Fair enough, even tough You’ll never be able to measure something properly if the device you use to measure have lesser specifications that the device you are designing,even with some cheat tricks.
Again, there are ways around the limitations of audio analyzers. Hell, even an APx555 isn't good enough to measure the distortion of modern op-amps.
Bob Cordell's distortion magnifier is a good example. Another way is to use a notch filter with a lesser-performing ADC like the QA401 or QA402. Obviously those notch filters aren't trivial to design, but in practice this solution works well. Is it a calibrated solution? No. Is it good enough for engineering use? Yes, although it's a lot more time consuming than an AP box.
But really that was not the core of my statement, there seam to be a feeling around here that products start to perform better because engineers started to measure their work. That’s not the case, every electronic designer need to know how it perform on objective metrics. Everybody measure their stuff, that’s an essential part of developpment. What happens is some prefer to not talk about measurments when they know they don’t score good on some metrics. I believe some engineers self compromise the objective performance of their stuff, for irrational belief in their own special way to do stuff that for them trumps the common metrics like SINAD, but the point is, everybody measures.
There are also a lot of engineers who hold the view that reducing distortion from 0.005% to 0.0001% won't lead to an improvement in sound quality because even if someone could pick out the difference, it's extremely unlikely that they'd be able to decide which is better. In a sense, it's the viewpoint that once a certain performance threshold is met, there's no point in trying to play the numbers game.
I'm not saying that this is right or wrong, but I wouldn't just assume that it's about an irrational belief. In some cases it is, but not all. And hey, if you can do it with discrete op-amps and still get 0.005% THD+N, it'll sound good
and give the marketing department something to talk about.
I myself like to see DACs and ADCs perform to this level because it makes them more useful, especially if they're capable of operating at 192k sample rates.