| Creating Surround Sound Mixes |
| EQ, October 1997, Volume 8, Issue 10, by Robert Margouleff and Brant Biles with Steve La Cerra Enough theorizing--here's how to get into surround sound mixing and, more important, why Since its introduction in the 1960s, stereo (2-channel sound) has dominated our industry. During the early 1970s, the record industry marketed recordings in quadraphonic (4-channel sound), but quad never really caught on for several reasons. One: groove geometry couldn't support four discrete channels of sound on a vinyl record. Two: the music business didn't involve the artist enough. Three: there were too many encoding schemes and no one could agree on a standard. While the various companies involved argued, the whole thing fell by the wayside. Well, it's more than 20 years later and audio technology has grown by leaps and bounds. Motion pictures are routinely mixed in surround sound and encoded into film soundtracks using formats such as Dolby Pro Logic Surround, Dolby Digital, DTS (Digital Theater Systems), and SDDS (Sony Digital Dynamic Sound). Anytime that you start mentioning something with three initials-whether it's DTS, AC-3, THX (which is not a codec), DVD, or the like -people tend to get a glazed look in their eyes due to the information overload caused by all this new technology. Since we are professionals working in the audio industry, we think it's a very good idea to define the parameters of what's going on in surround, and why it's important. Why Surround? The first thing you're probably asking yourself is, "Why would I want to mix a record in surround sound?" And given the checkered history of surround audio, it's a legitimate question. Here's a story to help you understand why: Last summer, we were mixing the Pavarotti and Friends for War Child concert in 5.1 surround, for release by HDS Records in DTS (we'll get to what "5.1" means later). Originally produced by Phil Ramone and recorded by John Pellowe, it had been recorded live with a full orchestra. We were given only six stereo pairs of premixed stems, thus making our job kind of a cross between mixing and mastering. There were a few problems initially, but when we dug into it and got it sounding good, a few people came in for a listen. Upon doing so, they would slip out of the control room in tears because they were so struck with emotion. It happened to us as well, and we had to actually stop, step back for a minute and then continue mixing. Surround greatly intensifies the emotional energy of the music--that is the bottom line. Over the course of the last few years, people have left music behind for other electronic stimuli. Our industry needs to bring them back through the true power of music. Millions of people already have surround systems installed in their homes to play back movies on video and laserdisc, so why not share the same platform for music reproduction. It seems we're at the beginning of a convergence of new technologies, where music, film/video, and computing will all come from (and be accessed by) the same system. Surround allows us to communicate in a whole new way (with a much larger canvas) and with many new colors of paint for our creative palette. First Thing's First For starters, let's make clear what we're talking about--discrete multi-channel audio. We're not talking about psychoacoustic surround systems, which all use two speakers to create the illusion of sound coming from all around you. This technology is called transaural audio, an example of which is QSound. Transaural audio is now being developed for the delivery of surround for multimedia computers. Sometimes referred to as "Fifty-Yard-Line" surround, it's good for only one person, and you have to be positioned on the center line of the two speakers to be able to hear the phantom surround images. If you're at exactly the right angle and the listening geometry is perfect, then one person will get 3D audio. What we are concerned with is digital, discrete, 5.1 surround sound. Some people in the industry refer to an audio-only surround mix as a "six-channel mix" to distinguish it from multichannel audio-for-video. For this article, we'll refer to it as "5.1." In and of itself, 5.1 is not specific surround format owing its allegiance to any particular company or codec. It is a listening platform and hardware concept for a surround monitoring system. If you want to listen to music in 5.1, you need six discrete audio channels in your listening room. These channels are Left Front, Center, Right Front, Left Surround, Right Surround, and Subwoofer--that's the ".1.". Since the Subwoofer, or LFE (Low Frequency Effect in the film industry / Low Frequency Enhancement in the music industry), channel isn't full frequency, the powers that be didn't feel it was worthy of its own number when naming this format "5.1." It's important, though, to note that 5.1 audio-only mixes can be played using the same amps and loudspeakers used for home theater surround. As audio professionals, it's also important to remember that the philosophies of mixing audio-only surround and surround for video, are completely different. Video requires a "frontloaded" mix to support a picture. In audio-only mixing, the constraints of film/video mixing can be broken. By regarding all channels as equals, the possibilities become endless. In the motionpicture world, the two most popular 5.1 formats are Dolby Digital and DTS, with Sony's SDDS trailing as a distant third. All three store data digitally and are delivered by varying methodologies. (It's worth mentioning that the recently approved HDTV standard also includes 5.1 audio.) Six Mix Within the 5.1 platform are several formats that use encoders to form a new bitstream from the six channels of digital information comprising the mix. This encoded bitstream is then burned onto a CD, laserdisc, DVD, or other storage medium. Upon playback, the listener needs a complementary decoder to translate data from within that new bitstream back into the six channels of digital audio. This process of encode-decode is called a codec. Now you're probably thinking, "Here we go again with different decoders and no standard," but the truth is that the 5.1 setup and hardware are the same for these systems except for the codec. Fortunately for the consumer market, the same DSP chip from one of several manufacturers can decode both the DTS and Dolby Digital formats. Soon, the discs themselves will have an ID flag to let the decoder know which codec was used. Once this is known, the decoder will automatically switch modes and process the incoming signal appropriately. This will overcome the main reason that quad failed, because it will be possible to have a single decoder unit that recognizes all incoming bit streams. As previously mentioned, there are two major 5.1 surround formats already in the marketplace. Both call for five full-range speakers (preferably all the same) plus a subwoofer. Let's take a brief look at them: Dolby Digital Previously known as AC-3, Dolby Digital is well-entrenched in the motion picture industry with roughly 10,700 Dolby Digital-equipped screens worldwide. For home use, Dolby has developed a data compression ratio of (typically) 11:1, originally developed to fit 5.1 audio onto the same disc with the video data for movies on DVD. The data rates for 5.1-channel Dolby Digital are designated as 320 kilobits per second (kbps) for film, 384 kbps for laserdisc, and 384 or 448 kbps for DVD. Dolby's current encoder can accept incoming data at 32-, 44.1-, or 48- kHz sample rates, with word lengths of 16, 18, or 20 bits (this can be extended to 24-bit word length in the future). As of this writing, Delos has released the first and only audio-only DVD in the Dolby Digital format (DVD Spectacular, featuring a complete performance of Tchaikovsky's "1812 Overture."). In addition to this audio-only disc, there are several music video DVDs available with Dolby Digital soundtracks, with 60 music video releases expected by year end. DTS In 1993, Jurassic Park became the first motion picture to be released in DTS. Since that time, DTS has become firmly established in the motion-picture industry. DTS is now currently installed in over 12,000 theaters worldwide. For home video, the DTS codec is scalable from 256 kbps to 1536 kbps, and DTS is focusing on 1411 kbps as the optimum for transparent sound quality. Even though DTS began exclusively in the movie business, its emphasis on optimum sound quality stems from a corporate commitment to bring 5.1 to the forefront of the music industry. Over the past year, DTS has been responsible for releasing over 40 titles on CD in 5.1 surround, including Boyz II Men II and the Eagles' Hell Freezes Over. In fact, DTS is the only company currently releasing product in the 5.1 audio-only market on labels such as Telarc, High Definition Sound, and Digital Music Products. A DTS compact disc carries six channels of digital audio (5.1) in 20-bit words at a 44.1 kHz sample rate. To fit this information onto a CD, DTS compresses the data at a ratio of about 3:1. As of this writing, there are around 10,000 DTS decoders in homes across the U.S., and that number is expected to grow to 50,000 by the end of this year. DTS was the first company to jump into the 5.1 CD and DVD-audio-only pool. Right now, only DTS-encoded 5.1 mixes are being released on standard CD. Although an additional decoder is needed, you can play a DTS CD on a standard CD or laserdisc combination player (and DVD player as well), provided that the player has a digital output for connecting to the decoder. Jumping In Artistically speaking, mixing in surround opens up a whole new world of creativity to an artist. Surround interacts with listeners by putting them in the performance. Instead of being observers, they are participants. For example, if you listen to a CD, the music sits in front of you as an object. One can then choose to regard it or not--sort of like looking at a painting hung on the wall. In surround though, the listener moves from that objective experience into a subjective experience, where the music and listener occupy the same space--you are now an element in the painting and not just an outside observer. This is a very different world, which will determine how people write, perform, and record music in the future. It's not just, "Let's take a stereo recording and make it into surround." We're talking about many new artistic production decisions like, "Are we going to put the congas in the rear and the rhythm guitar in the left front? Should the lead vocal move to the center channel while the background vocals move in rhythm around the room?" In traditional mixing, we use all kinds of devices, such as reverberation, echo, doubling, spatial processing, compression, and pitch-shifting--every trick in the book--to create the illusion of space and depth. In conventional stereo recordings, a person might double a part and pan one left and one right. In surround, artists might think along the lines of recording four (or five) different performances or orchestrations of the same part, and have one come from each loudspeaker. It's the same part, but you'd get this overwhelming immersion in the sound because it's being delivered from multiple locations instead of just two. A lot of recorded performances (especially in pop music) are not documentary reports of reality. For example, most symphony orchestra recordings are made in the documentary style. You put a dummy microphone head in an auditorium, the orchestra plays, and that's the whole recording (which is totally valid). This is not to say that orchestral recordings in surround aren't phenomenal, because they are! A pop record (on the other hand) is done in a more serial way, with lots of overdubs. Many pop records will never be (and never were intended to be) performed live. It's not a real-time event. So what happens is that you move into this realm where the medium is part of the performance. When you suddenly have six channels of audio to put the performance into (instead of two), the listening experience is incredible. As producers and engineers, we can now think in terms of six busses. Earlier this past year, we remixed Boyz II Men II for DTS. This was a major project that had been recorded on 24-track analog. Some songs had multiple slave reels for ganging up background vocal and percussion performances into submixed stereo pairs for the final stereo mix. One particular song, "Water Runs Dry" (originally produced by Babyface), had a total of 18 individual tracks of background vocals. Instead of using the available stereo submix, we set them up so that you sit in the center of the room and get this massive blend of the quartet. If you want to hear a little more of Nate you can move over to the left rear speaker. If you want to hear a little more of Sean, you can walk over to the right front. In this way, the mix becomes interactive and you actually change the mix depending upon your position in the room. The mix takes place in the room-not in the loudspeakers. Also worth mentioning is that there seems to be a lack of hearing fatigue when mixing in surround. We're all familiar with the fact that when mixing in stereo, the ears get fatigued over the course of the day. But to our surprise, with surround mixing, fatigue seems greatly reduced. We believe this is due to the nature of human hearing. When a sound is generated from two locations, there are phase peaks and valleys at different frequencies, at different places in a room. This phenomenon is more apparent if the two sources are located front and rear as opposed to left and right. So, as you move while working (nobody keeps perfectly still with their teeth on a bite bar), your ears get more of a variance in SPL at different frequencies. Also, with music in 5.1, the ratio of average SPL to perceived loudness is lower than that of stereo. The Politics of Surround In the past, surround has not been a very democratic medium because it has been expensive for both pros and consumers. But as audio professionals, we have to recognize two important things: First: 5.1 lives in what is now known as the home theater--which is the largest selling segment of the hi-fi business. In fact, it's the only area of home entertainment that isn't depressed. Second: Technological advances in recording equipment are making surround attainable in any home studio (there's the democracy). Tools of the Trade We have been mixing and mastering projects digitally at The Enterprise (Los Angeles, CA) using a Neve Capricorn that has an all-digital signal path and dynamic memory. Other producers and engineers have been using consoles such as those from Euphonix (digital control of an analog signal path with dynamic memory), SSL, and older Neve's. Though the older analog desks lack dynamic memory, it is still possible to do surround mixes on them. So where does this surround mix end up? Mostly on a TASCAM DA-88. In the case of mixing for the DTS coder, all six output busses must be digital (AES), 20bit, and at 44.1 kHz. In the case of an analog console, the analog busses must go to six A/D converters, which can put out a 20-bit word length. Once in the digital domain, all six channels get routed to a Prism MR-2024T. The Prism unit takes these six 20-bit channels and converts them to eight channels of 16-bit TDIF (TASCAM Digital InterFace) information. The MR-2024T achieves this by storing the upper 16 hits of all six channels on the first six tracks of the DA-88, respectively. The remaining 4 bits per channel are then combined and stored on tracks 7 and 8. On playback, the Prism recombines the first six tracks of 16-bit information with its respective bottom 4 bits and spits back out the original 20-bit information sent to it. Thus, we store a 20-bit, 5.1 mix on a 16-bit DA-88 tape. Although we use a Neve Capricorn, there is great news for you, the thrifty audio equipment buyer: the Yamaha 03D and revised 02R can output six digital busses. Both have dynamic memory and panning--just like the Capricorn and Euphonix--but cost under 12 grand (under five for the 03D). That, combined with the DA-88, means that a project studio can mix in discrete sixchannel surround with automated faders, panning, equalization, and dynamics (though the mix would still have to be mastered and encoded). As a matter of fact, you can set up a standard 8-bus analog mixer for 5.1. We recently mastered The Eagles' Hell Freezes Over, which was mixed by Elliot Scheiner at Capitol on a Neve VR with an analog film bus (see the accompanying discussion with Elliot on page 80). The point is, with a little forethought and creativity, it can be done. To maximize panning flexibility, you might want a joystick panner, but the Capricorn doesn't have one, and we've used that console to do many surround mixes (the tips of our fingers have calluses on them from twisting the knobs all over the place). If not a joystick, then certainly a console that can memorize panning moves in automation. If you were using an analog console and needed to move a sound from the left front to the rear right, you might try returning the track to two faders. Then assign each fader to its intended output bus and crossfade between them. You do it once and, like magic, you've got automated surround panning. Simplifying matters further is the fact that there are a lot of high-quality, self-powered loudspeaker designs available, which decrease the wiring headaches. We've been using Genelec 1032A's with a pair of M&K's M5000 subwoofers. In our monitoring situations, we try to set all of the speakers equidistant (90 to 100 inches) from the listening position. This places the left-front, center, and right-front speakers in an arc. If that cannot be done, then a compensation of level for the center channel is in order. We want to make sure that there are symmetrical and equal front and rear soundstages. One of the nicest tools we have found for surround mixing is the EMT 250 reverb, which--originally designed in the 1970's for quad--is single input, four output. We've also used Lexicon 480L's and two PCM80's with one machine returning to the front and the other returning to the rear. Generally you'll find that you use less effects in surround than in stereo, and the front-to-back assignment of sound is more dramatic than left-to-right panning. We have taken to mixing into either Sonic Solutions or Pro Tools, eliminating the need for tape, and this obviously makes assembling an album, level-matching, and mastering much easier. These masters really need to be prepared in some sort of standardized room, though. In fact, A&M and Abbey Road studios are currently building control rooms dedicated and equipped for surround mixing. It's all at the "bloody cutting edge" right now, and we don't have all the answers yet, but we're working hard to get as many as we can. We're in the lab right now and quite honestly, we haven't had this much fun in years! The surround bell has rung and it can't be unrung! The Near Future While we're writing this, the RIAA is working with the DVD Alliance to evaluate a proposed software/hardware standard for a high-density audio-only DVD. Among other features, it would store a discrete six-channel mix (bit depth and sample rate to be determined) along with a high-resolution stereo mix. A DVD can hold all of this information because of its high storage capacity. The RIAA is also recommending that this disc be dual-layer comprised of the high-density layer plus a standard-density "Red Book" layer to provide your standard 16-bit PCM stereo mix (which would ensure backward compatibility). Under consideration for inclusion in the software/hardware standard are codecs from Sony/Philips, Toshiba/Warner, DTS, and Dolby.
|