I wouldn't use a switch for several reasons:
1. I keep noticing that the differences between sources and between cables tend to be quite subtle sometimes, albeit still very important. Quick A-B-ing is not the best way to reveal them IME.
2. The logic behind using an A-B switch is something like this: if A>B (A and B are the components to be compared, > means "sound better to me than") then A+X>B+X is also true (where X is the sound impact of the switch and additional cables). This logic is flawed IMO. Just a crude example: say A and B sound very much alike, but A has better treble extension making it preferable. If switch X is rolling off the treble a bit itself, the advantage of A will not be heard anymore through the switch. Another example: A is a bit more refined and detailed, but also a bit boomy and PRAT is lacking compared to B, maybe because the bass of A is more extended and is exciting some room modes more, making it woolly and swallowing dynamics - so I'd definitely prefer B. The switch could reduce and tighten the bass a bit so that A+X (switch) sound more balanced and dynamic, while B+X sounds a bit bass light, thus making me make the wrong choice (A). Of course these are crude and oversimplified examples, but you get the idea.
I think your best bet is to listen to a song through A, than repeat the first 10-12 seconds, then listen to the same song through B, also repeat the first 10-20 seconds and finally listen again through A. I find the A-B-A testing much more revealing than the simpler A-B. This way you will quickly notice the main differences between the sound of A and the sound of B and then confirm and refine your findings through subsequent auditions (other, as different as possible, tracks to listen according to the A-B-A - or B-A-B sometimes, for a change - protocol).