I have zero experience building gear as a hobbyist, so although I never doubted in theory that passive parts could make audible differences, it was only when I got a phonostage with plug-in sockets for changing loading resistors that I was able to hear this for myself. After settling on the best loading value by substituting with decent but inexpensive resistors, I got some recommended premium resistors in that value. The sonic difference was far greater than I had imagined it would be. A resistor seems like the simplest of parts, and I wasn't sure exactly what I would hear when I sat down to do comparitive listening tests - I was mentally preparing to have to strain to hear any difference at all, maybe a touch more transparency, I guessed.
The expensive resistors did offer that (more clarity and resolution), but what surprised me most - and I was totally not prepared for - was that the premium jobs revealed the cheapies as being not tonally neutral. I don't know whether I would have even heard the colorations of the inexpensive resistors had I not done the comparision, but if I had, I sure wouldn't have thought to lay it at the feet of one measely resistor per channel. And the transient articulation, dynamics, and imaging palpability improved too. After getting a taste of the better stuff, the cheap resistors (and they weren't bottom of the barrel by any means) just sounded a distinct bit veiled and artificial in comparision - they simply didn't get out of the way like the premiums did. Now multiply this over the number of resistors in an entire circuit (OK, I know it's not quite that simple, but still...) and you begin to see how this sort of thing could matter quite a lot.
The expensive resistors did offer that (more clarity and resolution), but what surprised me most - and I was totally not prepared for - was that the premium jobs revealed the cheapies as being not tonally neutral. I don't know whether I would have even heard the colorations of the inexpensive resistors had I not done the comparision, but if I had, I sure wouldn't have thought to lay it at the feet of one measely resistor per channel. And the transient articulation, dynamics, and imaging palpability improved too. After getting a taste of the better stuff, the cheap resistors (and they weren't bottom of the barrel by any means) just sounded a distinct bit veiled and artificial in comparision - they simply didn't get out of the way like the premiums did. Now multiply this over the number of resistors in an entire circuit (OK, I know it's not quite that simple, but still...) and you begin to see how this sort of thing could matter quite a lot.