An engineering view on what's wrong with some remastering of Olivia's music

I was a broadcast TV studio engineer for a while in a previous life. So while I am not an audio professional I hope I have done the CD measurements competently. Unfortunately there's a lot of manual processing to get the right results linked to the right CDs, there is always the possibility of human error, in which case mea culpa1 and maybe let me know I goofed on the messageboard.

No Olivia album is going to be a straight transcription of the microphone signal to your ears - normally a smidgen of dynamic compression and EQ is applied to vocal tracks to sweeten them, and John Farrar had a penchant for reverb at times. You also have to add in the musicians, like all pop music the product is crafted in the studion from a mix of different sources. It's not like classical, where sometimes you can put a Decca Tree microphone array up in front of what is effectively a live performance. A certain amount of processing is a perfectly fine part of the mixing and original mastering process of making the mastertape for the stereo release.

Intentional and unintentional distortion

Remastering is where you take that stereo mastertape and do something else to it before putting it on CD. Think about that statement for a while. The stereo mastertape was good enough to make vinyl and cassettes that often sold in big numbers with Olivia, a lot of happy listeners. We now come up with CD that now surpasses most technical parameters of those analogue distribution channels and... all of a sudden you have to do something else to the master? It's not totally illogical, vinyl didn't really like some things like huge amounts of bass and had other foibles people had to allow for. Undoing that is not what remastering has come to mean.

Nowadays remastering involves compressing the bejesus out of the music. While I don't like that2, it is perfectly possible to do it technically without introducing unintentional distortion into the signal. Remastering for loudness is intentional distortion. It is an aesthetic and cultural call. Unfortunately, some remastering engineers ended up introducing extra distortion they didn't anticipate, and the evidence is on some CDs. Remastering catalogue is not a high-budget operation. So maybe sometimes they just remaster, fire and forget.

In the search for max loudness, some remastering pushes the signal beyond the maximum level. What happens then is at best you get clipping, but how that happens is undefined - it is different for different CD players3, and it's never going to sound good. Some fans said that there was audible distortion on the Physical CD of the Japanese box set Not all did. Sure, some fans just won't hear clipping distortion, but I was surprised because while I didn't care for the generally bright and harsh remastering I didn't hear massive peak distortion. I listened to Physical on vinyl in 1981 so I don't have a twenty-something's ears, but I would expect to hear this sort of thing. Turns out that that as the paper in the footnote says, not all CD players respond the same to signals that have gone over the top. Some people will hear more distortion because there is more distortion in the output of their CD players. Mine is perhaps one of the better ones.

IMO, if the process of remastering causes clipping, and particular if it pushes consumers' CD players into different sorts of distortion because they are being pushed outside the design envelope, that's unprofessional. That AES convention was in the year 2000, so the problem has been known about for over 20 years.

waveform of the Japanese Box Set 2010 compared with the Japanese 1984 CD underneath.

Compare the Japanese Box Set 2010 CD (UICY) with the Japanese 1984 CD (CP35). It's really not hard to see that something has gone very seriously wrong here on the later remaster.

If you want to discriminate between less dynamically compromised remasters and more compromised ones, the easiest diagnostic is to compare this parameter. You want it to look more like the left hand side for less compressed.

It's hard to describe what this actually does. MasVis have an understanding the graphs article.

The difference between maximum in the solid line and the dashed line correlates quite well with the amount of crest factor loss that is a result of level maximation of the 2-channel mix.

waveform of the Japanese 1984 CD compared with the Japanese Box Set 2010

Now contrast the Primary Wave remastering of Physical with the Japanese Box Set 2010. Primary Wave have clearly pushed loudness up compared to 1984, but rather less that the 2010 Box Set. This is easy to see visually.

the histogram of levels shows three spurs in both cases. The biggest, at 0, is the quiet runout of about 2s. Nothing controversial there. The two spurs at the extremes of the histogram show a problem, showing the waveform dwells at flat out positive or flat out negative for many samples. This is entirely absent in the 1984 copy of the track, so we can safely say that this shows distortion due to remastering. There's much less of it in the Primary Wave release compared to the box set.

Can you hear the difference?

I don't know what your hearing is like. But I can hear the difference between remastered versions, in an ABX trial, and I don't even care that much for the track Physical. Sorry, Olivia ;) I didn't strain to hear the difference, four minutes, quick-fire 16 trials, first go, the difference isn't that subtle, 100% accuracy. Log file is here and you may verify it here. I don't even care for remastered sound, but there's still a difference between these two to my ears. I listened around the part where it goes I gotta handle you just right - You know what I mean

The second line got a little screechy on the Japanese version. I didn't really like what either did to the drums, compared to the original, but I can let that go. Screechiness on Olivia, no. To my ears the Japanese remaster sounds brighter4 than the Primary Wave one, but not really in a good way.

I repeated the ABX exercise on the Australian 1993 version of If Not For You against the Primary Wave INFY5. I still prefer the older version (file A). To me there's a little bit less top and air around Olivia's voice on the PW relase. In fairness, this analogue mastertape (or possibly a copy, since INFY was recorded in England) was 22 years old when Festival made the CD in 1993. The original tape was 51 years old by the time Primary Wave got to transfer it, perhaps nearly three decades take a toll.

page top

  1. Let's face it, Festival have been known to miss tracks off a CD and Primary Wave have been known to get the bonus CD listing not quite in the same order as the tracks on the CD. Those guys get paid to get things right, whereas I am doing this in my free time, so I may screw up too. ↩
  2. I have listed my own view on remastering here. You may find it opinionated, obnoxious and at variance with your experience. That's why it's only linked in a footnote. ↩
  3. A more technical reference is 0dBFS+ Levels in Digital Mastering, Nielsen and Lund, TC Electronic A/S, Risskov, Denmark presented at AES 109. A copy can be found here as of Nov 2022. The authors' observation All of the domestic CD players investigated have shown difficulty dealing with 0dBFS+ levels that can easily occur on modern CDs. New models are actually worse than older types relying less on oversampling and more on analog filters.
    We have not investigated how seriously audio quality is subjectively affected, nor have we made any listening fatigue tests concerning 0dBFS+ levels. However, modern CDs contain these kind of signals and modern CD players are not designed to reproduce them without distortion.     may explain the anecdotal observation that some Olivia fans complain of distortion on the Japanese boxset Physical CD where others don't. ↩
  4. You see a hint of that brightness in the comparison in the relative heft between 1kHz and 4kHz on the box set version. ↩
  5. Log file is here. File A was the Australian 1993 version of Love Song in that case.↩