The Demery Chronicles (no.3)
The souls of the righteous
Following our demonstrations in New York I returned to Holland and waited for the call from Tom Jung. The SA-CD Scarlet Book specifications were not locked down at that point, and lots of concepts were being tried out, amongst them the idea of using DSD with a sampling frequency of 48 Fs (i.e., 48 x 44,100 Hz) for the multi-channel portion of the disc. To cut a long story short here, by the time we were ready to carry out comparative listening tests with the “golden ears”, the lossless compression experts had discovered that 64 Fs DSD compressed better anyway, and so an all 64 Fs disc was decided upon. There was lots of activity, so I wasn’t exactly sitting around with nothing to do: we had more test recording and development planned. In December 1997 I traveled to Vienna for the recording of Beethoven’s 4th Piano Concerto at the Musikverein. Conducting the Wiener Philharmoniker was Sir Simon Rattle, tinkling the ivories was Alfred Brendel. This session was far from easy.
The Musikverein is heavily in demand for concerts and recordings. There were so many events taking place that we would have to set up the microphones and lay all the cables each night when the public had left, record the following morning and early afternoon, then tear everything down and wait until 11 p.m., or later, and do it all again. As a result, balance engineer Erdo Groot decided against setting up a multi-channel monitoring room to try to cut down on the mayhem. Sadly, this would prevent us from being able to use these recordings for release. As the start of recording approached, producer Martha de Francisco was not 100% happy with the piano sound. Some last second piano microphones were added, cables were run, connections were made, Erdo made the necessary adjustments to his stereo mix, and recording began.
At the end of the day, we would realize that the two cables that had been pulled to make way for the piano were the feeds from the surround microphones. Due to the way the cables were routed, all the recordings of the first movement contained the piano spot microphones instead of the hall ambience. It was a terrible shame, but understandable under the circumstances, and the price you pay for recording “blind”.
Despite this hiccough, we continued to record both the sections with orchestra, and the solo piano patches. The 2nd and 3rd movements turned out beautifully. They are among my favorite DSD recordings, and heard in multi-channel are absolutely captivating. Like Rattle, Brendel vocalizes during his performance. The patching sessions for the solo piano parts of the 2nd movement were spookily real. You felt like a fly-on-the-wall eavesdropping on something very private. Brendel would talk to himself, try things, and continually refine his playing to his vision of this beautiful music. Then you start to notice something coming from the rear of the hall. His concentration now broken, he starts to lecture the cleaning staff who had inadvertently entered (despite the RECORDING IN PROGRESS signs on every door). Angry, he shouts, “Scheisse!” gets up from his piano stool and storms out. Perfectly captured in the multi-channel recording, you can follow him as he stands up, leaves the stage by the steps on the left, walks along your left side, until he cuts across the acoustic behind your head and exits through a right-side door midway in the hall.
Roll out the barrel
Vienna in December is beautifully picturesque. Watford, North London in February is not! Despite that, we were back at the Colosseum in February 1998 to record Haydn’s Schöpfungsmesse (The Creation Mass). Conducting once again would be Sir John Eliot Gardiner, performing would be Ruth Ziesak (soprano), Bernarda Fink (alto), Christoph Prégardien (tenor), Oliver Widmer (bass), the Monteverdi Choir and the English Baroque Soloists. This time we would set up two control rooms: one for stereo and one for multi-channel.
I’m sure most everyone has seen photographs or video of their favorite pop/rock artists in a control room at one of the World’s leading recording studios. Some of them can be very opulent; beautifully furnished with every convenience on hand. Most classical music is recorded in concert halls. Short-sighted 19th century architects did not think to add a spiffy control room with large cable ducts to their designs. As a result, an on-site classical music control room is usually some unused room at the back of the hall, far away from the stage and even further from the amenities (hot and cold running water in the better places!). At Watford, a control room has been fashioned from some upstairs offices. There is a winding stone staircase to get to it. Carrying two 400 lb Sony DASH multi-track recorders up these stairs is an ordeal!
When it is not being used as a classical music recording venue, the Colosseum is used as a discotheque. The multi-channel control room was to be the downstairs bar. It had the advantage of being close to the hall (easy for laying cables), on the ground floor (so no heavy carrying), had good dimensions for a listening room and was close to the bathrooms! The downside was that it stank of spilled beer and old cigarettes!! Still, after a few days, you stop noticing it!!!!
Is the coffee ready?
Erdo had brought an extra mixing console to this session and full 5-channel monitoring. Balance engineer Jan Wesselink was going to be Erdo’s ears, as Erdo manned the stereo (CD-release) room. Erdo fed Jan his stereo mix, and a telephone link was set up between the rooms. For this recording, the orchestral players were on the floor of the hall with the choir arranged on risers behind them. The four soloists were positioned behind the conductor, singing towards the orchestra. Erdo’s main pick-up came from 5 omni-directional microphones (probably B&K 4006) situated with the center over the conductor’s head, placing the soloists in the circle of main microphones.
We chose to use this recording as an experiment to determine how the microphone arrangement performed, and how to optimally combine the spot microphones for the best end result. Up to now, our tests had all been straightforward 5-microphone recordings. The purists will tell you that that is the only way to record. Indeed, they will claim that you really only need two microphones to perfectly capture a stereo image of an orchestra. Well, let me tell you that I do not agree with that concept at all.
Safely ensconced in our beer-smelling control room, we were free to experiment. We probably had some 30 microphones in use, and were using one of Philips Classics’ (now Polyhymnia’s) state-of-the-art custom analogue consoles, with custom microphone pre-amplifiers and custom microphone power supplies (in fact, only the capsule and housing of each B&K, Schoeps and Neumann microphone was original) all designed and built to Philips Classics’ specifications, and custom microphone cable made by van den Hul. We began listening to just the five main microphones. The sound was very good, very spacious, and, despite their real positions in the hall, the soloists “appeared” spread across the front of the image with the orchestra and choir behind them. However, there was a sense of something missing, that the “flesh” wasn’t really on the body in some cases.
This is where the skill of the balance engineer comes in. Despite having 30 microphones in place, no one says you have to use them all. However, it is far better to have them in place than have to interrupt the session when everyone is in the “groove” because you are not getting the definition you want! So, Jan would carefully blend some of the spot microphones into the mix. The set of microphones used would change from movement to movement depending on what the score required in terms of instrumentation and number of soloists. Through careful blending (at a relatively low level) of the spot microphones with the main “pick-up”, Jan was able to maintain the spaciousness while putting some “flesh” on the bones, so to speak.
As the session was being held for a stereo CD recording, control of the set up and orchestral/choral placement was a juggling act for Messrs. Gardiner and Groot. If the conductor, producer or engineer are not happy with the sound, it has to be decided whether the problem can be resolved by having a section play louder/softer, or by moving them closer to/further from the microphones, or by raising/lowering the soloists, or “simply” moving the microphones. We could follow along with the discussions by listening to the conductor/producer (Nicholas Parker) interchanges. What was fascinating was that recording good-sounding stereo seemed to be more difficult than recording good sounding multi-channel. Each time a discussion took place, we would switch to Erdo’s stereo mix which he had also piped to our room. Sure enough, there were occasions where something would be lacking in stereo (masking due to signal build-up in each stereo channel is a problem), yet when we listened in multi-channel the problem would be gone or considerably less present than in stereo. The music always seemed to open up more in multi-channel – it had room to breathe!
After each adjustment, Jan would re-balance, and then the phone would ring. It would be Erdo, explaining the changes and asking if things were still okay for us. Jan would report that we didn’t have a problem in the first place, and that having compensated for the changes, things were still hunky-dory. (Erdo would experience the same thing at later recordings in Budapest where he manned the multi-channel room while Hein Dekker would struggle with the balance in the stereo room.)
One thing that the recording team took as seriously as their art was their coffee. No instant Nescafé out of a jar for these seasoned professionals. They took their own coffee percolator in its own custom flight-case, and a separate large flight-case full of strong, Dutch coffee (Douwe Egberts, of course). On trips to some countries they would even take their own bottled water! It amused me that amongst all this high-technology, these two cases were always the first to be opened. My own colleagues at Philips were also coffee-fanatics, and seemed to drink a cup about as often as I drew breath. With the physical separation between the two control rooms, the phone link became the “coffee hot-line”. All conversations either started or ended with the questions, “Is the coffee ready?” or “Is there any coffee left?” For a non-coffee drinker like myself, it was amusing to see something so mundane being treated with such reverence. Fortunately, the strong coffee smell helped dull some of the odors down in our bar-come-control-room.
How to make friends and influence people
One of the side effects of being located so close to the bathrooms is that you receive a lot of visitors. During the pauses, the soloists and musicians would apologize for interrupting (as if) so that they could make their way to the restrooms. Naturally, the presence of a second recording team, and one with more loudspeakers, would pique people’s interest. After briefly explaining what we were doing, we would ask if they would like to listen to a playback. Every time there would be a look of amazement on the person’s face, and word quickly spread because pretty soon sections of musicians were dropping by – we were much less formal than the stereo room where the serious discussions between conductor, producer and engineer took precedence over a quick playback of the horn’s 2 bar solo! Overwhelmingly, the performers could not get over how much better the instrumental timbre is when listening in multi-channel – a quick switch back and forth between stereo and multi-channel was very convincing.
As the sessions continued, we fell into a regular pattern. For each movement, we would first record only the main microphones then begin to add the spot microphones on subsequent takes. In this way, it was easy for the balance engineers and producers to contrast the effect of one microphone set-up with another – of course, we were greatly helped here by the professionalism of the performers who were able to do take after take with considerable similarity in their singing and playing.
Expect the unexpected
With our first fully monitored multi-channel session in the bag, I headed back to Holland and a call from David Kawakami. Tom Jung (TJ) was ready to do the project that had occurred to him after hearing our first demonstration in New York. Could I travel to the US with the equipment? Of course I could! Plans were made, and I arrived in New York City in May. As the Philips recorder and I were unknown quantities for TJ, he had asked whether he could first spend a few days evaluating the equipment before the recording dates. No problem! TJ wanted to do the evaluation at his home studio in Stamford, CT.
A phrase that I came to hear often from David Kawakami is “due diligence”. It would come to encompass many aspects of our work, and, as the boss of all things SA-CD in the US, David too wanted the project to run smoothly. So, he kindly offered to drive me from NYC to Tom’s house. With my bags ready outside my Manhattan hotel, I waited for David to pick me up. Looking out for an executive-type automobile, my attention was broken by a beep-beep. There was David behind the wheel of a Ferrari-red Geo Metro! I threw my bags in the back and jumped in. “Welcome to the Red Rocket!” David declared. We weaved our way through Manhattan’s congested streets until finally, on the freeway, David was able to unleash the full power of its 993 cc engine. It was an exhilarating ride, and at some points (mostly downhill) we threatened to touch the speed-limit. However, top speed would be diminished from time to time as David pushed the cigarette lighter in, then each time it popped out the turbo-like rush of power would rocket us back up to close to 60 mph. I imagine this is what Chuck Yeager experienced in his X-1!
As we pulled up to TJ’s house a familiar sight greeted us – one of Sound Moves’ trucks. We unloaded the 15 flight-cases of equipment, and then started figuring out where everything should go. The noisy PCs and hard-disk arrays plus the empty cases would go in TJ’s garage, the ADCs, DACs and computer control in his studio. Since we were using very thin fiber optic cables to connect between the converters and the PC sub-system, it was an easy task to run the fibers around the walls and under the doors of TJ’s house. His wonderful wife, Jean, must be so used to “equipment invasions” as she did not bat an eye as I started running the bright orange fibers all over her lovely home.
With the equipment unpacked and connected, we hit the first snag. As an early advocate of digital recording, TJ’s home studio set-up was based around a digital mixer and editing system. At that time, no such “toys” for DSD existed, so we needed to create an all analogue monitoring chain, but TJ had no analogue multi-channel pre-amplifier or mixer and neither did I – in fact, this was a problem we would encounter over and over in the early days of SA-CD until ace designer Ed Meitner developed his first Switchman. So, what to do? Fortunately, TJ’s monitoring system was based on active Hafler monitors, though customized by him to include all Meitner internal wiring. We decided to go purist. We hit Radio Shack and bought a bunch of metal boxes, rotary controls and some 10 kohm potentiometers. Within an hour, or so, we had five nifty little in-line volume controls that we attached close to the balanced input of each active speaker. With the help of some pink noise, we found a reference level, marked all the boxes with a Sharpie™, and were set for some listening. We had no master volume control, so David and I would have to hop around the speakers to adjust levels as TJ measured and listened from the sweet-spot.
We spent a couple of days going over every aspect of the system, and the impact it would have on the recording session. The Philips DSD recorder had been designed as a research tool. It was not intended for production recording, and it had some major operational quirks, but since there was no alternative, we had to work around its idiosyncrasies. One of the most annoying aspects was that the switch from recording/input monitoring mode to playback mode was done via a set of switches on a small PCB attached to a parallel port on the front of the data interface unit. Each time you wanted to switch modes you had to mute the monitors (to avoid any nasty noises), walk to the noisy equipment, set the switches, walk back, then reset the monitor level – a pain when you have 5 volume controls! Of course, another key consideration was the fact that no multi-channel DSD editor existed, so to make a release out of the upcoming recordings would mean requiring “perfect” takes from the performers.
Tom & Tom and Paul & Paul
With the familiarization period over, we packed up the gear and TJ’s son, Paul, arrived with a truck for the drive up to Hartford, Connecticut where the recording would take place. The location was the Chapel on the campus of Trinity College. The Chapel is very pretty (reminding me of many traditional British churches), isolated from the rest of the campus buildings, and set back from the road a bit. The last fact would be the most painful, as it meant we had to hand carry all the equipment along an old stone path, into the chapel, then down a stone staircase into Crypt chambers below the Sanctuary, where the Chapel choir held practices. It was quite a hike! Anyone who is interested in the venue should take the virtual tour at:
Inside the Chapel, David Kawakami (who would also attend all the recording sessions) and I were introduced to Tom Bates who would be co-producing this project and handling all the discussions with the artists while TJ worked as the balance engineer. Like TJ, Tom is a music person through and through, and shares TJ’s quiet demeanor, grace and ability to never get ruffled even in the tensest situations.
For the recording, TJ had clear ideas about most things – he would use a Decca “tree” to handle the front three microphones, would add surround via two spaced microphones according to the ITU multi-channel specification (though the antiphonal pews in the Chapel would make it impossible to freely choose the position of the rear microphone stands), and would record a sixth height channel to boldly go where no recording had gone before! For his microphone pre-amplifiers he had turned to Grace, while Millenia had provided the 6-to-2 mixer needed to create the stereo mix. The only question-mark was which microphone(s) to use. TJ’s reputation is such that both Shure and Earthworks had each lent him six brand new prototypes. The KSM-32 cardioid in Shure’s case, and an omni-directional model (possibly the QTC-1, I do not recall exactly) in Earthworks’ case.
TJ set up two stereo pairs, as close to co-existing as possible. Then he and Tom listened as the choir (Gaudeamus, though known at that time as the Paul Halley Chorus) rehearsed. Switching back and forth they both listened intently. As is often the case, they found themselves liking different things in each microphone, and really wanted a combination of both, but, in the end, they decided to go with the Shure. So, all 6 were unpacked and hung in their special webbed holders (to minimize feedback from vibration).
I was as surprised as anyone to find myself in the bowels of a church about to record a variety of traditional and contemporary religious choral works, especially with an engineer whose record label was known for jazz. However, this is typical of TJ, and he would go on to surprise us with other projects that aimed to test the capabilities of converters and recorders. Female voice, in particular, is tough to record, especially when the sopranos trigger standing waves due to the reflective nature of the Chapel’s structure. Standing in front of them at rehearsal, the sound could sometimes become so intense that it felt like someone was sticking knives in your ears.
Holidaying at the Holiday Inn
The first day of any recording session is typically the most stressful. The equipment has to be unpacked, assembled, tested, microphones need to be positioned, cables must be laid, the correct sound balance has to be found, and, if that wasn’t enough, the producer needs to coax killer performances from the artists. So, at the end of a long first day, everyone was ready to crash. Since we had gone directly from TJ’s house to the Chapel, it must have been after 10 p.m. when we arrived at our Holiday Inn. This wonderful hotel was built in the intersection of several freeways, one of which was under round-the-clock construction. It was also holiday time, and the hotel was full of teenagers on school outings. The foyer was packed with noisy kids as we showed up to check in. Tom Bates was helped by one receptionist, TJ and Jean by the other. After a prolonged check-in, it was my turn. Before beginning the process I insisted that everyone get off to bed, and we made arrangements to meet the next morning. Then I presented my reservation details, my passport, my credit card, filled in the necessary forms, and eventually received a room key.
Surrounded by lots of wound-up teens, I rode the elevator to my floor. Ahhhhh, peace at last! I strolled down the corridor to my room. Checking the door number with the number on my key, I was a little surprised to hear voices coming from inside MY room. I checked the details again (after all, I was pretty tired at this point). After a few minutes of wondering what to do, I knocked on the door. A few seconds later, a surprised TJ opened the door. “Hi Dem, did you forget something?” “Uh, no, it seems that we’re rooming together!” The look of bewilderment on TJ and Jean’s faces was a picture. I said, “Leave this to me, I’ll sort it out; see you in the morning.”
Back in the teen-filled lobby, I explained my predicament. The receptionists looked at the computer, they gave quizzical looks at each other, they excused themselves and called for the night manager. She listened, she looked at the computer, she tapped some keys, she apologized, and issued me with two new keys: one for my room, and one for the elevator. Funny, I didn’t need an elevator key the first time.
I inserted the key, punched in the floor number and in seconds was on their “executive” floor. It was 100% teenager free. Bliss! I found my room, inserted the door key, and stepped into a gigantic suite. I had been upgraded. Having seen how tiny TJ’s room was I immediately called him. After apologizing for yet another disturbance, I explained that I had been well taken care of this time, and offered to swap rooms. However, he explained that they were far too tired to care at this point, and told me to enjoy my stroke of luck.
Over the course of three days, we recorded 18 pieces of music. Some of the more complicated pieces were impossible to capture in a single unblemished take, so did not make it onto the final album. Of course, among them were three of my favorites. Barber’s Agnus Dei is a sung version of his most famous composition: Adagio for Strings (or the theme music to the movie Platoon). Then there was Vaughan Williams’ Lord, Thou Hast Been Our Refuge. This was the only piece that had musical accompaniment. The choir was arranged in the Sanctuary of the Chapel with the front microphones about level with the pulpit. A trumpet player stood mid-way in the Chapel playing towards the Sanctuary, and the organist was situated at the West End – the full length of the Chapel from the choir. The piece is about 8 minutes long, but the trumpet and organ do not make an entrance until about 4 minutes in. When listened to on a properly set up surround system, you can clearly locate the trumpet directly behind your head, while the organ seems to fill the acoustic behind and alongside you. It was quite thrilling.
The final unused piece that sticks in my mind is Tavener’s Song for Athene, a haunting tribute to a young family friend it is now, perhaps, best known for being sung while Princess Diana’s body was carried out of Westminster Abbey. The work calls for the basses to sound a wordless, continuous drone, while the sopranos, altos and tenors sing over the top. To avoid gaps, the basses were staggered in their breathing, and were placed further back in the Sanctuary so that the voices would meld better. We recorded a few takes, but faced an unforeseen problem in the sounding of the Chapel’s clock every 15 minutes. For the final planned take Tom and choirmaster Paul Halley discussed what they felt needed to be done. I indicated to TJ that we needed to get started, otherwise we would have to wait. TJ diplomatically informed Tom, who diplomatically informed Paul. With the recorder rolling we were all ready to go when one of the sopranos decided to ask a question. The sound of the groans in the control room was sadly not recorded, however, with the question answered, the choir began singing.
The piece is about 7 minutes long, and it was about 7 minutes to 5 o’clock. It was going to be tight. After the crescendo towards the end, calm is returned as the tenors sing “Alleluia” over the unwavering bass drone. As the tenors finished their last “Alleluia”, the clock started striking. Thinking quickly, Paul Halley kept the bass drone going while holding everyone else silent. The clock chimed 5 times, and when the sound of the bells had decayed, he brought the basses down to silence. Then follows about 5 seconds of absolute silence before everyone spontaneously bursts into laughter. It was a complete fluke that the chimes occurred just at the right point in the score. Unfortunately, the piece is so difficult to sing correctly that none of the unedited takes could be used, but we did demonstrate this version to a few lucky people in the early days of SA-CD promotion.
Of the songs that did make it onto the Sacred Feast (DMP SACD-09) album, two stand out for different reasons. The first is Messiaen’s O Sacrum Convivium. I think it is the prettiest piece on the album, but what makes it special is the rain. Some thunderstorms moved in one day, and we had a few spells of very heavy cloudburst. If you listen very carefully to the SA-CD, especially in multi-channel, you will hear the heavy rain beating on the slate roof and the stone courtyard outside.
The other song, of course, is Durufle’s Ubi Caritas. One of the biggest problems in demonstrating SA-CD was finding music that anyone could appreciate and instantly hear the improved sound quality. You’re looking for something that will knock people’s socks off and make them either want to make SA-CDs (if they are an artist/producer/engineer) or buy them (if they are a consumer). Too many audio demonstrations, in my opinion, either use inappropriate music or set the playback level far too high. We quickly built a reputation in the SA-CD project for striking the right balance, and one of the pieces of music that set the mark was Ubi Caritas. At about 2-and-a-half minutes, it is an ideal length, as it can be played unedited even when demonstrations needed to be kept short. Also, as a pretty, unaccompanied choral work, it is a style that every listener will be familiar with. We first started using it in demonstration in the Summer of 1998, and, in contrast to most other hi-fi demonstrations at the time, we used it as the finale. Its simplicity and beauty totally disarmed many listeners, and brought several to tears. I can’t count how many times I have played this track since then, but it is a testament to the music, the performance, and the recording that I still enjoy it every time.
The bleeding edge
When our 3 days of recording were over, it was time to move back to TJ’s house for a quick listen to all the takes before I took all the equipment back to Holland. This was a bizarre situation for TJ as, after waiting all this time to do the recording, I would have to take the only machine that could play it away from him. Fortunately, Tom Bates had made some parallel PCM recordings, so they could do a more leisurely take analysis that way.
As I stated earlier, there was no workstation on which to edit, so TJ had to choose the best complete takes, then tell me the running order, and I pieced the album together with my colleagues in Holland using a DSD crossfade package that ran on a powerful Sun computer. However, this was a case of “razor-blade” editing the files, overlapping them slightly, and then having the computer calculate the crossfades. Like our first recordings, this process was blind. We only knew how the files sounded after the program had run and the data had been transferred back to the recorder for playback. We also had no easy way for TJ to assess the fades.
However, after a few iterations, we had done it, and the World’s first fully-loaded stereo CD, stereo SA-CD and multi-channel SA-CD hybrid could be made. It had taken 8 months from first meeting TJ to actually making the recording, it would take another 18 months, or so, before the physical discs were made. It had been a struggle, but life on the bleeding edge always is. I’d like to finish this part of the story by saying that after this things only got easier, but nothing could be further from the truth as TJ and others like him dreamed up projects to push the technology to the absolute limits. One thing was for sure, life in the world of SA-CD recording was going to be anything but dull!
Next: A date with Princess Leia