Back to no-feedback, non-inverting, Class-A architectures (which I think the Valhalla is a good quality example off), Solderdude mentioned that linear behavior is difficult given the lack of FB. However, the Valhalla seems to have pulled a decent job with the HD600 which has a relatively large impedance. What is it about such architecture that seems to perform well with some high impedance/hi-fi cans, but no so with lower impedance/lo-fi ones?

I know I'm reaching back fairly far into the post stream.

I don't think this is a global feedback related issue. And, I'll bet that both of these amps have multiple forms of local NFB.

It is a class A issue, but that's not all.

Respectable linear behavior without NFB is not that difficult if the operating and loading conditions are right. It becomes much more difficult when the amp needs large excursions of either voltage or current or both around the various device operating points.

One of the tricks of amp design is to keep these excursions as small as possible in the first stage so that the THD generated there is small and not amplified by the following stage(s). Obviously, at some point you have to bite this bullet and, consequently, generate more THD. But, the later in the circuit you can do this the better (most of the time). And, in many cases, the distortion products of the output stage swamp even the amplified products of the input stage.

It is not the Zo of the amp that is the principle issue either. As I understand these amps, it's simply a loading issue.

At high headphone impedance the output stage is not required to deliver much current and so, while it may be swinging a lot of volts it is not moving that much current. It operates in a more linear regime which may be linear enough to reduce the distortion products to below or near audible threshold.

But, as the load impedance goes down the O/P stage is required to deliver more and more current. For most devices this significantly increases the distortion products and, in this case, seems to push them into the audible, but maybe not discretely detectable range (that is it just starts to sound a little worse, but you can't pick out exactly why).

At some point the amp begins to approach its class A limit and as it does it becomes much more non-linear until it clips or current starves.

So, if I may return to my earlier thought, the tube topology of the Valhalla was probably designed with these issues in mind (without presuming to speak for its designer). It is excellent for its targeted circumstances. It works really good for less optimal conditions and, like every other amp, breaks down outside of its design envelope. In this case the amp seems to give in gracefully.

None of this behavior is exclusively the property of tubes.

What is the property of tubes are the specific non-linearties of the chosen tube(s) and how they create the THD as the O/P stage leaves its "linear" regime. These non-linearities determine how the distortion spectrum grows as the output levels are increased. They also determine what the "less good" sound is like.

Outside of this topic, i do agree about the psychoacoustics.

I think this obvious in other circumstances not related to audio/music. For example, someone may bring us some news. The news may not be clearly good or bad, just news. Maybe it's just facts. Depending on our frame of mind at the moment we will react, maybe negatively or positively. The news is just information, but our brain makes a whole set of connections (logical and emotional) that we notice. And if we are in a really bad frame of mind when the news comes we can easily not even hear parts of it and so misinterpret it.

The news is information and we process it in complex ways. Music is nothing more than information. What we hear depends on how we process (from ear to brain to chemicals). Human hearing has, more or less, a mean which gives us a common ability to hear the same things in about the same way. But there is enough variation around this mean (both in time and across different people) that the receiver of the information can become nearly as important as the transmitter.

Good audio design might turn out to be analogous to what researches have learned about faces. IIRC, there is a perfect face that is seen as perfect by nearly all people. I think they call it the "mask". The closer a face conforms to the mask the more beautiful it seems to be and the more it deviates, well...it is the way the human mind works.

Maybe there is a "mask" for audio that works universally. The job then, is to find it.