Honestly, that Japanese CD sounds like it has a high end eq boost at minimum. It should be able to be bettered. This new set is not it though. If the tape was in rough shape or something like that... I don't know. I hear more of the recording and with less damage in the 1st Japanese CD edition. I mean... I obviously don't know that because I don't have the master here to A/B with! Perhaps I agree with an eq alteration done on that CD version?
All I know is this new copy sounded muffled and unbalanced frequency wise. And then it was boosted and limited for no good reason and when it had actual problems that needed correcting first. And then the goofy artifact soup upmix. This is just childish.
Whatever generational madness happened or didn't behind the scenes, that 1st CD has the most naturally clear sound and dynamics. I'm not normally a fan of the hyped high end eq boosts that are in vogue for CD versions.
That SACD version sounds even worse. I don't care what DR numbers someone made the meter read. (You can make those read anything at any volume. Check out TV or radio commercials. You can skew LUFS numbers almost as easily. Almost...)
So much speculation. Or, there's facts.
1990 = Japanese CD
2014 = hybrid SACD
2021 = latest remaster
These are data for the entire album considered as a single track (the SACD rip was downconverted to 44khz/16bit to make the measurements fully comparable)
Waveform view : 1990 blue, 2014 red, 2021 green. Why so skinny, 2014? it's because there's only one clearly highest 'peak' value in that one (in the track 'Solid Space'), aka old school mastering, versus the other two. Typically meaning: more dynamic range. Though neither of the other two are 'loudness wars' atrocities by any means.
Numbers. Include those darn untrustworthy Average RMS and LUFS numbers
| 1990 | | | 2014 | | | 2021 | | |
Peak Amplitude | 0.00 | 0.00 | | -0.04 | -0.13 | | -0.10 | -0.10 | dB |
TRUE Peak Amplitude | 0.16 | 0.01 | | -0.02 | -0.10 | | 0.09 | 0.17 | dBTP |
Maximum Sample Value | 32767 | 32767 | | 32634 | 32267 | | 32393 | 32393 | |
Minimum Sample Value | -32768 | -32022 | | -30642 | -29698 | | -32394 | -32394 | |
Possibly Clipped Samples | 7 | 1 | | 0 | 0 | | 0 | 0 | |
Total RMS Amplitude | -18.21 | -18.49 | | -21.52 | -21.49 | | -17.43 | -17.12 | dB |
Maximum RMS Amplitude | -7.69 | -8.36 | | -10.37 | -11.00 | | -6.70 | -7.13 | dB |
Minimum RMS Amplitude | -50.10 | -49.97 | | -96.57 | -88.75 | | -96.48 | -96.62 | dB |
Average RMS Amplitude | -21.28 | -21.47 | | -24.95 | -25.02 | | -20.87 | -20.61 | dB |
DC Offset % | -0.64 | -0.99 | | 0.00 | -0.07 | | 0.00 | 0.00 | % |
Measured Bit Depth | 16 | 16 | | 16 | 16 | | 16 | 16 | |
Dynamic Range | 42.41 | 41.61 | | 86.20 | 77.75 | | 89.78 | 89.49 | dB |
Dynamic Range Used | 35.25 | 32.25 | | 59.80 | 62.55 | | 62.50 | 62.65 | dB |
Loudness (Legacy) | -15.31 | -16.08 | | -19.72 | -19.48 | | -15.06 | -15.01 | dB |
Perceived Loudness (Legacy) | -9.61 | -11.38 | | -14.57 | -14.08 | | -10.46 | -8.64 | dB |
ITU-R BS.1770-3 Loudness (LUFS) | -14.59 | | | -17.79 | | | -13.65 | | |
0dB = FS Square Wave | | | | | | | | | |
Using RMS Window of 50.00 ms | | | | | | | | | |
Account for DC = true | | | | | | | | | |
Conforming with the darn untrustworthy (/sarc) numbers above, replaygain indicated that that the 2019 and 2021 were perceptually louder than the 2014 SACD, and would need to be lowered by 3-4dB to 'match' it.
So, I used the replagain values to adjust overall levels for the two louder ones, to level-match to the 2014. Then , a frequency scan (Blackman-Harris, FFT=65536), graphed on a logarithmic scale (reflecting how we hear); the Y-axis is dB the X is frequency
For much of the midrange, these three are similar -- they overlap almost completely (though not identically). But note here the bass and treble. Below 60 Hz the 1990 JCD (blue) bass begins to drop off (has a lower level) significantly compared to the other two. The 2021 (grey) bass also drops off little bit compared to the 2014 SACD (orange), notable at 40Hz and less. Above about 4K the 1990 (blue) is trebli-er than the other two, and above about 2021(grey) has *less* treble than the 2014 (orange).
Details? Here's a closer look. This graph is the *difference* between 1990 (blue) and the 2014 SACD (black line at 0)
yep, there's the comparably steep bass drop off (JCD sourced from an LP-equalized master, you think?) , and there's the treble rise above (compared to 2014), actually starting near 6K, and showing a 'hump' between 8-9K. Also some lesser bumps and troughs in between (which should be audible too, in a level-matched comparison) .
And the 2021 (grey)?
yep, there's the smaller bass drop below 60, and the treble drop starting around 10K (compared to 2014).
Those aren't radical difference -- none of these mastering sounds 'bad' to me , 'badness' is mainly in the mix itself -- but they are all differences you can hear. Stuff people speculate about, maybe not so much.