Back in my ill-spent youth, we'd balance and "blueprint" engines to get more performance and reliability. Blueprinting involved tightening every bolt to exact torque specifications, keeping all clearances to spec, etc. Balancing kept vibrations our of everything that moved. The idea wasn't that any single bolt being a bit too tight or loose or any single spark plug gap being a tad too narrow or wide would have much effect, but an engine that had everything done right would operate better than an engine with a lot of tiny imperfections. We called it tolerance stackup.
As I read these threads, I see references to very small improvements or differences in each individual component, many below the ability to hear. But I wonder if they all add up to something that is within our hearing.
In building a solar system, especially a low-voltage system, it's critical to keep all connections to a minimum because each connection drops voltage. It's critical to use properly sized wire because wire that is too small drops voltage. No single connection is all that critical because the voltage drops are usually very small, but they add up. With solar, we can measure the voltage drop exactly and so there is no debate.
But with audio, I'm not seeing any way people can measure the pleasure factor of music. Yes, I see frequency-response charts, but they don't always correlate to the sound we're looking for. Is there a way to measure warmth, transparency, detail?
I read in Robert Harley's "High-Performance Audio Systems" that THC, while certainly capable of being measured and is certainly a good design goal, is often a misleading piece of evidence in the actual sound being reproduced because certain feedback designs can reduce distortion while degrading the actual sound quality. Some amp manufacturers would design for enhanced specs so their equipment could be more easily marketed, even though it didn't sound as good.
It's interesting to see that people in double-blind tests don't seem to be able to find improvements in individual component improvements, I wonder if they can reliably hear differences when all these improvements are lumped together in a "blueprinted" system.