First I would measure how well the existing pot tracks. A pot is simply a variable resistor. A stereo volume control is two pots on one shaft. To give good performance, the two pots should be at the same level of resistance at any point in the rotation of the volume control. This keeps the stereo balance accurate at all volume settings. Unless a pot is REALLY BAD, it is not going to add noise to the signal, and there is no difference is sound quality between a 99 cent surplus bargain pot and a $140 machined metal custom pot. The difference is that a $25 Alps Blue Velvet pot will have closer to identical values of resistance between the two pot sections (left and right gain controls) at various points along the pot's rotation. The Blue Velvet pots also have a nice smooth action that is pleasant to use in terms of ergonomics, and it is made of quality materials that should last longer without wear-related failure. A stepped switch with hand-selected matched resistors gives the most accurate tracking.
So, if the existing pot tracks pretty well, there is no sonic reason to replace it.
Some people claim they hear differences in resistor types. However this has never been shown in double-blind tests to the best of my knowledge. (Anyone knows of any such studies, please chime in with a reference- )
