What will you give me if I can build 2 that sound different? It wont be hard considering that you have IGNORED the importance of output impedance and THD: both aspects which any test ever made to prove that everything sounds the same has mandated be withing vanishingly small tolerances
The nay-sayers argue that an amp should have a zero ohm output impedance. thats great, this is a headphone board, there are several standards which state 120ohm output impedances. Which is it to be? I could list a couple headphones that sound VERY different driven from varying output impedances.
Next the testers bring up THD. thats cool, you have effectively eliminated ANYTHING which runs without global feedback by allowing this standard to be imposed. come on, lets let the standard sag from the normal 0.001%THD MAXIMUM to 1%... If we follow a distortion scale which observes increasingly weighted values as the order increases 1% is not so bad compared to what some SS amps pump out (weighted by order of harmonic, as opposed to direct readings: objectiveists HATE this one, considering psycho-acoustics it gives a better indicator of perceived SQ this way, but its bad bad bad for marketing)
Here is what you are really saying/asking, expanded to include the caveats and conditions generally imposed.
All solid-state amps which have lots of global feedback, zero ohm output impedances, and similar THD of barely measurable amounts sound similar enough that no human can tell them apart by ear. Some SS amps should be excluded for failure to be electronically identical to the others and some tube amps can be included in our testing because they can meet our electronic standards.
You will kindly note that these normally imposed restrictions GAURANTEE that there are differences in amplifiers which fail to meet the standards of electronic measurements. Why else would they be so restrictive when there have been very good sounding amps made which dont measure up to their purely electronic standards?
This is like me taking a sample of 10 Olympic sprinters as they run for their country, and saying that since all of them (100% of my sample set) can run the 100M in 12 seconds or less 100% of the people in the world can. You cant run the 100M in the olympics if you cant do it in 12seconds or less (maybe 13, you get the point) I have restricted my sample to people who already meet the point I am trying to prove! its not even close to scientific.
Further expanded, the REAL question asked of this WHOLE thing (and cables, they include the requirement that the cables be indistinguishable by electronic testing) is this: is the human better than the machine at discerning SMALL objective differences: the answer is of course no.
Is the machine capable of telling the human what will subjectively sound good: again no - it offers good guesses, and points out REALLY obvious flaws (even applies to gear which on its best day measures poorly by unweighted THD scales), but there are some sounds that can hide from the machine.
The human ear is something of a crude tool, BUT it excels at making many many measurements simultaneously which is a place electronics tend to fail abysmally.
now here would be an interesting test since "good SQ" is something of a subjective point, and electronics have shown some difficulty measuring that. It should be clearly evident that I would argue that amps which fail the standards imposed above can sound good so why not run THIS test?
A group of discerning listeners (audiophiles, well respected recording engineers, etc) could be asked to give a yes/no mark using double blind methods to whether various amps from a large sample set of all manner of amplifiers (vintage tube, new tube, new audiophile/pro SS, new consumer SS {older SS is a carp-shoot, throw some in anyways though}) sounded "good" to them. There would probably be SOME amps that were clearly "approved" by the vast majority, and probably a few that were universally hated. Take the best 25% or 6 (for the sake of making this test not take all week...) which were approved by the majority and apply the "do all amps sound the same" test to these. There will probably be more than 2 that are "approved by the majority", so break them out into several random pairs for the DBX testing. Now, it may happen that some of these random pairs accidentally meet the standards of electronic similarity imposed above! In this case, the results from THAT test should be ignored: it will be someone trying to hear differences that the above tests have proven cant be heard! this is not what we are testing...
When the group of amplifiers to be compared are selected by humans rather than machine, I would almost guarantee that other humans can identify differences in DBX.
what irritates objectivists is that EVERYTHING in this "do they sound different" test was agreed to sound good previously... and there will be very obviously different pairs in this test. sorry objectivists, SQ is subjective.