Audio is a fun hobby for both listening and enjoying music with the thrill of connecting to the music but also for the exploration and experimentation of the equipment. I started out decades ago with modest gear making upgrades and modifications based on what I learned from others as well as applying my own engineering expertise. Making improvements in the sound is both rewarding and provides me with a sense of control. "I am the master of my own domain" sort of thing.
On the other hand, as I had the means I moved up to buying higher levels of hifi gear. That has its own type of satisfaction as well. Good gear sounds even better than what I was able to do on my own with tweaks and upgrades. But still, having that experience I believe, made me savvy as to which high end components were worth the price of admission.
The engineers in the business have many more resources and capabilities at hand to develop an audio component with much more performance capability than most of us. A good example of that, is the experience I had once while working with Ford engineers at one of their proving grounds in the early 1990s. The Mustang development crew was working in the bays adjacent to where I was working. They had a very large stack of tires in their bay- many different brands. A pack of 5 or 6 Mustangs would go out to the tire test track every morning with new tires. It got to be annoying, I would hear tires squealing on that track all day long. At the end of the day these Mustangs would return with bald tires. I said, "Wait a minute. Weren't those new tires just this morning?" The engineers confirmed to me that they were. By the end of the week that large stack of tires was gone. These engineers had assessed the performance of a large group of tires on a sample size of 6 cars. And this was just the hot summer testing.
I'm not saying audio gear is tested with the same type of rigor but the point is the developers typically know what components work best in their gear but they also often have to fight with the bean counters to come to a cost/performance compromise. Knowing which components to replace with what is the real trick.