The differences in the chip's noise figures themselves? You cannot hear them. I doubt you can hear them at a gain of 100. And unless you've got one serious scope, you can't measure them... A couple nanovolts, or even microvolts difference?
It's configuration specific. I can't tell you what you'll hear in advance of swapping without knowing the total layout, especially the gain and the resistor values, topology, layout, the position of Mars in relation to the direction your house faces, etc...
Sometimes, it's nothing correct but something wrong you hear... that "air" sound that people call hiss, it CAN be high frequency oscillation, really! I didn't believe it the first time I experienced it either.
Also take into account bandwidth... if a given opamp is dropping back to unity around 30kHz full bandwidth, and another is dropping back around 210kHz, which do you think will sound quieter? Technically you can't hear either, right? But test it. You can.
Also if you're controlling the bandwidth yourself, with compensation pins or a shunt capacitor across feedback resistors... you can control which sounds better, or at least "quieter" as you say. (or a multiloop feedback scheme, same diff)
Just dropping in chips, you might not have a great experience... but tailoring each design to a
specific set of critera (volume needed, load, gain, source, etc)... then you can optimize it so that different chip swaps affect the sound less and less...
This is why you gotta respect commercial designs - these engineering guys are designing amps to be everything to everyone... and that cannot be done. Getting close, even marginally close... is godhood.
Which is the best? Heh. I won't go there.
