As I mentioned on the thread about the recent box set reissue of APP's Pyramid album, I was so annoyed by how the 5.1 mix of The Turn of a Friendly Card sounded, that I took it upon myself to try and de-turd-ify (highly technical industry jargon) the 5.1 mix because it's by far my favourite APP album, and it seemed like underneath the piercing treble, weirdly scooped-out midrange and lack of low-end impact that there was actually a pretty good mix.
I also felt like (as mentioned in the post linked above) that most of this damage (for lack of a better word) was done during the mastering phase - or at least as a global EQ applied to the entire mix during export - because it wasn't like only specific elements within the mix like vocals or hi-hats were affected by the scooped mids and boosted top end, it was everything. To me this suggested that the mix could probably be salvaged with judicious (extreme) use of EQ, based on the premise that if the 5.1 mix had a similar instrument and vocal balances to the stereo mix with the only difference being that they were spread between five channels instead of two, that if I make the tonality of the 5.1 mix match the stereo mix then the 5.1 mix should sound good.
Just as a sidenote, and you can skip this paragraph if you don't want to hear me talk about myself and my feelings: I find myself having to do this kind of thing, remaster, adjust, or somehow "fix" surround mixes released by the major labels (or licensees thereof) and I've grown to become really resentful of it. It's incredibly frustrating and disheartening to feel like these companies - three of which own what feels like 90% of all the popular recorded music of the last century - who should be doing their utmost to present the part of our shared cultural heritage that they're lucky enough to be stewards of in the best possible light, but instead choose to use either mastering engineers (and other people involved with these projects) who either don't understand or don't care about the preferences of their target audience, or even indeed what objectively "good" sounding audio is, or in many cases no mastering engineer at all. Granted I'm not party to the financial, corporate/managerial or artistic pressures on any of these projects, but to me it seems insane to push stuff like this out the door without having a recognized, qualified expert run the auditory ruler over it, especially when you're marketing these products as an upgrade on the versions that have come before. Every LP cut back in the day was done by a craftsman whose life's work was cutting LPs, and during the golden age of CD reissues the same handful of guys who had 15 or 20 years of LP mastering were responsible for mastering thousands of CDs and as listeners we all benefited from that accrued experience. It feels like we have far too many jacks-of-all-trades in the production chain these days: tape transfer specialists who also do mixing; mixing engineers who master their own work; Blu-Ray authors with a sideline in mastering, and guys who do even more of those things all at the same time, and the result of having all these "lowest quote wins the contract" people is that we have no masters of any of the individual disciplines unless you're Pink Floyd or something. I appreciate that the industry has contracted a great deal in the last couple of decades and that the budgets for projects like these have gone down accordingly, but there has to be a better way to do things, because I refuse to believe that "economical" and "good" are entirely mutually exclusive, when a no-talent like me can detect and remedy what seem like obvious issues like channel assignment errors, out-of-phase channels and out of sync LFE tracks, for example.
But anyway, I digress - the point being that (outside of Mike's D-V masterings) I find myself regularly having to "fix" new surround releases in some way or another to make them personally palatable, and while I've posted my fair share of them here on QQ, there are so many more that I've kept schtum about, partly because I feel like some of these releases need to fail in order for some of these companies to change their ways (and that they won't if there's a 'Catcher in the Rye' like me saving them before they go off the cliff) and also because I'd like to see them paying me (or someone like me) to act as a mastering/quality control consultant. But, I'm breaking my self-imposed rule with this one because I really like this mix and I'm hoping that a few more people the the capability to replicate my work will get to enjoy (rather than endure) it.
To do this work, I used a VST plugin called CurveEQ by Voxengo. I'll spare the tutorial because you don't know how to use it in order to take advantage of my EQ curves, but in a nutshell it's a 30-point graphic EQ where you can play it two different recordings (in this case the stereo mix and the 5.1 mix) and then it will generate an EQ curve that "describes" the difference between the two. So I can play recordings of each track (stereo and 5.1) and then end up with an EQ curve to make the 5.1 version sound (tonality-wise) like the stereo mix. The one caveat is that CurveEQ isn't free - it's about $80 USD, but Voxengo are currently running a promotion (15% off + 8% "flash discount") that knocks it down to $62, at least in my web browser. Having used it extensively over the last couple of years, I can say for me anyway that it's a very fair value at that price.
On a related note, for the purposes of comparison I used the 2004 Classic Records HDAD, which is a double-sided DVD-V with 192/24 stereo on one side, and 96/24 stereo on the other. Supposedly the master used for this release came from Parsons' own collection, and for me it's easily the best-sounding mastering of this album. Like a lot of albums, the 80's CD versions of ToaFC are bright and brittle sounding, and the modern remasters are compressed, but this one has just about perfect tonality. Really nice midrange and great deep bass extension that no other version has, and it's 192/24 to top it off. Having said all that, however, there is one problem with this remaster - the left-right balance skewed slightly to one side (sorry, can't remember which way now, but it's obvious in the waveforms) for some reason, by somewhere between 1.5 and 2dB. Not enough to be catastrophic, but it pushes the phantom center (lead vocals in particular) slightly to one side of 12 o'clock. So if you have this, it's worth fixing as it really improves the phantom center.
If you don't have (or don't want to pay for) CurveEQ, applying similar looking curves should yield a positive (though not as precise) result, and for those with CurveEQ, I've attached the .CSV (save) files for the 10 tracks. Seeing as .CSV (comma-separated value) are just text files, if you don't have CurveEQ, you can open these in your spreadsheet software (like Microsoft Excel) and it will display them in two columns, one column for the frequency band in Hz, and another column for the boost (or cut) in dB to that frequency band.
(So this is the lower frequency range of track 1, +6.64dB at 20Hz, +6.32dB at 25.38Hz, and so on...)
Anyway, without further ado, here are the screenshots of the EQ curves for all 10 tracks - the green (front) waveform is the 5.1 mix, the pink (rear) waveform is the stereo, and the orange line running through the middle is the actual EQ curve applied.
01-May Be a Price to Pay
02-Games People Play
03-Time
04-I Don’t Wanna Go Home
05-The Gold Bug
06-The Turn of a Friendly Card, Part One
07-Snake Eyes
08-The Ace of Swords
09-Nothing Left to Lose
10-The Turn of a Friendly Card, Part Two
Obviously the curves vary somewhat from track to track, but you can see that generally they all need way more bass, an injection of midrange somewhere around 5kHz (to give the vocals and instruments body and presence) and a big treble roll-off starting around 10kHz to tame the sibilant hi-hats and piercing nature of other instruments like horns and cymbals.
It's also worth noting that if you are using CurveEQ, because it boosts as well as cuts in generating curves it sometimes results in the output track having louder peaks than the original - in the case of this album everything is normalized to 0dB so to avoid clipping you need to reduce your master volume to avoid clipping. These are the reductions I jotted down when I was doing mine, though your results may vary:
Track 1 - Master bus: -1.6dB
Track 2 - Master bus: -2.7dB
Track 3 - Master bus: -2.0dB
Track 4 - Master bus: -2.8dB
Track 5 - Master bus: -3.9dB
Track 6 - Master bus: -4.1dB
Track 7 - Master bus: -4.1dB
Track 8 - Master bus: -3.9dB
Track 9 - Master bus: -3.9dB
Track 10- Master bus: -2.3dB
Track 6 segues into Track 7, and Track 8 segues into track 9, so whatever you choose to do, the reductions for these two pairs of tracks should be the same (6 same as 7, 8 same as 9) otherwise you'll get a "volume hiccup" as one track switches to the other.
When I finished exporting my new "remastered" 5.1 mix and listened to it back, I still wasn't entirely happy - aside from the EQ problems, this mix was also hit with a really unnecessary brick-wall limiter, and maybe other compression, I don't know, but the clipped peaks are visually obvious when you look at the waveforms. Normally I'm not one to bother with de-clipping audio as I think it has minimal impact (and I don't believe it actually does anything for dynamic range) but this album is so immaculately recorded I felt like I could hear a kind of "halo of distortion" around particularly loud transients, notably snare drum hits and cymbal crashes.
Because this 5.1 mix was normalized with everything peaking at 0dB, if I did want to reconstruct the clipped peaks I knew I needed to reduce the overall volume of the whole album so there was headroom for them to expand into, so I had to come up with some kind of way of deciding how much to turn it down by. It was sort of a three bears situation - if I didn't turn it down enough, it wouldn't be possible to fully reconstruct the peaks at maximum amplitude, and if I reduced it by too much I was basically losing fidelity by decreasing the signal-to-noise ratio unnecessarily. I don't know if it's flawed logic, but the "just right" methodology I went with was that I looked at the DR of the original stereo mix (DR14) and then looked at the DR readings of individual channels in the 5.1 mix (FL and FR were about DR9, and then the C, Ls and Rs were at or above DR14) and came to the conclusion that the difference between the clipped FL & FR channels (DR9) and the other channels (DR14) which was 5dB seemed like a good number.
Here's a sample of my work on Track 2 ('Games People Play'):
(Original track with peaks at 0dB)
(Track volume reduced by 5dB - I think you can really see how much brick-wall limiting there is in this view)
(With the clipped peaks restored, and probably a lot like the track looked before it was originally mastered, with the louder parts allowed to 'breathe' properly.)
Now I'm fully willing to admit that it may be placebo effect to some degree, but to my ears, de-clipping the album seemed to help with the previously-described "halo of distortion" issue. It's not 100% gone (you can't unburn the toast, but you can put butter on it) but it feels like this sonic therapy has taken the edge off it and pushed it into the background.
Now (in closing, finally!) I'm not saying this mix is 100% perfect - it feels a bit sterile occasionally, probably as the result of being done in a DAW in 2023 and not in a multi-million dollar recording studio, and there are times where Parsons could've been more Dark-Side-of-the-Moon-quad aggressive, but overall it's a very satisfying listen once put the EQ and compression issues into a headlock and punch them into submission. If anyone else cares to try and replicate my work (and to the handful of compadres I've shared samples of my work on this album with) I'd be interested to hear what you think.
I also felt like (as mentioned in the post linked above) that most of this damage (for lack of a better word) was done during the mastering phase - or at least as a global EQ applied to the entire mix during export - because it wasn't like only specific elements within the mix like vocals or hi-hats were affected by the scooped mids and boosted top end, it was everything. To me this suggested that the mix could probably be salvaged with judicious (extreme) use of EQ, based on the premise that if the 5.1 mix had a similar instrument and vocal balances to the stereo mix with the only difference being that they were spread between five channels instead of two, that if I make the tonality of the 5.1 mix match the stereo mix then the 5.1 mix should sound good.
Just as a sidenote, and you can skip this paragraph if you don't want to hear me talk about myself and my feelings: I find myself having to do this kind of thing, remaster, adjust, or somehow "fix" surround mixes released by the major labels (or licensees thereof) and I've grown to become really resentful of it. It's incredibly frustrating and disheartening to feel like these companies - three of which own what feels like 90% of all the popular recorded music of the last century - who should be doing their utmost to present the part of our shared cultural heritage that they're lucky enough to be stewards of in the best possible light, but instead choose to use either mastering engineers (and other people involved with these projects) who either don't understand or don't care about the preferences of their target audience, or even indeed what objectively "good" sounding audio is, or in many cases no mastering engineer at all. Granted I'm not party to the financial, corporate/managerial or artistic pressures on any of these projects, but to me it seems insane to push stuff like this out the door without having a recognized, qualified expert run the auditory ruler over it, especially when you're marketing these products as an upgrade on the versions that have come before. Every LP cut back in the day was done by a craftsman whose life's work was cutting LPs, and during the golden age of CD reissues the same handful of guys who had 15 or 20 years of LP mastering were responsible for mastering thousands of CDs and as listeners we all benefited from that accrued experience. It feels like we have far too many jacks-of-all-trades in the production chain these days: tape transfer specialists who also do mixing; mixing engineers who master their own work; Blu-Ray authors with a sideline in mastering, and guys who do even more of those things all at the same time, and the result of having all these "lowest quote wins the contract" people is that we have no masters of any of the individual disciplines unless you're Pink Floyd or something. I appreciate that the industry has contracted a great deal in the last couple of decades and that the budgets for projects like these have gone down accordingly, but there has to be a better way to do things, because I refuse to believe that "economical" and "good" are entirely mutually exclusive, when a no-talent like me can detect and remedy what seem like obvious issues like channel assignment errors, out-of-phase channels and out of sync LFE tracks, for example.
But anyway, I digress - the point being that (outside of Mike's D-V masterings) I find myself regularly having to "fix" new surround releases in some way or another to make them personally palatable, and while I've posted my fair share of them here on QQ, there are so many more that I've kept schtum about, partly because I feel like some of these releases need to fail in order for some of these companies to change their ways (and that they won't if there's a 'Catcher in the Rye' like me saving them before they go off the cliff) and also because I'd like to see them paying me (or someone like me) to act as a mastering/quality control consultant. But, I'm breaking my self-imposed rule with this one because I really like this mix and I'm hoping that a few more people the the capability to replicate my work will get to enjoy (rather than endure) it.
To do this work, I used a VST plugin called CurveEQ by Voxengo. I'll spare the tutorial because you don't know how to use it in order to take advantage of my EQ curves, but in a nutshell it's a 30-point graphic EQ where you can play it two different recordings (in this case the stereo mix and the 5.1 mix) and then it will generate an EQ curve that "describes" the difference between the two. So I can play recordings of each track (stereo and 5.1) and then end up with an EQ curve to make the 5.1 version sound (tonality-wise) like the stereo mix. The one caveat is that CurveEQ isn't free - it's about $80 USD, but Voxengo are currently running a promotion (15% off + 8% "flash discount") that knocks it down to $62, at least in my web browser. Having used it extensively over the last couple of years, I can say for me anyway that it's a very fair value at that price.
On a related note, for the purposes of comparison I used the 2004 Classic Records HDAD, which is a double-sided DVD-V with 192/24 stereo on one side, and 96/24 stereo on the other. Supposedly the master used for this release came from Parsons' own collection, and for me it's easily the best-sounding mastering of this album. Like a lot of albums, the 80's CD versions of ToaFC are bright and brittle sounding, and the modern remasters are compressed, but this one has just about perfect tonality. Really nice midrange and great deep bass extension that no other version has, and it's 192/24 to top it off. Having said all that, however, there is one problem with this remaster - the left-right balance skewed slightly to one side (sorry, can't remember which way now, but it's obvious in the waveforms) for some reason, by somewhere between 1.5 and 2dB. Not enough to be catastrophic, but it pushes the phantom center (lead vocals in particular) slightly to one side of 12 o'clock. So if you have this, it's worth fixing as it really improves the phantom center.
If you don't have (or don't want to pay for) CurveEQ, applying similar looking curves should yield a positive (though not as precise) result, and for those with CurveEQ, I've attached the .CSV (save) files for the 10 tracks. Seeing as .CSV (comma-separated value) are just text files, if you don't have CurveEQ, you can open these in your spreadsheet software (like Microsoft Excel) and it will display them in two columns, one column for the frequency band in Hz, and another column for the boost (or cut) in dB to that frequency band.
(So this is the lower frequency range of track 1, +6.64dB at 20Hz, +6.32dB at 25.38Hz, and so on...)
Anyway, without further ado, here are the screenshots of the EQ curves for all 10 tracks - the green (front) waveform is the 5.1 mix, the pink (rear) waveform is the stereo, and the orange line running through the middle is the actual EQ curve applied.
01-May Be a Price to Pay
02-Games People Play
03-Time
04-I Don’t Wanna Go Home
05-The Gold Bug
06-The Turn of a Friendly Card, Part One
07-Snake Eyes
08-The Ace of Swords
09-Nothing Left to Lose
10-The Turn of a Friendly Card, Part Two
Obviously the curves vary somewhat from track to track, but you can see that generally they all need way more bass, an injection of midrange somewhere around 5kHz (to give the vocals and instruments body and presence) and a big treble roll-off starting around 10kHz to tame the sibilant hi-hats and piercing nature of other instruments like horns and cymbals.
It's also worth noting that if you are using CurveEQ, because it boosts as well as cuts in generating curves it sometimes results in the output track having louder peaks than the original - in the case of this album everything is normalized to 0dB so to avoid clipping you need to reduce your master volume to avoid clipping. These are the reductions I jotted down when I was doing mine, though your results may vary:
Track 1 - Master bus: -1.6dB
Track 2 - Master bus: -2.7dB
Track 3 - Master bus: -2.0dB
Track 4 - Master bus: -2.8dB
Track 5 - Master bus: -3.9dB
Track 6 - Master bus: -4.1dB
Track 7 - Master bus: -4.1dB
Track 8 - Master bus: -3.9dB
Track 9 - Master bus: -3.9dB
Track 10- Master bus: -2.3dB
Track 6 segues into Track 7, and Track 8 segues into track 9, so whatever you choose to do, the reductions for these two pairs of tracks should be the same (6 same as 7, 8 same as 9) otherwise you'll get a "volume hiccup" as one track switches to the other.
When I finished exporting my new "remastered" 5.1 mix and listened to it back, I still wasn't entirely happy - aside from the EQ problems, this mix was also hit with a really unnecessary brick-wall limiter, and maybe other compression, I don't know, but the clipped peaks are visually obvious when you look at the waveforms. Normally I'm not one to bother with de-clipping audio as I think it has minimal impact (and I don't believe it actually does anything for dynamic range) but this album is so immaculately recorded I felt like I could hear a kind of "halo of distortion" around particularly loud transients, notably snare drum hits and cymbal crashes.
Because this 5.1 mix was normalized with everything peaking at 0dB, if I did want to reconstruct the clipped peaks I knew I needed to reduce the overall volume of the whole album so there was headroom for them to expand into, so I had to come up with some kind of way of deciding how much to turn it down by. It was sort of a three bears situation - if I didn't turn it down enough, it wouldn't be possible to fully reconstruct the peaks at maximum amplitude, and if I reduced it by too much I was basically losing fidelity by decreasing the signal-to-noise ratio unnecessarily. I don't know if it's flawed logic, but the "just right" methodology I went with was that I looked at the DR of the original stereo mix (DR14) and then looked at the DR readings of individual channels in the 5.1 mix (FL and FR were about DR9, and then the C, Ls and Rs were at or above DR14) and came to the conclusion that the difference between the clipped FL & FR channels (DR9) and the other channels (DR14) which was 5dB seemed like a good number.
Here's a sample of my work on Track 2 ('Games People Play'):
(Original track with peaks at 0dB)
(Track volume reduced by 5dB - I think you can really see how much brick-wall limiting there is in this view)
(With the clipped peaks restored, and probably a lot like the track looked before it was originally mastered, with the louder parts allowed to 'breathe' properly.)
Now I'm fully willing to admit that it may be placebo effect to some degree, but to my ears, de-clipping the album seemed to help with the previously-described "halo of distortion" issue. It's not 100% gone (you can't unburn the toast, but you can put butter on it) but it feels like this sonic therapy has taken the edge off it and pushed it into the background.
Now (in closing, finally!) I'm not saying this mix is 100% perfect - it feels a bit sterile occasionally, probably as the result of being done in a DAW in 2023 and not in a multi-million dollar recording studio, and there are times where Parsons could've been more Dark-Side-of-the-Moon-quad aggressive, but overall it's a very satisfying listen once put the EQ and compression issues into a headlock and punch them into submission. If anyone else cares to try and replicate my work (and to the handful of compadres I've shared samples of my work on this album with) I'd be interested to hear what you think.
