Okay, interesting (kinda) news.
I took a small chunk from the beginning of both LD tracks, normalized that chunk, cut away one single sample from the Japan LD's beginning for a better sync and then subtracted both tracks. To my surprise, there was a clearly visible and also audible waveform left. However listening to it revealed that whatever differences there are between the tracks, they must be pretty constant, because there was no fluctuation of any kind in how the difference signal sounded.
So I think, without any doubt, the two are the same master.
But, I wanted to find out why I was still getting a visible waveform, so I fiddled with spectral views a little bit and they looked pretty much identical, save for what I think must be dithering noise; the JP LD seems to have the tiniest amount of more noise in the high frequencies, whereas the US LD seems to have a bit more in the mids. Since this observation was so very minimal, I concluded that it must simply be slightly different noise shaping parameters for the dithering that caused this.
But I was also seeing some more noticable changes in the low frequencies, so I chose a strong logarithm that shows the low frequencies in more detail and I found that the US LD seems to, for some reason, "smear" under 20 Hz and also have some kind of constant undefined ultra low frequency hum that likely isn't audible, whereas the JP LD seems to have nicely defined low frequencies throughout.
I'm guessing they might have just run some kind of anti-rumble filter on the US LD that removes anything below 20 Hz.
Very likely this doesn't affect perceptible audio quality in any way, but explains why subtracting the waveforms didn't work properly.
Nevertheless I'd argue that this objectively makes the JP LD the superior, more "raw" track that may even theoretically sound better in some subtle subconscious way because EQing/highpassing with most EQs likely introdudes phase distortions, which might also explain why subtracting the tracks results in a track that still appears to have a rather full frequency range - phase distortions would prohibit a proper cancellation of all frequencies. However I have yet to do an A/B side-by-side test to see if there's actually any perceivable difference; I'm guessing there isn't. Still might be the better choice for any projects that aim to edit the audio or upmix it, as it's arguably a more faithful "raw" source. Thoughts?
You can see it best in a screenshot comparison:
http://www.framecompare.com/screenshotco...n/2F11NNNU
US LD:
JP LD: