Q&A with Barry Diament of Soundkeeper Recordings
Some of you already know about Barry Diament from his CD mastering days at Atlantic Records where he was the mastering engineer for too many classic CD releases to do justice to with a short list (here's one longer list and I'll cherry pick Led Zeppelin II, Physical Graffiti, and the majority of the Bob Marley and The Wailers catalog to give you a peak at the tip of the iceberg). And some of you may know Barry from his label Soundkeeper Recordings or his audio recording, production, consulting, mastering, and editing services offered through Barry Diament Audio.
You may also recall our Soundkeeper Recordings Format Comparisons post that talks about their free downloads of the same recordings in various levels of resolution (16/44, 24/96, and 24/192) so you can listen for yourself. And others still may have seen Barry's often informative comments on various forums.
Barry Diament was kind enough to agree to this Q&A and as you'll see, his answers are thorough and illuminating.
You have a very interesting resume that includes being one of the first CD Mastering Engineers. Can you talk a bit about your history as a recording engineer and bring us up to date with your Soundkeeper Recordings label?
First, I want to say Thank You for your interest, Michael. I’m honored to be on AudioStream.
I started recording when I was eleven, playing with my brother’s Concord reel-to-reel deck. While recording the weekly jams I had with a few friends, I found that I could play drums and then, using the sound-with-sound feature, add a guitar part later. I’d found “overdubbing” without knowing what it was.
During my college years, I was reading all the audio magazines I could get my hands on, from Audio to Stereo Review to High Fidelity, later finding some of the British magazines, such as Hi-Fi News in a local magazine store. Then came the discovery of Stereophile and The Absolute Sound, both of which (in addition to Bert Whyte’s columns in Audio) became a sort of audio “school”, in addition to the other reading and experimenting I was doing.
"While the studio experience was wonderful, I noticed early on that what I heard in the control room did not sound like what I heard out in the room with the musicians."
After college, I got my first studio job as an assistant engineer, setting up microphones and operating the tape machines during recording sessions. Once I was promoted to senior (then chief) engineer, I was doing recording, overdubbing and mixing and later learned to cut vinyl in the in-house mastering room. While the studio experience was wonderful, I noticed early on that what I heard in the control room did not sound like what I heard out in the room with the musicians. So began my fascination with monitoring and the realization that this was, in my view, the single most important factor in the studio. After all, if one could not hear what they were doing, nothing else really mattered.
Then I started to ask what I call “The Questions”, things I was never taught and which, to my knowledge, are not taught in the real “audio engineering schools” that have arisen in the intervening years. Questions like “Why this microphone?” and “Why place it here?”
When I heard an editor’s position opened up at Atlantic Records, I jumped at the chance and luckily for me, landed the job. My specific task was to make long songs shorter (to create “single” versions of album length songs, radio stations preferring to keep songs not much longer than three minutes) and to make short songs longer (to create the “dance” versions that were popular in the clubs).
"In January of 1983, Atlantic purchased the gear to create CD masters, built a mastering room and I was made the “CD mastering department”."
In the mid-late ‘70s, I’d heard talk of a new future format for recorded music called the Compact Disc. I remember a cardboard mock-up someone had given me, of a rainbow colored disc only about 5 inches in diameter, in what seemed like a miniature LP jacket. By 1982, Atlantic’s sister company, Warner Brothers, was already creating CD masters for the Warner family of labels (Warner, Atlantic, Elektra, etc.). In January of 1983, Atlantic purchased the gear to create CD masters, built a mastering room and I was made the “CD mastering department”. To my knowledge, at the time, only Sony in Japan, Polygram in Germany and Warner in Los Angeles and perhaps one or two other facilities had CD mastering engineers.
I remember hearing my first CDs then and thinking that the removal of hiss, crackles and wow and flutter (speed irregularities) was a good thing but the sound of the instruments themselves was not the great improvement it was promoted as being. In fact, I found my vinyl records to much more closely represent what I was used to hearing in the studio.
Over the years, some significant improvements in CD sound quality followed, from better players, more successfully able to reproduce the sound of the digital CD master to better CD mastering gear. When Atlantic installed Apogee’s retrofit filters in the Sony mastering gear, there was a nice step up in sonics but I still felt there was a long way to go.
In order to do the best job I could, I set about making some improvements in the mastering room, starting with the monitoring. I had the studio replace the little box speakers in the room with a pair of Dahlquist DQ-10s, later followed by DQ-20s. I treated the reflection points on the walls with absorbent material and brought my own cables to work every day in order to bypass the console, switching, patch bay and other standard accoutrements of a typical mastering room. Typically, I would wire directly from the output of the Studer reel-to-reel deck to the input of the Sony 1630 A-D converters. If I felt some EQ was necessary, I’d add only the EQ, with nothing else in the signal path.
"By the mid-1990’s I realized that many clients were starting to evaluate my work using the level meters instead of the loudspeakers."
In 1987, I left Atlantic to form BDA and worked primarily as an independent CD mastering engineer. I worked out of a few different studios, choosing them based on my assessment of the monitoring accuracy. By the mid-1990’s I realized that many clients were starting to evaluate my work using the level meters instead of the loudspeakers. At this point, I had to stop and ask myself just what I sought to accomplish as an audio engineer.
Now, I enjoy loudness when it is appropriate but in my experience, if you want to shake the walls with AC/DC (or with Mahler), the best way to achieve this is with the playback volume control. Any other way, such as arbitrarily increasing the level on the recording itself involves a host of sonic trade-offs. First among them, is the sense of Life that comes from musical dynamics. Since my goals as an engineer are sourced in my love of music, I didn’t want to participate in the ongoing Loudness Wars. All the truly great sounding records and CDs in my collection had much lower average levels than what the majors were releasing. I wanted to preserve all the musical Life in every source I mastered too and never used compression myself. While some say it increases “punch”, the sonic evidence tells a quite different story. Besides, how does one increase punch by reducing dynamics, where the punch “lives”? So, I started accepting only those jobs where the client’s prime interest was the musical presentation and the preservation of musical dynamics.
I also started to think more about the Questions I mentioned earlier. I asked myself if I ever wanted to listen to a great piano player with my head under the lid of a grand piano and my ears just inches above the hammers. Or if I wanted to listen to a ripping guitar solo with my ear up against the grill cloth of a Marshall (amplifier) stack. Or a great horn player with my ear in the bell of the horn. How about a great vocalist with my ear an inch from their lips? This in fact, is where the mics in a typical studio recording “listen” from. I realized that those recordings I found truly amazing sonically used considerably fewer microphones and none were ever placed as close to what they were capturing as is done in a typical studio recording.
I came to realize that 90-95% of a recording’s ultimate sound quality has already been determined by the time the signals are leaving the microphones. This was the genesis of Soundkeeper Recordings. I knew then that my greatest interest in audio and recording is in making records that give the listener the feeling they are in the presence of the performance, in the room where the performance took place.
To be clear, there is a very large library of recorded music that I love, which was made using the typical studio techniques involving multiple, closely placed microphones. Over the years, the recording art has evolved to the point where musical magic can be created which bears little relationship to the original performance. Wonderful and magical as many of these recordings are, they were made to sound like "records"; my goal is to explore the idea of records that sound like performances. Two different approaches, each with its own rewards.
Soundkeeper Recordings are made with all the musicians playing live, in real time. In order to keep the signals as coherent as possible, there is only one microphone per playback channel. The mics are arranged in a stereo array and are separated by a baffle of my own design. Musical balances are achieved by moving players and instruments physically, rather than moving faders on a mixing console. The recordings are captured in stereo with no overdubs or further processing. What leaves the microphones is essentially the finished recording. (Because I record with a lot of headroom, final levels are adjusted in the mastering room.) The results sound very much like what I hear when standing at the position of the microphone array at the recording sessions.
Provenance is a main concern for many HD download customers. Some people would like this provenance information to include things like what generation master tape was used for a given reissue. This is certainly a new requirement as I'm not aware of many LPs or CDs that provide this information. In your experience, how often is it the case that an album, regardless of format, was created from the original master tape and why is this important?
The first issue of an album is generally created from the original master. In the case of huge sellers, there will often be subsequent masterings. This is particularly true for vinyl because the vinyl lacquer created in the mastering room is only good for the production of so many “mothers” used for creating of the final pressings.
"What this means is that with hit albums, there is no guarantee a given purchased copy was mastered from the original master tapes."
In vinyl mastering, it has been common practice to make a separate tape recording during the initial mastering stage, which contains the same signal sent to the cutting lathe. In other words, this separate tape recording would capture the level adjustments, EQ and any other processing the mastering engineer uses while cutting the lacquer. The labels would keep this “EQ’d copy” (or “EQ’d limited copy”) and use these when subsequent lacquers are required. These could be cut “flat” (i.e., with no further alteration) and reflect the changes made during the original mastering. What this means is that with hit albums, there is no guarantee a given purchased copy was mastered from the original master tapes.
Aside from EQ’d copies created during mastering, sometimes the label is not in possession of the original mixed tapes. This is something I encountered a lot while at Atlantic. The originals might be in a musician’s private library or they might be in an overseas tape library. The label receives a “flat” copy of the original tape for use in creating the records or CDs they sell.
Is this important? There is no hard answer. A well made analog tape copy will exhibit a slight increase in background hiss and perhaps a very slight loss of transient speed. However, in my experience, these are of much smaller magnitude than the changes the mastering engineer is going to make. The question is much like asking whether purchasing produce from a gourmet shop will produce a better meal than purchasing it from a supermarket. If all other things are equal, I believe the answer would likely be yes. But all other things are rarely equal. If two different chefs prepare meals from the same ingredients, the results will often be quite different. Personally, I’d rather choose the chef than the ingredients. Put another way, I’d rather listen to a record George Piros mastered from a third generation copy than some other engineer might have mastered from the original studio mixes.
"Put another way, I’d rather listen to a record George Piros mastered from a third generation copy than some other engineer might have mastered from the original studio mixes."
To the issue of provenance, I believe what many folks really want to know is less about what mic cable was used in the original recording and more about whether the resolution of the download they are purchasing is truly the resolution it is purported to be. Is it really 24/96 or really 24/192? Or is it merely 16/44 (CD resolution) delivered in a high resolution “package”?
On your Soundkeeper Recordings site, you recommend CD-Rs over CDs for those buyers who will play back their music on a transport or player yet you also recommend the CD for those buyers that intend to rip their music to hard drive. Could you explain why a CD-R is better than a CD when spinning a disc and why this difference doesn't matter when ripping and playing back from hard drive?
I wish I could explain why. I’ve read a number of theories and some of them may or may not make sense.
From my earliest days in CD mastering, I always noticed that the finished CDs from different replication facilities all sound different from each other and none sounds indistinguishable from the CD master used to make it. Often, CDs made on different production lines within the same plant don’t sound like each other either. In all cases, there is a loss of “focus” and fine detail, usually subtle, sometimes not so subtle.
When it came time to choose a plant to do Soundkeeper’s CDs, I spoke with a few dozen facilities. The one I ultimately chose was the only one which, without any prompting from me, did not claim their CDs sound exactly like the masters. It turns out, their CDs are the closest in my experience. I can still distinguish between the CD and the master from which it was made but with their discs, I need a synchronized playback against the master to discern the differences.
This plant cuts the glass master (the first step in CD production) in real time, instead of the more typical 4x or faster used by most other facilities nowadays. They also use a ~9 second injection molding cycle, rather than the more common ~4 second cycle. Whether these account for why their discs are more faithful, I don’t know. Some say procedures like this make for better formed “pits” in the disc, making it easier for the player to read the disc with less “jitter” (i.e., timing errors). I don’t know if this is the case but I do know I like the results.
With a well made CD-R (burned at relatively slow speed on a high quality blank), I find the results of playback in a CD transport or player sound closer to the CD master than even the best pressings in my experience.
I think something similar occurs with processes such as SHM, Blu-Spec and HQCD, where the processes are different from usual and sometimes the materials in the disc itself are different. I recently compared some of these with their plain CD counterparts. I was pretty surprised by the degree of difference I heard and found it to be so obvious, I would have bet I was listening to two different masterings, with different EQ!
To “prove” this, I extracted both the “special” disc and the plain CD to computer hard drive so I could perform a “null” test. In a null test, two digital files are synchronized (to the sample) and mixed together. The polarity of one of the files is reversed. What results is that everything the two files have in common, i.e., what is the same in the files, is cancelled (or “nulled”), leaving only what is different between the files. To my surprise, the result of the null test was dead silence. Listening to the two files from the computer resulted in both sounding indistinguishable from each other. It was a slightly clearer version of the “better” disc heard from the CD player. Whether commercial CD, “special” material or process CD or a fine CD-R, my experience has consistently been that extraction to computer and playback from there (as a raw PCM file in .aif or .wav format) gets me the true sound of the master.
"What I do know is that as an audio enthusiast, I’ve always wanted to hear “the master” at home. With computer audio, this is finally a reality."
What is the difference between playback from a transport or player and playback from the computer? To create a CD, those “ones and zeros” of digital code must be further encoded, using a scheme referred to as “8:14 modulation”. This is used to create the nine different length “pits” and “land” (the space between the pits) on the finished disc. Among other things, the player must spin the disc at the correct speed, track the spiral of pits, keeping the laser properly focused, read the disc, decode the 8:14 modulation, decode the resulting binary code, apply any necessary error correction, convert it to stereo analog signals and feed it to the outputs, often using a common power supply for all these functions. The computer, given something like a raw PCM file in .aif or .wav format, has a much simpler job. Whether all this accounts for the audible differences, I don’t know for sure. What I do know is that as an audio enthusiast, I’ve always wanted to hear “the master” at home. With computer audio, this is finally a reality.
With this in mind, if computer playback is the goal, the advantages CD-R has in transports and players are no longer there, hence, my recommendation of the less expensive CD to Soundkeeper customers who listen via their computers.
You have been vocal about your preference for recording, as well as delivering, your Soundkeeper Recordings in 24-bit/192kHz format. Why 24/192?
The reason is because I feel properly done 24/192 crosses a very important threshold. Over the years, I’ve used all sorts of analog recorders and digital recorders but the output of these devices was always quite different sounding from the signal they received at their input.
Even the best 24/96 digital I’ve heard, while certainly much better than 16/44 CD in terms of fidelity the input signal, still sounds very different to me than the input that is coming directly from the microphones.
"For the first time in my experience, those reservations I have always had about digital, where I felt there were some things the best analog did better, simply evaporated."
When I first heard properly done 24/192, it was a jaw dropper. For the first time in my experience, those reservations I have always had about digital, where I felt there were some things the best analog did better, simply evaporated. This is, to my ears, a bigger jump up in quality over 24/96 than that was over 16/44. It no longer feels like a great digital recorder or a great analog recorder. It feels like the recorder has been effectively removed from the equation and I am listening directly to the mic feed.
I mention “properly done 24/192” because I’ve heard a number of converters with these numbers on their spec sheet, which actually sound worse to me at this rate than they do at 24/96. This, I attribute to the significantly increased demands made by the higher rates on clocking accuracy and for wide band performance from the analog stages.
When the higher rates are well executed, the results are simply magical. Though I hear it throughout the range, perhaps surprisingly, I find many of its benefits particularly audible in the bass. The only downside I’ve found so far is that I can no longer blame the gear for any flaws in my recordings. Of those, I must take full ownership.
There is continued discussion and controversy over dynamic range compression. Mainly, how do we know if a given recording has been compressed to the point of not being worth listening to or buying. Some people are using sites like the "Unofficial Dynamic Range Database" to gauge the level of dynamic compression used in a given recording and to help inform their purchase decisions. Do you have any advice for how to interpret things like a "DR Value" and generally does loudness, which you've written about, automatically equate to a poor sounding recording?
I very much applaud the efforts of the folks behind the Dynamic Range Database. I have long felt that preservation of musical dynamics (as opposed to their eradication in the ongoing Loudness Wars) is one of the last frontiers in recorded sound.
That said, I would hesitate to assign quality to a recording based on a number. While a low number will certainly indicate what many, including myself, would deem an undesirable curtailing of dynamics, a high number, in my view, is no guarantee of quality. It is important to remember tests like this are looking at only one aspect of what is a complex reality. For example, who cares if the dynamic range is high if the treble has been boosted to the point of being able to loosen dental fillings?
"Ultimately, in order to evaluate a recording, particularly how any individual might feel about that recording, I don’t know of any substitute for actually listening."
As to whether loudness automatically equates to a poor sounding recording, I think this is a matter of degree. I have some recordings I consider good sounding but which are still compromised, in my opinion, by dynamic compression. I can only wonder how they’d sound if their dynamics were left unhindered. Personally, I think they’d be better but I should add that these recordings were not at all eviscerated to the point that many others are nowadays.
Ultimately, in order to evaluate a recording, particularly how any individual might feel about that recording, I don’t know of any substitute for actually listening.
Beyond dynamic range and provenance, are there other important factors people should factor in when making music purchase decisions with an ear toward sound quality?
Keeping in mind that it is possible to have a high DR rating, with the source being the original recording itself and still end up with something that doesn’t sound very good, we’re back to listening as the only real way to tell.
I think it would be a mistake to interpret a DR rating of 20 as, in and of itself, being better than a rating of 19. If they are two versions of the same recording, I’d want to know the recording and mastering engineer for each, as that would tell me more than the numbers ever could.
Still, in the end, the recording has no other purpose than to be listened to, so I’d rather hear a sample in order to know if I want to buy it.
Some audiophiles have a stated preference for a direct-from-master approach, meaning no EQ applied during the final mastering process. Is this the ideal?
In most cases, I would say it is not. While I understand and appreciate the sentiment, this is invariably suggested by folks who have not heard many (perhaps any) masters.
When I first started as an engineer, having been an audiophile first, I too believed EQ is bad and that all masters should be transferred “flat” (with no alteration). Then I learned how most records are really made and got to hear how many masters really sound.
One must consider the microphones used in most recordings and the colorations they bring to the results. Combine this with where the mics are placed (generally in places where the listener would not want to place their ear) and the fact that the signal will then be sent over long lengths of not-so-great cables, through dozens and dozens of switches and patch points and then “adjusted” to make the result, played back on not-so-great monitoring, sound “right”, is it any surprise that most recordings need help?
"Another thing to keep in mind is that there is EQ and there is EQ; I think it took me 15 years to learn how to use it invisibly."
Put another way, if I’ve got a master that sounds thin, with a treble hyped to the point of keeping insects away and I can make the results sound less painful with EQ, I would find EQ to be quite a good thing. Another thing to keep in mind is that there is EQ and there is EQ; I think it took me 15 years to learn how to use it invisibly. The idea is, for example, to bring out the “bite” in a horn section, not to bring out 5 kHz. All too often, when we hear an EQ’d program, we hear the EQ more than we hear what it was trying to accomplish program-wise.
All that said, with some recordings, there is no need for EQ. I’ve worked with mixes that came in sounding so right, they needed absolutely nothing. And with some recordings, which were made with the intention of capturing life - usually the minimally mic’d, “purist” recordings - there is also no need for EQ. In such cases, if no minor level adjustments are necessary, a straight from the master approach is the one I’d choose.
I've seen you mention George Piros, the mastering engineer, on a number of occasions and was wondering if you could tell us a little about your experiences with him and why he's made such a lasting impression.
I had the very good fortune to work with George when we were both at Atlantic. He would often tell me of his days with Bob Fine and Wilma Cozart and the work they all did on the Mercury recordings, which I only later came to hear and then to love so much.
George is one of my engineering heroes, being one of the only mastering engineers I ever met who did not routinely add a compressor or limiter to his signal path. In fact, I never saw him patch one into his mastering channel.
I spent many an afternoon in his mastering room, watching him work and discussing audio and music. George always called it as he heard it, no matter who he was speaking with. I loved his directness and his passion. And no matter what, he always served the music.
One memory that will always stay with me came from a time shortly before George retired. I was walking past the outside of the padded, double-door “airlock” outside George’s room and heard some rocking music coming from inside. I entered the room and saw George bent over the lathe, peering into the microscope, looking at a groove he was cutting in a test lacquer... while AC/DC played at a level that could peel the paint from the walls.
Through George, I also got to meet Bert Whyte and Joe Grado, both of whom came up to George’s mastering room to work with him and share audio stories.
In addition to being involved on the recording and mastering side, you are also an audiophile. Do you think this helps inform some the choices you make in the recording studio?
It has always been key in everything I’ve done as an engineer. From the realization of the critical importance of monitors and their setup to all the other components in the chain, being an audiophile has helped me develop as an engineer.
"It is why, when on an AES panel of CD mastering engineers in the early ‘80s, mine may have been the only voice in the room to (shyly at the time) declare my vinyl records still brought me closer to the music."
It has shaped my sensibilities in terms of what I want to achieve in my work and is the reason I refused to take part in the Loudness Wars. It is why, when on an AES panel of CD mastering engineers in the early ‘80s, mine may have been the only voice in the room to (shyly at the time) declare my vinyl records still brought me closer to the music.
And it is what led me to form Soundkeeper Recordings, where I believe I’ve done my best work to date.
Are there an upcoming projects you'd like to share with us?
With Paul Beaudry & Pathways’ “Americas” released just last month, I’m still in the “promo” phase for that project but there are a few new ones under consideration, which might come to pass later this year and see release in the next.
One is a reggae project - perhaps the world’s first “purist” reggae recording. I’ve also proposed a project with a young alto saxophone player I heard last month, who really entranced me with his playing. I’m always on the lookout for artists whose music moves me and who are interested in trying the “recording without a net” approach.