I recently listened to Beck's Sea Change, and it reminded me of how much bottom end a CD can actually incorporate. Some early transferred CDs — or digital recordings, for that matter — are missing low end and front-to-back depth.
There are a lot of reasons for this. I started working with digital in 1978, when I mastered and cut a recording for Telarc, for vinyl, on the Soundstream digital machine. Until the invention of the Neve DTC-1 digital domain mastering console in 1987, and the Daniel Weiss BW-102, there was no way to master and stay in the digital domain. One always had to play back the digital master through a not-so-great Sony PCM-1610 digital-to-analog converter [DAC], do all the mastering in the analog world, and re-record it back into digital through an even less-good analog-to-digital [ADC] Sony converter for CD — first the Sony PCM-1600, then 1610 and finally the 1630. For a while, as everything was 16-bit — even for post-production and mastering — simple level changes sounded dicey in the digital domain. Digital equalization was initially so horrible and brittle; no one would use it. The PCM-1600 converters were a "ripped from the textbook" industrial design converter that had all those sound qualities you describe. Some great digital recordings, like the Rush's Moving Pictures CD, were recorded with it; but the artists and producers mixed while listening through its output, so they adjusted EQs and levels to accommodate the sound of the converter. The invention of the CD meant that there were now tens of thousands, and then millions, of DACs being sold and all R&D went into developing better DACs. I remember when there were no more than a handful of ADCs in all of New York City! The analog to digital converter took a while for developers to really pay attention to it. In my opinion, it wasn't until the middle 1990's that some really first-rate converters were invented. The first Sony digital editors, which were necessary for creating the CD masters, did not have dither. Even their manual suggested passing through the -60 dB area of the fader as quickly as possible to avoid digital distortion when one had to do a digital fade! The widespread use of dither — which makes low-level digital go from horribly distorted into extremely low distortion, and gives the ability to hear sounds below the least significant bit, was a major leap in the improvement of CD sound.
I read that when transferring some '60s and '70s records to digital for the first time, some engineers kept EQ settings originally meant to accommodate vinyl. Do you think that's part of the problem?
Well, what happened was merely a matter of economics. Mastering engineers had always mastered with the vinyl disk in mind. When they made mastered, equalized tape copies for cassette duplication, it contained all the engineering optimized for vinyl; which means making sure the inner bands had adequate top end to counter the diameter equalization losses inherent in the physics of vinyl disks. Sometimes the very lowest bass frequencies were somewhat "mono-ed" to make the groove easier to cut, and to prevent skipping from the groove thinning out. CDs only sold 800,000 copies total for the USA in 1983, jumping to 5.8 million in 1984, to over 100 million by 1987. CD sales didn't exceed vinyl until 1992. The Dire Straits Brothers in Arms CD I mastered in 1985, one of the very first albums recorded on the Sony 24-track digital machine, was the first CD I mastered that was totally mastered for the CD medium. It was also longer than the vinyl version. That original version had to be mastered in the analog domain, in spite of how great Neil Dorfsman's mixes were. Please re-buy my latest all-digital remastering of it; it's how I always wanted it to be.
Some remasters of old albums shoot for a contemporary "bright and loud" sound. How would a flat transfer benefit today — due to modern converters, playback electronics, etc. — in comparison to the first CD version some 30 years ago?
In the past 30 years, every aspect of the recording chain has dramatically improved. I have five different analog playback electronics from which to choose for the playback of the masters, including a Studer A820 and a modified Ampex ATR-102, with both balanced and un-balanced outputs, as well as the discrete Class A Aria solid state electronics and the Esoteric Audio Research tube electronics. I recently mastered Elton John's 40th Anniversary Edition of Goodbye Yellow Brick Road; that involved choosing the correct 1/4-inch tape playback head. There are two standards for 1/4-inch tape: the USA head, which has a 2 mm guard band between the left and right heads, versus the European 0.75 mm head — which was used when they recorded Goodbye Yellow Brick Road — so reproducing the tape meant that there wasn't the extra bass bump that would have happened with a mismatched playback head. The new Transparent Audio cables I use throughout the studio simply pass more musical signals than even their original cables did when I first outfitted Gateway Mastering with them in 1992. The Apogee Symphony, Pacific Microsonics, dCS, and Horus converters I use are truly superior to what we had 30 years ago. My [SPL] 8-channel analog console is a state-of-the-art design, with 124-volt DC rails — something unheard of 30 years ago, and still almost never pushed to that extreme in most audiophile designs. It never distorts. The [Merging Technologies] Pyramix workstation is also state-of-the-art. In addition, our studio is properly grounded; we run the studio off huge batteries and create our own 60 Hz AC power.
Let's talk about the starting point for a remastering job. I think the term "master tape" often gets confused with the actual mastered tape. When remastering, do you create a new mastering instead of the old one, or do you work on top of it?
A "master tape" is the original mix approved for the production of the album, ready to be mastered. A "master take" would be the choice of the sometimes hundreds of mixes that were created and used for mastering the final album. An "EQ master" or "production master" would be the approved mastered files [or tapes] that were used for production. I just did a big remastering project, and the record company lost the master tapes. We had to use an analog copy that the artist had for his personal files. Now that safety copy would now be re-labeled as a "master." When I do a remastering job for a client, the ones with a budget and time try to find me all the original master tapes, or digital files, for me to do my best, like the recent Elton John Goodbye Yellow Brick Road or all the Queen remasters from a few years ago. This is the best situation. Sometimes the record label does not want to ship the original master tapes to me from their library and insist on sending me a flat digital transfer for me to master from. This removes my ability to choose which different tape electronics and head combinations sound best to my ears, which is a big problem for me. In addition, sometimes not every cut from an album is in proper azimuth, and the label often does not correct for this when making the transfer. An even worse situation is when the label loses the master tapes and only has the commercial CD for me to work from, in which case I have to master on top of what is there. If they over-compressed it originally, there is no way to undo that.
How do you approach the remastering of a record sonically, especially those that you had mastered originally?
When I'm asked to remaster a project I originally did, I try to have a discussion with the artist about what they wish to accomplish with the remastering. Some artists, like John Mellencamp, wished that the remastering could update the sound and make the old recordings sound like they were new. Others, like Bruce Springsteen, want to pretty much keep what he originally approved. Goodbye Yellow Brick Road had been re-mastered quite a few times in its long history. A friend of mine, who knew Elton John's producer Gus Dudgeon, told me which one of the remasterings had been personally supervised by Gus. When I bought that version, I found that it sounded more like the flat master than all the other remasterings I'd checked out, so I used Gus's work as a guide for my approach. Honestly, it was right in line with my original thoughts. Sometimes I wish I had been able to remaster recordings I originally did, like AC/DC's Back in Black, Led Zeppelin II and Houses of the Holy, or all The Band records I originally mastered. I'd love for people to hear the recently re-mastered early Bruce Springsteen titles that are Mastered For iTunes [MFiT] from the 24-bit sources, especially Darkness on the Edge of Town, which is one of my favorite albums by Springsteen. Those analog MFiT titles were transferred using the Plangent Processes [Tape Op #94] to lower the FM distortion, as well as correct wow and flutter on those early tape machines. I think the difference is quite amazing. [see page 80]
Where do you draw the line between noise removal and "sound reduction"?
For me, even the best de-hissers can cause artifacts that have to be traded off with the amount of de-hissing. On sparse intros or final chords I might throw in a few dBs of hiss reduction, if it gives me more than I lose. Even when mastering a new recording, there are often small ticks and mouth noises that are annoying. By selecting only the tick, which can be a fraction of a millisecond long, and remove it with intelligent de-clicking software, it is 100 percent impossible to hear an artifact. I use programs like Sonic Solutions NoNOISE, Cedar, Waves, or iZotope, depending on the problem. I recently removed the bleed of a click track coming through the artists' headphones during the ring out of a guitar note using iZotope RX 3, but I'm sure the Cedar Retouch or Algorithmix reNOVAtor could have done it as well. I would never try to draw a click away [on the waveform]. I know it works sometimes, but often the waveforms are so complicated that it would be impossible to draw out the click without further eliminating the good part of the music. Good restoration software looks at the problem area of the tick, does an FFT [Fast Fourier Transform] evaluation of the music before and after the tick, and creates an interpolation of what would have probably been there. If one owns several pieces of restoration software, usually at least one of them will solve the problem.
How would you rate the relationship between sound quality and getting the song across to the listener?
This is a really complex question! Suffice it to say that Marshall McLuhan's "the medium is the message" applies here. There are so many approaches for how one can present one's new artistic vision to the public. Some of the songs on the new Beck album, Morning Phase, are purposely grungy; he wanted "a little grittiness" in the album. We also had clean, non-distorted mixes. On most songs he chose the clean ones, but some are gritty on purpose. That was his artistic choice as to how he wanted his music to be heard. Super-dynamic, normal dynamics, or super-compressed — those are all decisions the artist needs to make, and hopefully I can assist them with their vision.
On the other hand, there's music that works especially within the boundaries of limited sound quality. Bruce Springsteen's Nebraska comes to mind, which is haunting, despite its "lo-fi" 4-track cassette sound.
With Nebraska, Bruce and his producers remixed the 4-track Portastudio cassette through a Neve console and an SSL console, but Bruce liked the cassette. I got a Nakamichi Dragon, a state-of-the-art cassette machine. I corrected the azimuth and speed of the tape, but Bruce liked it left alone — running quite slowly when played on a proper cassette deck. A few times during the years, when "Atlantic City" became part of a compilation, I would ask Bruce if he wanted the speed corrected to the same 440 [Hz] pitch he used when he sang it, and the answer is always, "No." It is what it is, and it is a work of high art. Just recently I did a project with Jack White and Neil Young, called A Letter Home. Jack White refurbished a 1947 Voice-o-Graph vinyl recording booth, used to make one-off disks to keep, or mail to someone, to hear on their phonograph. It's almost impossible to keep on-speed, and, by design, the recording sounds not only like a 78-rpm disk, but a well-worn and ravaged one at that! Jim Farber at the New York Daily News brilliantly said it was "like some audio equivalent of a message in a bottle, long ago tossed into the sea." The record is sometimes hard to listen to, with lots of ticks, pops, and other noises, but it is 100 percent vibe and it works beautifully. A project totally worthy of the extreme effort they went through to create it.
