DVD PRO Conference
EMedia Expo
Current Issue
Article Archive
News Archive
Buyer's Guide
nbsp; Home Magazine eNewsletters Events Contact Navigation

EMedia Buyer's Guide

Current Issue Current Issue
DVDs in the Works


News Indices

CD TrackerCD/DVD-ROM IndexFact, Figures & FindingsConference Calendar

eNewsletters
DVD TodayDigital EarfulTechVideo NEWSSubscript
Ad Links


Surround from the Ground Up: The Making of a DVD-Audio Title

Mark Waldrep

The call from Karl Held, an audio producer I knew from the early days of CD-ROM development at Warner New Media, came only two weeks before the sessions. Larry Kraman, the owner of the Newport Classic label, had recently visited me in California and was impressed with the work that we were doing with DVD and subsequently mentioned AIX to Karl. Was I interested in doing a recording of Beethoven's 6th Symphony and Respighi's Pines of Rome for the New Jersey Symphony Orchestra (NJSO) at the New Jersey Performing Arts Center? It seems that another record company had pulled the plug on their project at the last possible moment and he needed a quick solution. Musicians had been booked, the union guys put on alert, and the hall readied for a two-day session in September.

Of course I was interested. But, as I suggested to Karl and then to orchestra director Karen Swanson, maybe they should consider making a DVD-Audio recording rather than releasing yet another version of two fairly well-known symphonic works. As a strong advocate of the new DVD-Audio format and the owner of a specialist recording label (AIX Records), I suggested making a high-resolution (96kHz/24-bit), multichannel (24-track) recording as well as a video document of the session for release through my distribution arrangements. After further discussion of format and strategy, they became convinced that the NJSO could get more attention from a DVD recording than from releasing yet another compact disc. So began one of the most challenging recordings I've ever been involved with.

Step one was to secure the equipment and personnel required to pull off a high-resolution recording in Newark, New Jersey. AIX Media Group, our parent company, has an office in Manhattan, but the equipment necessary for a session like this is all located at our recently upgraded, high-resolution studio in Los Angeles. The initial plan was to fly the production team to Newark, rent the gear locally, and transport it to the New Jersey Performing Arts Center (NJPAC) for the two days of recording. The plan looked great on paper but as we were soon to discover, securing high- resolution audio equipment is no small task. We were good on the microphones and preamps, cables, video equipment, switchers, and monitoring gear; however, the actual recording machinery presented a bit of crunch. It seems our pioneering attitude hasn't yet captured the imagination of New York's rental houses.

HiRes: A Brief Introduction

For those unfamiliar with the world of high-resolution recording, allow me to present the following brief primer. Since the introduction of the compact disc back in 1982 with the release of 42nd Street by Bill Joel, all digital audio that has been delivered in that format has been encoded using Pulse Code Modulation (PCM), operating at a sampling rate of 44.1kHz and creating "words"-a.k.a., quantities of sampled information-16 bits long. This is considered standard resolution, meets or exceeds the ear's frequency range, and is close to the dynamic range of our human equipment. However, from the start audiophiles and consumers alike have questioned the actual fidelity of the CD products being released by most record companies. The reproduction of the high end has been described as "harsh" or "brittle." Other complaints and the increasing sophistication of computing equipment (hard drives and processor speeds) have driven equipment designers and audio engineers to seek methods for achieving better fidelity. Using PCM as the preferred methodology, the consortium of companies that launched the DVD-Video specification four years ago has introduced a new flavor of DVD discs which increase both the sampling rate and word size. The recording industry now has a high-density optical disc specification of its own: DVD-Audio.

The higher the sampling rate and the more bits allocated to each sample, the higher the fidelity of a digitized analog signal. Current CDs deliver frequencies up to half of the sampling rate or 22.05kHz, well above the range of human hearing. On the dynamic-range side of the equation, each digital bit adds 6dB to the potential signal-to-noise ratio. For CDs, that number is 96dB, or 6 times 16 bits. It turns out that real-life loudness can exceed 96dB SPL (even though you wouldn't want to experience it for very long), thus the need to allocate bits beyond 16. Recent trends in professional equipment have increased the number from 16 to 18/20 and now to 24. Using 24-bit equipment has the potential to deliver signal-to-noise ratios nearing 144dB, well in excess of the human ear's ability to accept it.

The DVD-Video specification can produce audio that is of very good quality, although when delivering 5.1-channel surround it does so using "lossy" encoding schemes developed by Dolby Labs or DTS. Prior to the DVD-Audio format, the only ways to hear digital audio with better fidelity than CDs was via SACDs (Super Audio Compact Discs from Sony/Philips) and the 96kHz/24-bit stereo capabilities of most DVD-Video decks. In fact, some high-end, specialist record companies (including AIX Records) are making a business of providing "original source quality" by transferring "classic" two-track masters at 96kHz/ 24-bit for release on DVD-Video. The discs contain no video, just PCM audio at high resolution.

DVD-Audio offers a large range of sample rates and word sizes using PCM to support older digital masters and the newer high-resolution releases. Sampling rates available for DVD-Audio producers include 44.1, 48, 88.2, 96, 176.4, and 192kHz with word sizes of 16, 20, and 24. The two highest sampling rates are restricted to two channels, but the others can accommodate up to 5.1 full-range channels. When choosing 96kHz/24-bit, it's necessary to make use of a clever "lossless" algorithm developed by Meridian, a high-end equipment manufacturer located in England (its cheapest DVD player goes for something over $10,000). Its algorithm is known as Meridian Lossless Packing (MLP), or sometimes as Packed Pulse Code Modulation (PPCM); all DVD-Audio machines decode MLP audio streams internally and deliver high-resolution music at the analog outputs.

One final bit of background on high-resolution audio: The recording industry has come up with a new slogan and logo to identify 96kHz/24-bit releases: "Advanced Resolution." The "S" shaped logo is prominently displayed on the front of most DVD-Audio releases from Warner Bros. The use of such a logo would seem to serve notice to consumers that the product behind the booklet was "recorded" using high-resolution equipment and that the full dynamic range offered by using 24-bit word lengths is in effect. Unfortunately, however, nothing could be farther from the truth. If the source recording was done, say, 30 years ago, when the dynamic range of analog tape machines was limited to about 60dB, there is simply no way to increase that number to take advantage of the additional resolution offered by DVD-Audio. What you get is 10-12 bits (remember 6dB per bit?) of music and roughly 12 bits allocated to reproducing the tape noise. Any older master tape remixed and transferred to 96kHz/ 24-bit is limited in dynamic range as compared to recordings that are made using high-resolution recording equipment when the musicians are present. Thus, the need for an entirely new generation of recordings that originate at 96kHz/24-bit.

Planning for two days

at NJPAC

Euphonix, a Palo Alto-based manufacturer of high-end professional audio recorders and consoles, supplied the equipment that AIX Records recently acquired for its high-resolution studio in Los Angeles. The company had previously donated a high-resolution digital multi-track machine called an R-1 for a show we produced in Los Angeles featuring Kim Wilson and the Fabulous Thunderbirds, and worked with us during our purchase of an R-1 and System 5 high-resolution digital console. As luck would have it, Euphonix was able to make available a 24-track rig for use in the NJSO project on the dates in question. AIX packed a flight case full of formatted 18GB drives ready to plug into the R-1, and off we went.

Meanwhile, Dreamhire in Manhattan supplied most of the microphones and preamps, a collection of large-diaphragm Neumann U-87s, and several B&Ks; the NJPAC provided a couple of Schoeps and AKG mics to round out the total of 24. We used preamps from John Hardy, Neve, Focusrite, and Audio Upgrades, all world-class examples of electronic design. The recording path was to be as direct as possible: no recording console of any type was placed between the mics, the preamps, and the R-1. We did use the small mixing deck in the control room at the venue to fold the 24 outputs of the recording machines to stereo for a DAT recording and to monitor in through the house system. Producing a multitrack recording destined for 5.1-channel surround means that you don't get to hear what you've got until you get back to the studio and play back the tracks through a console into your surround monitoring system. Our primary responsibility during our six hours with the orchestra was to capture the performances with as much fidelity as possible and figure out what we had later...a task that proved much more difficult than anyone could have imagined.

So far, there are very few DVD-Audio machines in the homes of music fans in the U.S. The first units were introduced in June 2000, and at this writing some eight months later, only 30 or so software titles are currently in release. Almost every available title is a rehash of an older recording and was produced without video. The discs might include a music video or interview along with the audio tracks of a DVD-Audio title, and most record companies are including Dolby Digital versions of the high-resolution mixes to ensure compatibility with DVD-Video players. Meanwhile, DVD-Audio is being promoted as the successor to the compact disc.

For the NJSO orchestra session, the plan to produce a disc uniquely and explicitly designed for DVD-Audio included producing a four-camera video shoot on the second day. The AIX Records release offers the high-resolution audio with still photos on the DVD-Audio side, and then presents the same tracks encoded in Dolby Digital and DTS with video on the DVD-Video side. Once you go to the trouble of assembling the musicians and compiling all the recording equipment, it's not much more effort to take it further by videotaping the sessions as well. On Sunday, we asked the orchestra musicians to wear concert dress in preparation for the cameras. Bexel in New York supplied the cameras and a local video director and crew were hired to handle the cameras. The production was getting bigger all the time.

Friday Evening:

The Concert

Karen and Karl invited me to the concert on Friday evening to hear the program that we were to record the following two days. On the program were Beethoven's Symphony No 6 in F Major, Op. 68 and Respighi's Pines of Rome. I've known both of these works for many years and was looking forward to hearing the orchestra, the tonal qualities of the hall, and to get to know the facilities that would be our home for the next two days. The Beethoven symphony is a standard repertoire, well known because of its use in the original Fantasia of 1932 produced by Walt Disney with Leopold Stokowski and the Philadelphia Orchestra. The Respighi is less well known, but an extremely popular work from his trio of programmatic works about Rome. Pines was composed in 1924 and is notable in that it directs members of the brass section to play from off stage and from within the hall as well as in its use of a prerecorded tape of nightingales. Perfect for a 5.1-channel surround recording, Respighi was obviously an advocate of music coming from all around the listener (he's not alone in that regard; Gabrieli did it repeatedly in the 16th century).

After the concert, I had the chance to meet Maestro Zdenek Macal and explain what we were going to try and accomplish the following day. I was initially worried that Karl and Maestro Macal would hesitate at the thought of 24 microphones and my plan for an aggressive mix, but both turned out to be extremely enthusiastic and supportive of the strenuous requirements of making a mix of this type. It seems only audiophile "purists" regard instruments coming from behind the listener as somehow inappropriate. For most musicians, being immersed in the middle of the music is a technological and creative luxury that DVD-Audio and 5.1-channel surround can offer for the first time. The plan was never to simply "document" a performance within the NJPAC. Instead, we wanted to present the music of Beethoven and Respighi in a fresh and interesting new way-one that takes advantage of the available technology.

Saturday:

Load in and Launch

The rented van showed up with all of the gear at about 10 a.m. The session was to start promptly at one p.m. and continue until five p.m. with the mandatory union breaks interspersed throughout the afternoon. The R-1 consists of about four or five individual units that were each packed in their own cardboard box. There is a pilot computer, a studio hub, drive units, and two units for analog-to-digital and digital-to-analog conversion. A flat-panel screen, keyboard, and mouse were also needed to interface with the system. All of the equipment was loaded in the control room or placed on the stage depending on its function, and members of the team went to work assembling the Euphonix rig or placing microphones and running cables.

On stage, the microphone layout diagram called for 12 pairs of mics for stereo recording to be dispersed among the various sections of the ensemble. We placed five pairs in the string section (Violins I & II, Viola, Violoncelli, and Basses), three stereo pairs in the woodwind section (Flutes, Oboes, Clarinets, Bassoons, and Horns), and the remainder over by the brass and tympani. We also put a couple of microphones in the hall to capture the room ambiance. The intent was to capture areas of the orchestral sound in discrete pockets so that they could be spatially placed around the listener back in the studio. Most classical recording is done with minimal microphones situated around the conductor's podium or slightly behind it to capture the sound properly balanced as the conductor intends. Even large orchestral movie scores eschew the multitrack recording methodologies so common in the commercial music world. We knew we were doing something against the traditional standard, but then again, there are not many models around in the DVD-Audio format.

Around noon, the equipment was up and running, and things were ready for the obligatory mic check and test-record. The complexity of the digital equipment and the extreme interdependencies of the system made everyone a little more nervous than they would have been otherwise. Everything was set, the recorder was placed into record mode, the metersÝ were bouncing up and down, and signal was flowing through to the monitor console and speakers. After a couple of minutes, the machines were stopped, rewound (a holdover from tape days-actually the R-1 is random access), and the material played back. The system was working. The orchestra was already beginning to assemble on stage, tuning their instruments and warming up. Everything was set.

At precisely one o'clock in the afternoon, the union guy walked off the stage and we were set to go. The orchestra was informed about the project and asked to turn off pagers and phones, settle in, and get ready for an unusual session. This was to be one of the first high-resolution recordings, and was certainly the first to utilize 24 mics and multitrack recording models. Beethoven was up first. As is generally the case in classical recordings, the piece is broken down into the individual movements and recorded piecemeal. The first movement of Symphony No. 6 is an Allegro (fairly fast) section of about 10 minutes in duration. The slates were slated, the machines engaged, and we began. The conductor and members of the orchestra played as a first-rate ensemble stopped at the final bar and waited for instructions from the booth. We asked them to hang on for a moment while we did a quick playback to verify the recording.

At that moment, the computer screen displayed one of those error messages that chill you to the bone. I don't recall the exact words but it was something like, "Communication Buss Error 13-Linkage Fault/Recheck Connections and Try Again.Ý OK?" Not OK. The room was suddenly awash in questions and attempts to rectify the problem. Calls went out to the Euphonix tech support line and equipment was reset for another try. Maestro Macal and the orchestra were informed that we needed the movement one more time and once again we pressed the record button. After another ten minutes, the same thing happened again. With $15,000-per-hour musicians on the stage, the situation was rapidly spiraling from bad to worse. Everyone was calm and focused, but it was apparent that the magic of high-resolution recording was more difficult than anyone had imagined it was going to be. The entire first hour went by without success and we had no solution in sight. The DAT recording was starting to look like it would have to serve as the master unless a miracle happened.

As it turned out, we failed to make a recording the entire first day due to a fried component in the R-1 system. With the substantial help of Euphonix technical personnel and a lot of scrambling overnight, we secured replacement equipment and got the system running during the early hours of Sunday morning. This was to be our redemption day. With the entire orchestra in black formal dress and with video cameras rolling, the conductor was asked to run through the Beethoven for the videotaping. We would follow that recording with the Respighi. It meant recording both works, almost 48 minutes worth of music, during a single three-hour session. To accomplish the feat, the orchestra would have to be flawless most of the time. We would be able to pick up only the most glaring performance mistakes after the movements were played continuously. It is a testament to the quality of the New Jersey Symphony Orchestra that they were able to rise to the occasion and deliver two stunning performances. The Respighi was recorded essentially in one take and has the magic and excitement that is often lost in the editing room. They pulled through even though they never knew the importance of Sunday's sessions.

Postproduction:

The DVD-Audio/Video Disc

The mixing sessions were planned for early December in our Los Angeles studio using our Euphonix System 5 high-resolution console. We've set our room up with a matched set of B&W speakers driven by a new Bryston 9B amplifier. Tomlinson Holman of TMH Labs supplied the "Profunder" subwoofer and also did a complete room tune and calibration. There are some fundamental philosophical issues to contend with during the creation of a 5.1-channel surround mix of any type of music, but the issues are even more sensitive when you're dealing with "classical" repertoire. It seems that hundreds of years of traditional performance practice have created some rather stringent expectations on the part of "classical" listeners, especially those in the audiophile community. They want and expect to hear a "sonic document" of an actual performance. That means capturing the acoustics of the space and hearing the musicians play from a proscenium stage. It's what we refer to as the "audience" mix perspective. Most audiophiles, some of the most likely individuals to acquire DVD-Audio/ Video hardware and associated systems, enjoy tweaking their expensive stereo (two-channel only!) playback equipment to get the last ounce of fidelity out of it. It seems they aren't completely convinced of the necessity of 5.1 channels of amplification and additional speakers. And more importantly, if one accepts and adopts multichannel playback, how should it be determined what is placed in those surround speakers?

For DVD-Video mixes, where the attention of the viewer needs to remain focused on the character displayed on the screen, any audio that comes from the rear is usually sound effects (like the "fly-over" sound of the arrival of an off-camera helicopter). In fact, the rear speakers in a "home theater" are usually "dipoles" (having two outlets for sound on the sides of the units) instead of the matched set of direct reflectors needed for music. Film mixers use the surround speakers for ambiance and effects, but would never consider placing dialogue in the back of the space. The center channel contains the majority of the dialogue, whereas the center channel in music mixes is the subject of a rather heated debate.

DVD-Audio uses a slightly different arrangement of the same 5.1 speakers found in "home theater" systems. The five main speakers are placed in a perfect circle around the listener at specific angles. The front-left and -right speakers form an equilateral triangle with the center (30 degrees from the center) with the left and right surrounds between 110-120 degrees off the centerline. All five speakers should be at the same height and pointed directly at the center of the circle.

The philosophical question regarding mixing music in surround focuses on the degree to which an engineer/producer can or should place individual instruments in the rear channels as opposed to simply placing room ambiance behind the listener. For audiophiles and most music fans, who are accustomed to hearing everything in stereo, being immersed in the midst of a music presentation might seem odd and uncomfortable. After all, if you accept that the ultimate goal of recording is to capture an acoustic event, then placing instruments in the left and right surrounds would destroy this "documentary" ideal. As a composer, musician, and engineer bent on doing more than merely recreating the concert experience in the home, I am completely comfortable with sitting on stage with the musicians. The thought of having layers of rich sound coming at me from all directions is a luxury that this new DVD technology offers us for the first time (with the exception of the failed period of quadraphonic recordings). Maybe the best part of the DVD format is its ability to provide both sorts of mixes on a single disc. AIX Records is making both "stage" and "audience" perspective mixes available on its releases. We really want to know what people think. It might be that musicians prefer being in the middle of the music and traditional audience members like it better in the best seat in the house.

A large part of my initial enthusiasm for this project came about because I dreamed of having multiple microphones at my disposal during the mixing process that I could place anywhere I want. The mix that is on the DVD-Audio side of the disc puts the listener right in the middle of the ensemble, a position that the music director would find very comfortable. I was extremely gratified that Karl, Karen, the Maestro, and the New Jersey Symphony Orchestra encouraged a "radical" approach to their first DVD project. I'm not sure the management of most orchestras of the world would endorse my philosophy; we'll see what the general public thinks. So far, the vast majority of people who have heard the mixes prefer the "immersive" one (a free sampler disc is available on the AIX Web site).

In December, Karl and I worked on the balance and placement of the instruments, making sure that everything was clear and appropriately situated in the mix. At my urging, we placed the listener directly in front of the woodwind section. The first and second violins come from the left surround and left-front speakers respectively. The violas, celli, and basses occupy the opposite side of the mix, with the woodwinds and brass straight ahead. Things are generally in their correct positions relative to the conductor-the breadth of the group is simply stretched beyond the normal limits. At a recent playback for the trustees of the NJSO and the New Jersey area media, the results were met with uniform approval, although I played the tracks back for a prominent commercial musician who thought that I hadn't gone far enough. He wanted to hear the clarinet flying from corner to corner and the strings constantly shifting location. Sometime, maybe, but for now the NJSO has taken the first step.

Implications for Future DVD-Audio Projects

The sampler disc that AIX Records has compiled demonstrates what can be accomplished with high-resolution recording and aggressive 5.1-channel surround mixing. From the center of a Brahms piano quintet to the experience of hearing music from inside a nine-foot grand piano (15 microphones on a single instrument!), the flexibility and creative control offered by the DVD format is incredible. The sound quality is beyond anything you've ever heard before. We're recording 3-5 new projects per month for release in this format, using musicians that perform using acoustic instruments as a live ensemble. There's nothing quite like hearing the sound that comes off the front of a great guitar rather than a paper speaker cone found inside an amplifier. I believe that listeners will hear the difference only when they are presented with compelling materials that can demonstrate the difference: making new recordings with live musicians at 96kHz/24-bit and mixing from a "stage" perspective worth the pain and suffering we endure when producing at the bleeding edge.


Copyright 2000-2001 Online, Inc.
213 Danbury Road, Wilton, Connecticut 06897-4007
203/761-1466, 800/248-8466
Fax 203/761-1444
info@onlineinc.com