PP amp  with EL 84 output tubes. The first cassette is not "loudly recorded", so I had to rise the volume up to almost 1/2 of the max power. This was an optimal level since rising higher the volume did not really result in a louder coherent reproduction. The other cassette is recorder louder, so that I had to drop the volume at about 1/4 to get the same loudness as on the first cassette. Rising the volume higher had the same affect as with the first cassette at the level 1/2.  A natural implication here is that the distortion did not actually come from the amplifier:  The sound itself becomes "distorted" if it is higher than an acceptable (in this case, for me) level. So, it looks like a completely acoustic (non-electronic) matter. 

I played my 5.5 W SET amp (EL34 tubes) alternatively with 91db and 87db speakers. Honestly, I noted only a slight difference, much less than i normally note between two different CDs recorded at slightly different loudness levels. And the SQ with the large (less sensible) Thiel speakers is much better. 

I would love to audition a non-SET SS amp with the same clarity as my SET amp (at this point, I am a bit skeptical).