Why you can’t trust measurements

[stereogeek]

I only trust my ears. If It sounds good to me then nothing else matters. I know my speakers don't test very well but I bought them anyway. Why would I buy sketchy testing speakers? Because folks that have heard them love the way they play music. So do I.

 

[CPP]

I don't listen to measurements

 

[atmasphere]

You rather misunderstand my points (and perhaps I do yours as well). As to "proving" it, that's not really possible for one person (listener) to do. I could convince myself that it is true for me, but that is kind of my point. A careful listener with good hearing and discrimination (and in my experience this unfortunately applies to few "audiophiles", even reviewers) can learn a lot about his own preferences, which is all that matters when choosing his own components. However, this doesn't "prove" anything in a more general sense, or in the sense of "evidence" as verified by experimentation and appropriate analysis.

I would like to read about more investigation into why 2 amplifiers (for instance) which measure similarly in whatever ways one deems relevant often still sound different, and what measurements might better explain this.

A good example of how the measurements can explain how two amps can seem to 'measure similarly in whatever ways' is if the higher ordered harmonics are not masked by the lower orders, that amp will sound bright. Yet on the bench, they will both measure perfectly flat. Its not a frequency response error- its the ear converting the distortion to tonality, also on account of how sensitive the ear is to the higher orders since it uses them to sense sound pressure. So if you look at the measurements and see a higher 2nd and 3rd harmonic, that amplifier will likely not sound as bright. Again, this is why tube amps can sound smoother and less bright than solid state, even though they actually have more higher ordered harmonics and otherwise have similar bandwidth. The lower orders are masking the higher orders.

The brain has a tipping point where tonality induced by distortion is favored over actual FR. If you know what to look for in the measurements (and the measurements actually exist...) then you can predict which amp will 'sound' smoother.

 

[rbbert]

Often?? Name one such non-imaginary, controlled instance. Unless you mean this type measurement
Just a moment...

Ignoring (for now, at least) my hypothesis, what amps do you use to test your speakers?

 

[AJ Soundfield]

The NCores in the Fusions, mostly.

 

[rbbert]

The NCores in the Fusions, mostly.

What does the distortion spectrum look like for those? Is it safe to assume that the speakers' impedance curve is easily handled by these, or are all your speakers fully active?

 

[AJ Soundfield]

What does the distortion spectrum look like for those?

Well below any human hearing capability before outright clipping. Not linking ASR, but there are innumerable online measurements of the NCore (amp/ps) modules, 99.9% of which are just modules placed in fancy cases.

Is it safe to assume that the speakers' impedance curve is easily handled by these, or are all your speakers fully active?

The amp sees only the driver impedance, which is far less complex/reactive than "speaker". Not all, as mentioned, mostly. I have amps for testing luddite passives, both SS and high output impedance SE's etc., but 99% of what I do is multiway active. If you know how to do a variable directivity. cardioid bass, room mode adaptive passive, I'm all ears.
Btw, my question was rhetorical. The Carver Challenge showed nicely a 70db null is more than sufficient for physical reality. Now for the imagination....

cheers,

AJ