The subjective measurements by listening are rubbish AFAIK. How do you subjectively cover the whole frequency spectrum? What is the reference point for the whole spectrum? And this is my POV. Need not be others.
For you, they probably are. Your question is indicative of that. But there is a parallel with measurements as well. Putting so much faith in measurements means you're making assumptions and generalizations just as well. The sad truth is it doesn't tell you how it is going to sound. Just like when people here are saying they measure first and tweak after.... do you get the part where they tweak after? Which measurements are they using then? Do you notice they skip over the aspect of how their tweaks will affect change? How do they maneuvre according to prediction ? Swinging in the dark?
What If I asked you to show me a decoupling network that sounds good on paper or simulation. Would it be complete enough to seem realistic in practice? Would it measure as per simulation? How do you correct it??
What if your measurements are telling you that you have a flat and clean noise floor, and you tell yourself that it's low enough not to be audible with a little hand waving about the capability of CD audio encoding or the realities of a 32bit dac only likely equating to 16 in practice unless you're really good at what you do you might get up to 24 but probably not.
I'll tell you the real danger is in thinking it can only be achieved a single way, and very unlikely with a diy hobbiest have either the understanding to use measurements properly or to extract meaningful information from them.... let's be honest, most professionals don't. They don't have to because they are going to sell you junk and they know it. Then they'll tell you it's the room that's to blame.
But that mistaken belief you share all but rules out any hope for a hobbiest to get anywhere. Look what the guru's do in audio anyway. They dove off the highboard head first into a dry deep end of euphonic systems because when they tried to make an accurate one it was "dry" and "analytical". To them the only way to get "music" is with painted layers of euphony which means it only sounds good playing one song and any other type of music will always sound like that same song... horrible. Not only have measurements failed them, but their ears have as well.
If you want the ultimate in lifelike realism that transcends any type of listening material then I assure you, you will not get there with measurements alone. Either approach takes serious dedication, methodical testing, and a whole lot of understanding. That all comes with experience, and plenty of second guessing. Interestingly, I've learnt that immediate and first impressions are often dead on.
Your reference is ultimately realism. In my last major upgrade to my DAW which I already generally described here, I was reducing the sick effects of high jitter according to my ears alone. Was it worth attempting to measure? No. I didn't have the equipment and had I made it a life's goal to acquire it and learn it I still would have been stuck with the compromised junk that I had and still would have had to resort to the same types of fixes that I did.
Because of an ill design which was far less than perfectly ideal, the flavor of the component itself came through in full force. I discussed my surprise with a very highly regarded digital audio designer and was told if there's any kind of noise leakage the nature of the components will show through despite it being a purely digital stage. Generally the same principles applied there as everywhere else. You're just not going to use garbage components that suffer high leakage and self noise. Do I need to measure it to now that I'm using probably the world's best component? Nope. The audio grade capacitor manufacturer took care of those oddball measurements for me and I'm more than satisfied with their efforts. Where I'm not, I figured out a way to capitalize on their advantages and compensate their weaknesses in order to fully exploit them.
I still tried half a dozen alternatives for comparison. "Baseline" simply becomes the junk you started with, and then it becomes the version you're intent on settling on which you compare everything else against.
From that, I focused purely on its ability to create a 3D image with realistic formation and very specifically the ability to actually move that focused image. The baseline for comparison then was very specific reference recordings. If it's off in any great way, it may sound like a completely different song. It is important to resist the beginner mistake of simply boosting base or some other narrow view... this comes with training and discipline. Focusing on image formation and moveability makes that rather easy. Fact is, once it is pulled together correctly, it is unmistakable. I may even use half a dozen 30 second sections of different songs, for each test.
Of course it is still a junk circuit at the end of the day in the sense that really nothing is ideal so there's always some necessary compromise. The goal is still to get the best of it. But when you get the balance correct, such that the image has indeed pulled together towards ultimate realism, then you find it's not such a compromise after all, because it does everything equally well.
Of course it is also all verified with actual test tracks. I defy you or anyone to make a system with measurements alone and have it reproduce a believable piano scale. You may think you've done well but that will break you every time. There is seriously nothing my system could play that would embarrass me in front of anyone at this point, and again, the joke is still that it's done in a way that people said quality audio could never exist.
These cries of subjectivity remain a farce. People have more in common than they think. Some are just a little more deaf to their understanding is all. People at wine tastings don't cry about subjectivity for example, and nobody chews rusted nails preferring it to a steak. Realism is realism and that's all there is to it. Let that be your baseline.
Just as well, if someone had to reproduce the mona lisa as realistically as possible, would they do it using auto cad? Maybe for part of it, but that alone would only give them a narrow representation.
Measurements are simply a human construct that we overlay on reality in order to help analyze it using narrow interpretations. Our ears do an impressive job of that already. It's more about the method than anything and I find that there is a method in common amongst most people responding here.
Also, you don't pay for failures in diy. You pay for lessons.