Unless you've been asleep under a rock for the past ten years, you've at least heard about surround sound. During the past several years, it has had a serious impact on movie soundtracks, and it could replace stereo as the de facto music-listening standard. But surround is a new format, and lots of myths, misinformation, and misconceptions about it are floating around. So if you're not sure what terms like AC-3, DTS, DVD-A, and 5.1 mean, you're in good company.
You can provide customers with a multichannel surround-sound experience in two different ways. In one method, a number of audio processors synthesize a multichannel signal from 2-channel stereo sources, but no magic surround processor will transform an existing stereo mix into a proper surround mix. Most home-theater systems have some sort of "ambience" setting for this purpose, but it's the worst-sounding effect you can imagine. A few of the more advanced processors include "matrix" modes, some of which work pretty well, but these modes will never equal a true 5.1-surround mix.
A much better option is to create original multichannel mixes from scratch. Mixing in 5.1 surround requires some extra equipment, which means learning some new techniques. But it's worth the effort; mixing music in 5.1 surround is the most exciting thing I've been involved in. We are witnessing a milestone in audio history. Millions of home-theater systems in the United States can play back surround music mixes and movie soundtracks. It's time for the recording industry to put out surround product - and it's not as difficult as you might think.
EARLY BIRDS AND WORMSI'll start with a little historical perspective. You're probably familiar with monaural (aka mono) sound: you deal with only one channel of music information. The information can be as simple as a cheap clock radio with a tiny 1-inch plastic speaker or as complex as a concert sound system with dozens of triamped cabinets in large speaker stacks.
Of course, stereo is our standard listening format today. In stereo, you work with two distinct information channels, and you can position various instruments between two speakers typically arranged in front of the listener in an approximately 60-degree spread. This technique allows for some pretty cool psychoacoustic tricks which make the sound seem to emanate from a position between the speakers, even though no sound source exists in the center. This is called virtual, or phantom, center.
You can produce a stereo sound field, or soundstage, in a variety of ways. The easiest is to use a pair of microphones on a live group and direct these two channels to the speakers. But most music today is recorded in multitrack format. The channels are panned left, right, or in between at mixdown, creating an artificial soundstage.
Of course, stereo has been king for more than 30 years. The industry made a brief excursion into quadraphonic (4-channel) sound back in the 1970s, but squeezing four channels of music into a single record groove was beyond the technology's capabilities. A small group of home experimenters actually set up quad systems, and a few music recordings were released on LP and 4-track reel-to-reel tape. However, quad died an ignoble death, and many people ended up with expensive gear and nothing notable to play on it. The industry never really recovered from this brief affair with quad. To this day, you can make many record executives jump by mentioning the "Q word."
SURROUND REBORNThe movie industry revived the idea of surround sound. Aside from the incredibly ingenious multichannel soundtrack of Disney's Fantasia, the first real breakthrough was Dolby Surround, which offered left, center, and right front channels as well as a monaural, limited-bandwidth rear channel for special effects, such as the sound of Superman flying overhead. This mono rear channel was normally reproduced with two speakers to the sides of the listening area.
However, squeezing four channels of sound information onto the two audio channels of 35 mm film proved an imperfect solution. Playback with different Dolby Surround decoders could vary radically. More advanced decoders, such as Dolby Pro Logic, were designed, but they all suffered from the dreaded "phase-steering" problems, in which a level change in one channel could affect the mix in the other speakers.
Enter the digital age. The development of the compact disc in the early 1980s provided a way to deliver large amounts of digital data. Bits is bits, so the same bits could represent a graphic picture, your accounting information, or more audio. Tomlinson Holman (the "TH" in THX) was one of the leaders in surround sound in those days, and from his experiments with movie soundtracks, the term 5.1 (pronounced "five point one") was born.
The 5.1 format defines six discrete channels: five full-bandwidth channels (20 Hz to 20 kHz), and one low frequency effects (LFE) channel (the "point one" in 5.1) with a frequency response rated from 5 to 125 Hz. The LFE channel requires a specialized speaker, called a subwoofer, which reproduces only low frequencies. Few, if any, subwoofers can reproduce 5 Hz; most can reach down to 30 or 35 Hz before they roll off, and a few of the more expensive ones can go to 20 Hz. The channels are designated left, right, center, left surround, right surround, and LFE.
The bright people at Dolby Laboratories figured out how to digitally compress these six channels of information into a form that would take up less bandwidth than two stereo PCM tracks, and the Dolby Digital codec (coder-decoder) was born. Also known as Dolby AC-3, this codec is used on many current DVD movie soundtracks, and it is part of the High-Definition Television (HDTV) standard.
The situation remained static for a few years, but with the release of the movie Jurassic Park, a competing codec was introduced by Digital Theater Systems (DTS). The DTS codec (DTS is the name of both the format and the developer) uses less data compression and requires more bandwidth and data-storage space than Dolby Digital, so some DTS movies don't quite fit on a single DVD. However, the tracks have the potential to sound more like the discrete PCM tracks from which they were derived than is possible with Dolby Digital.
DTS pioneered a way to use the same format on a Red Book CD, but with compressed DTS data in place of PCM stereo audio. DTS also formed a record label to produce remixed 5.1-surround versions of stereo releases. Many of these remixes were done by the engineers who handled the original mixes. Currently, you can buy more than a hundred 5.1 DTS titles, including work from such artists as Steely Dan, Lyle Lovett, and the Eagles.
To play these CDs, you need a CD or DVD player with a digital audio output that can send the DTS bitstream to a DTS decoder, which extracts the six channels of information and converts them to analog. (Early DVD players have a digital output, but they don't recognize the DTS bitstream. Most consumer CD players don't have a digital output, and those that do might not recognize the DTS bitstream.) You also need six channels of amplification and speakers.
BACK TO BASICSI'll start at the beginning of the mixing chain and go through it step-by-step. You need some special items to mix in 5.1 surround, but most studios already possess 90 percent of the needed equipment. Once you add a few select pieces, you could be mixing surround music in your own studio.
The first thing you need is a multitrack master of the tune you want to mix (see Fig. 1). The multitrack format is not an issue; it can be as simple as an 8-track analog tape deck or as complex as a pair of 48-track digital decks. I've done some really cool 5.1-surround mixes using 16- and 24-track ADAT systems. The source tracks can be in any digital or analog format, including a computer workstation. Of course, you want tracks with excellent production values.
You need to route the recorded tracks into a mixing console that lets you pan them between five output channels. (There should also be a sixth output for the LFE channel, but you don't pan anything from the main tracks to this output.) If you have a Yamaha O2R or O1V, Panasonic DA7, Tascam TDM4000, or Mackie Digital 8-Bus, you're already in business. Each of these digital consoles lets you patch the outputs from the surround matrix (outputs) to an 8-channel mixing deck (more on this shortly).
If you don't have a console with built-in surround panning, it's relatively simple to patch the equivalent using sub-groups or aux sends (see the sidebar "Mixing Surround Without a Matrix"). But for ease of mixing, nothing beats a screen with a picture of the room and a cursor that shows where the sound ends up. Some consoles, such as the DA-7, provide a pair of controls on the work surface to pan left/right and front/rear, while others, such as the Mackie D8B, use a trackball or mouse.
Patch the console's outputs to an 8-track deck, where your final surround tracks will reside (see Fig. 1). The Tascam DA-88 has become the standard multitrack deck for surround due to its popularity in the film industry, but any common 8-track format will work, including a computer workstation. You don't need a Dolby Digital or DTS encoder to mix surround tracks; encoding is the last part of the process. Whatever you record your mix on, carefully note the track assignments. Unlike stereo, surround gives you many different track-assignment methods to choose from. The table "Surround Track Assignments" shows a list of the most common track-assignment systems, or modes.
Often you won't be able to choose which channels end up on what tracks. For instance, both the Panasonic DA7 and the Mackie D8B mixers are set up in mode 4, whereas many large-format consoles and some mixing programs (such as Minnetonka's MX 51 for Windows) are designed to be used in Mode 1. Try to pick one output format and note it on the label of every 5.1 tape you make. Eventually, someone will have to figure out your track assignments to encode them on a disc, and you don't want your sloppy work habits to jeopardize a project.
SPEAKING OF SPEAKERSYou'll need to upgrade your monitoring system to include five speakers and a subwoofer so you can hear what you're doing in 5.1 surround. The speakers are arranged in a circle around the listener's head (see Fig. 2), and the channels are labeled L (left), C (center), R (right), Ls (left surround), Rs (right surround), and LFE. Feeding these speakers requires six channels of amplification. Keep the channel definitions in mind as I discuss patching options.
The simplest and perhaps best setup for music mixing is five matched near-field reference monitors used with a subwoofer. I like the JBL LSR-28P and M&K 150 speakers, but I've heard some great mixes on the little Alesis Monitor Ones and Yamaha NS-10s.
Try to match the main speakers as closely as possible, because the relationships between the center, left, and right levels are critical, and the relationships between the front and surround levels affect the final mix much more than you might imagine. You could mix and match speakers - many excellent surround mixes have been done with mismatched center and rear-surround speakers - but I recommend you use a matched set to avoid putting yourself at a disadvantage.
The physical speaker layout is pretty simple (see Fig. 2). Just put a mic stand at the mix position and measure the distance from that position to the location of the center-channel speaker. Then cut a piece of string to the same length and tie one end to the mic stand. Mark the center-speaker position and go 30 degrees to the left and right for the front L/R speaker positions. Next, go 110 degrees to the left and right of the center for the Ls and Rs (surround) channels. This setup is the AES standard for monitor placement.
There is one exception to the "five matched speakers" rule. When mixing movie soundtracks, you don't want direct-radiating, point-source surround speakers, because most movie surround effects need to be diffuse. As a result, the surround speakers should be dipole designs, which have drivers that fire forward and backward, creating a diffuse sound field. In this case, the surround speakers should be placed directly to the sides of the mixing position, 90 degrees from the center-speaker location. If you're mixing multichannel music, use a pair of direct-radiating speakers matched to the front three speakers and placed at the 110-degree positions.
YOU NEED MORE POWERYou'll probably need to upgrade your power amps, too. Many big studios can afford separate amplifiers and a bass-management controller to run them, but personal-studio owners should instead consider buying a large home-theater receiver. Priced between $500 and $1,000, such receivers provide more than 100 watts per channel and a single level control that adjusts all channels simultaneously.
All the main speakers should get the same amount of power, but the subwoofer probably needs as much juice as all the other speakers combined. Giving the subwoofer two to three times the power of one of the main speakers will probably suffice. It's important to understand that many subwoofers include their own internal power amps matched to their particular drivers. As a result, virtually all modern home-theater receivers include five channels of amplification and a line-level output for the subwoofer.
Make sure the receiver sports Dolby Digital and DTS decoding as well as discrete analog inputs that bypass the decoders. Meant to accommodate future decoders, these inputs are ideal for monitoring the six discrete channels of the mix. With such a receiver, you can compare your mix with commercial mixes from DVDs or DTS CDs. (To hear commercial discs, you also need a DVD or CD player with a digital audio output that passes both DTS and Dolby Digital bitstreams to the receiver.) If you're doing your own DTS or Dolby Digital encoding, the receiver provides the only way to listen to your final mix as the consumer will hear it. These receivers also offer bass management (more on that later) and speaker-calibration options.
I'LL LEVEL WITH YOUProperly setting the relative volume level of each speaker is extremely important. It's very easy to tell when the left and right levels are wrong in a stereo mix; you can hear the sound leaning one way or the other. Hearing the balance with a surround system is not so simple. You'll need to purchase an SPL (Sound Pressure Level) meter to do this properly, and you'd be amazed at how many people use $50 Radio Shack meters for the job.
I could write a book just about speaker calibration, but for now, here's the quick and dirty. Grab some limited-band (100 Hz to 2 kHz) pink noise from the mixing-console noise generator, a home-theater receiver, or a test CD and patch it to an input channel on the console. Make sure to set the mixer's output gains to unity. Set the console's input strip so that the output level going to the mixing deck is at -20 dB (below 0) on the deck's meter. This is the standard reference level for surround mixes.
Feed the pink noise through one speaker at a time and point the SPL meter in the speaker's direction while holding the meter in the mixing location (see Fig. 3). Trim the gain of the corresponding amplifier channel so the meter reads 85 dB SPL. Repeat this process for each full-bandwidth speaker, one at a time, until you have the same output for each. (Most home-theater receivers include individual level controls for each channel, but they are often buried in a menu system that might require a video monitor for viewing and navigating.)
Next, run low-frequency pink noise (25 to 80 Hz) to the LFE channel. In theory, the gain of the LFE channel should be set 10 dB higher than that of the main channels (95 dB SPL), as read by a real-time analyzer (RTA). But a Radio Shack meter has a lot of low-frequency rolloff, and you're feeding it less than two octaves of audio information in this case. These factors cause the meter to read lower than the true output level. As a result, when the LFE level is correct, the Radio Shack meter will show approximately 90 dB SPL - 4 to 6 dB higher than the level for the five full-bandwidth speakers.
For diffuse surround speakers used in cinema mixes, the rear surround levels are set to 82 dB SPL (-3 dB relative to the other full-bandwidth speakers). For really small mixing rooms where you can literally reach out and touch the speakers, Dolby recommends setting the surround speakers 2 dB down; that is, at 83 dB SPL. It can be a bit confusing. For most music mixing, having all five speakers set at the same level is close enough.
If you don't get these levels correct, all the mixes you do will have incorrect surround- and center-channel levels, or the LFE level will be out of control. Those problems will force listeners to jump up and adjust the levels on their home systems.
BOTTOM FEEDINGBass management is probably the least understood part of surround mixing. It's very important to comprehend how it works, lest you make mixes that sound great in your studio but prove unlistenable on a standard home-theater system. As noted earlier, in 5.1 surround, each of the main channels is rated at 20 Hz to 20 kHz, while the LFE channel is rated at 5 to 125 Hz.
Any of the main program channels can go as low as 20 Hz, but very few home-theater speakers can produce output that low. A clever circuit in the preamp/processor or receiver of a home-theater system removes any low-frequency information below a certain point (80 Hz is the THX standard) in the main channels and reroutes it to the subwoofer.
In 5.1 surround, the subwoofer does double duty: it handles all the bass below 80 Hz for all the main channels, as well as the "point one" LFE channel, which might be earthquakes or explosions.
Why do you need bass management (sometimes called redirection) in your studio? It's the law of inverse mixing. A speaker that's deficient in part of the audio spectrum forces you to overcompensate for the missing frequencies by adding them to the mix in disproportionate amounts. Try doing a mix with rolled-off tweeters and you'll see what I mean. Because you hear fewer highs than are really going to tape, you'll overcompensate with too much high-frequency level in the mix.
Let's apply this concept to bass management. Suppose you have five NS-10 speakers and a big subwoofer that reproduces only the LFE channel (explosions and such). Each speaker directly monitors a final output track. The NS-10 has a small woofer in a small cabinet, so it naturally rolls off anything below 60 Hz or so. If your source tracks have any sonic material with extra bass in the 20 to 40 Hz region, you'll never hear it on these monitors.
This is particularly troublesome if the tracks have some undesirable low-frequency information - maybe some vocal plosives or air-conditioner rumble you weren't aware of. The natural filtering action of the NS-10s might make you think all is well, but when the mix is played back in any home-theater system, the bass-management circuit will faithfully reroute this low-frequency garbage into the LFE subwoofer, where it will be available for all to hear. So if each of your main speakers can't produce down to 20 Hz and you don't have a bass-managed monitor system, you have a potential mixing disaster on your hands.
Many mixing engineers think bass management has something to do with filtering the signal before it goes to the final mix-tape tracks. That's simply incorrect. In fact, each of the L, R, C, Ls, and Rs channels should get the full 20 Hz to 20 kHz program signal. You don't want to reroute the lows in these channels to the LFE channel; that's what the bass-management filter in the playback system is for. Instead, you want a bass-management filter in your monitoring system that emulates the home-theater playback system (see Fig. 4). This filter is placed after the mixing tape deck and directly feeds the monitor amplifiers, as shown in Fig. 1. If you're using a home-theater receiver for monitoring, it has an integral bass-management circuit, which does the job just fine.
Another misconception is that you must match the 80 or 120 Hz bass-management cutoff points in home systems to properly monitor in 5.1 surround. But because this filter is for playback and monitoring in your studio only, you have to do just what is needed to extend the low-frequency capability of your own monitoring system. Just as we don't care about the crossover frequency of the midrange driver in a home speaker, we don't know and don't care about what the exact bass-management frequency is in a consumer system. We just know that somewhere around 100 Hz, all the bass energy will head to the big subwoofer cabinet.
My main monitors go down to 35 Hz, so I like to adjust my bass-management cutoff frequency down to 50 or even 40 Hz. This would limit the bass-localization effect (yes, you can localize 80 Hz bass, contrary to popular belief) and take some of the power load off the subwoofer, which attempts to reproduce the bass from the five main channels as well as the LFE information. That's why you should buy the largest subwoofer you can afford and fit in your studio. For excellent information on bass management, download Steve Harvey's paper Secrets of Doing Surround Sound on Your Existing Console, published by Martinsound (www. martinsound.com/ lb_rp.htm).
Alas, there's no inexpensive way to put a separate speaker controller with bass management into a small studio. Martinsound makes something called the MultiMax controller, which does all the calibration, downmixing, and level control you could want, along with rudimentary bass management. (Downmixing refers to creating a stereo mix from a 5.1-surround mix.) It's certainly a great controller, and I take one along on my surround seminars because it's easy to use under stress. But its price of $3,000 is beyond many small-studio budgets.
Studio Technologies makes the Model 68/69 StudioComm ($1,599), which can be combined with the Model 65 Bass Management Controller ($899) to provide a $2,500 solution. These seem like expensive pieces of gear until you start shopping and find devices offering basic functions for up to $15,000. Gaaaaa! What's a personal-studio owner to do?
Using a home-theater receiver starts to look better all the time, because it handles bass management and speaker calibration, it decodes DTS and Dolby Digital discs, and it gives you a big knob that changes the levels of all channels at once. (Remember, you also want a receiver that includes six analog inputs that bypass the decoders so you can monitor the 6-channel mix directly.)
You could combine something like a Yamaha RV-1105 receiver ($700) with five small monitors and a self-powered subwoofer for an affordable speaker/amp/bass-management/calibration system.
DOWNMIXING BASICSThere's one other bugaboo to watch out for when doing surround. All 5.1 mixes might be downmixed to stereo at some point. For instance, if the consumer listens to a Dolby Digital DVD or DTS CD of one of your mixes and selects the stereo option on the receiver or surround processor, the six channels of information are mixed down to a pair of stereo channels and sent out the main left and right outputs.
In a downmix, the center-channel information is added into the left and right front channels equally, while the left and right surround channels are added into the left and right front channels. Some systems add the LFE channel into the stereo pair, and other systems throw away the LFE information.
This would be fine in a perfect world, but lots of potential phase conflicts crop up. For instance, if you put some sort of delay between the left-front and left-surround channels, you could end up with a huge, phasey sound when those channels combine into one. In such cases, a mix that sounds great in 5.1 surround can be unlistenable when downmixed to stereo.
If you think your carefully crafted 5.1 mix will never be heard in stereo, think again. Just as we need to check stereo mixes for mono compatibility, we also need to check 5.1 mixes for stereo compatibility. At the very least, a stereo version of your songs might be needed for radio play, and Lady Luck will probably choose the downmix of the one song that sounds as if it were mixed in your washing machine. Some very high-end surround mixes done by the most famous engineers on the planet sound horrible when auditioned in stereo. Guess what? Those engineers didn't understand the effects of downmixing.
What can you do about it? At the very least, you need a way to monitor the downmixing-cancellation effect. Many monitor controllers, such as the MultiMax, have a button that engages Downmix mode. Then you can easily hear phase problems.
Furthermore, always do a separate stereo mix of any 5.1-surround mix. This is particularly important for the new DVD-Audio (DVD-A) format, which has enough data-storage space to include both a 5.1-surround and stereo version of your mixes. That way, when consumers select stereo mode on their receiver or processor, they hear your stereo PCM version rather than a downmixed version of the 5.1 mix.
The proper place for this true-stereo version is on the same 8-track master tape that holds the 5.1 mix, on tracks 7 and 8 (see the table "Surround Track Assignments"). Although you have to do the mix in two passes - one for surround and the other for stereo - you'll be way ahead of the game when someone requests a stereo version of the mix.
ENCODE THYSELFThe final part of making a 5.1 mix is the encoding process. Fig. 5 shows where the encoder fits in the mixing chain in a hardware-based system.
Up until a few months ago, making a one-off DTS disc of a 5.1 mix required sending a six-track tape to DTS; the company would then encode the tracks and send you a DTS CD-ROM. Now you can get Minnetonka Audio's SurCode DTS ($499), a software DTS encoder for Windows that lets you encode your six discrete surround tracks as a DTS file with a WAV extension. This DTS file can then be burned onto a standard CD-R disc with any CD-burning application, such as Adaptec's Easy CD Creator. The file will play back through any home-theater system that includes a DTS decoder and a CD player with a digital output. (Don't try to play the CD from the analog outputs; all you'll hear is ugly noise.) Unfortunately, Minnetonka only develops software for Windows. I hope a Mac equivalent of Surcode DTS will be developed eventually.
Although a DTS disc looks like a Red Book CD-R (technically, it's an Orange Book disc), it probably won't play back in all DVD players. That's because the color of the dye and reflective layer in the CD-R medium itself might be incompatible with the wavelength of the laser in the DVD player's pickup. In addition, gold-colored CD-R media (such as the Kodak discs) seem to have a better chance of universal DVD playback than the dark-blue or green CD-R discs. Some of the newer Pioneer DVD players are advertised as having dual-laser pickups, and they seem to digest any color (and chemistry) CD-R I make. Interestingly, a CD-RW (rewritable) disc will play back on nearly every DVD player, even the old ones. So if you can't get a CD-R disc to play in your DVD player, burn a CD-RW disc and give it a spin. Of course, you can always play a DTS disc in any CD player with an S/PDIF output and a DTS-equipped receiver.
Those who want to use Dolby Digital encoding for surround mixes can burn a 44.1 kHz version of an AC-3 file and put it on a CD-R disc. You can get Dolby Digital encoders from Minnetonka Audio (SurCode Dolby Digital, $995) and Sonic Foundry (SoftEncode 5.1 Channel, $995). The encoding procedure is the same: load the six discrete audio tracks into the computer and toss them into the encoding software. Selecting the AC-3 WAV output creates an AC-3 file that's been padded out to fit in the exact same space as a stereo PCM file. Again, this file can be stored and burned on a CD-R or CD-RW for playback in many home-theater systems.
This method isn't recommended by Dolby Labs, and it's not 100 percent reliable, because some Dolby Digital decoders don't expect an AC-3 file to come from an S/PDIF bitstream with the Audio/Data flag bit set to Audio. Nevertheless, it works perfectly on many decoders and receivers. Until Minnetonka licensed the DTS encoder algorithms a few months back, this was the only way to hear a one-off version of your surround mixes on a home-theater system without dragging around a multitrack deck and mixer.
SURROUNDED BY INFOThere is much more to say about surround-mixing basics, but I've given you enough to get started. In future articles, I'll discuss subjects such as tracking for 5.1 mixdown and exotic surround microphones. But for now, it's time to start exploring the new world of multichannel music. It's a trip worth taking.
Even if your console has only stereo outputs and some extra auxiliary sends, you can get into the surround-mixing game. Of course, doing fancy spins around the five main speakers is challenging, if not impossible, without a true surround panner that employs a joystick or mouse. But some of my first experiments with static surround mixes were done with limited tools.
For instance, on a console without subgroups, if you have four extra aux outputs, you can patch the stereo bus to the front left and right channels, aux 1 to the center channel, aux 2 to the LFE channel, aux 3 to the left surround, and aux 4 to the right surround. (Of course, if your console has only four aux sends, this leaves you no way to easily add effects.) The LFE output should be sent through a lowpass filter with the cutoff set somewhere around 80 Hz so that the LFE information is completely out at the 125 Hz "speed limit." This patching works fine for static mixes, such as symphonies, in which you're only setting the relative levels in each channel for the duration of the track.
Getting an audio source to pan across the sound field takes a few tricks with subgroups. On a console with at least four subgroups, you can assign buses 1 and 4 to the front left and right channels, bus 2 to the left rear, and bus 3 to the right rear. Patch separate aux buses for the center and LFE channels.
Now, by panning between odd and even buses, you can perform front-to-back moves in the surround sound field. Also, by selecting a combination of buses, such as 1, 2, and 4, you can even manage a diagonal pan. This system can work out quite well with consoles that have pan and subgroup automation.