I've been producing a commercial DVD titled The Universe, and one of the most challenging aspects of the project has been building the 5.1-surround soundtrack.
The primary content of The Universe is a huge number of animated stills and movies showing the amazing astronomical images being sent to Earth by various space-based telescopes. The well-known Hubble Space Telescope captures images of cosmic objects ranging from within our own solar system out to unbelievably distant extragalactic regions. The less famous TRACE and SOHO telescopes shoot incredible close-up videos of the surface and corona of the Sun.
From the initial stages, I planned to create the entire project in my home project studio. The tools were readily available, and I believed I could push the project through — at least to the replication stage — with only moderate additions to my audio-video setup. I knew it would be a challenge but that it would ultimately be worthwhile for a number of reasons.
I cut the video, including the music and narration, in the Windows-based Vegas Video from Sonic Foundry. I use Vegas because its interface is remarkably intuitive and fast, and its real-time preview capability saves me huge amounts of time. My plan was to encode the edited video from Vegas to MPEG-2 and then move it to a Macintosh G4/867 MHz to author the DVD with Apple's affordable DVD Studio Pro software.
I knew that to be competitive the DVD would have to make the most of its surround-sound capabilities. I also knew that it would take some work. The first hurdle I encountered was that most of the music I was using, composed by Mix magazine's Paul D. Lehrman, was mastered in stereo. In concept, converting Lehrman's stereo mix into surround was simple. I wanted to avoid any aural trickery, such as extracting musical parts and flying them through figure eights behind the listener's head. Instead, I decided that the most effective approach would be to sweeten and process the music in ways that took subtle advantage of the surround channels.
The first and simplest step would be to put reverb and possibly some delay in the rear speakers. That would strongly establish the viewer's sense of being in the same space as the images. As you may know, there is no sound in outer space. Yet this literal reality has been violated in so many sci-fi TV shows and movies that an entire vocabulary of outer-space sounds has become common parlance. For example, big nearby ships emit a loud, deep hum; lasers make a quick, resonant filter sweep; passing spacecraft and various meteorites go whoosh; and so on.
Those conventions are so well established that people not only accept them but expect them. That's fine with me, because I can use that vocabulary to enhance the audio experience of my DVD. My goal, therefore, was to envelope the viewer in a sizable environment with a big, smooth ambience that would suggest the feeling of being in an immense, intergalactic space.
The next step would be to add percussion tracks, principally to the sides and rear. Adding melodic and harmonic sweetening would also be possible, but much more difficult without sheet music or even chord charts to go by.
To turn my stereo studio into a surround room I had to get some new tools, including Mark of the Unicorn's Digital Performer 3 sequencer and 828 digital-audio interface to run on my Mac G4.
Digital Performer 3 boasts serious surround capabilities: the program has several types of panners, each with different strengths and functions, and it even lets you mix with a joystick. The panning functions were especially important because I use an analog mixing board, which I've been reluctant to part with for various reasons (good busing, lots of aux sends, onboard EQ, and it's quiet). I also needed to use several MIDI instruments to get the sweetening and sound effects I wanted.
The new 828 seemed like an ideal solution. It uses FireWire (rather than the slower USB) to communicate with the computer. FireWire has been a blessing when I have worked with the greater bandwidth demands of video. The 828 also has eight analog inputs and outputs — more than I would need — and it offers excellent software control and compatibility with Digital Performer 3.
For speakers I picked a 5.1 set from Cerwin-Vega. They're consumer speakers, but they are accurate in addition to being compact and affordable. A more difficult task than choosing the speakers was finding a way to feed them the signals.
If you're turning a stereo studio into a surround room, one of the things you absolutely need is an amplifier that supports surround sound. On the output side, that's not too difficult. For as little as $175, you can find an AV receiver with five outputs for your left-front, right-front, and center-front speakers, as well as your left-surround and right-surround (rear) speakers. You also typically get a line-level output for a powered subwoofer.
Somewhat more difficult and expensive is finding an amplifier with six discrete line-level inputs. Some AV receivers have them, and some don't. Project-studio budgets usually dictate that you go for consumer gear. Because you can't count on help from the salespeople at consumer stores, you have to do some research to find the right box. (The Web is invaluable.) Check the manual and, more importantly, the actual rear panel to make sure the unit has six analog inputs (usually RCA jacks) labeled something like DVD Six Channel or SDDS Inputs.
In the end I chose a Kenwood VR407. Because it's about a year old, I was able to get it for $185. It's rated at 100W per channel, has a reasonably simple setup (and an awful manual), and even says 6-CH Input on the front.
Once you install the driver for the 828, remember to open the 828 control panel and set it for a 48 kHz sampling rate (see Fig. 1). That's important because Apple's A.Pack, which is a good tool for encoding your final mix to MPEG-2, doesn't work with any other sampling rate. You don't want to get all the way through a project only to realize it's set to 44.1 kHz, and you have to reburn it at 48 kHz. The good news is that once you import a sound file as a Soundbite into Digital Performer 3, it converts the sampling rate simply and quickly.
THE HANDS GET DIRTY
Configuring all that gear took some time. Although Digital Performer 3 and the 828 are perfectly suited for surround mixing, their manuals focus primarily on recording in stereo to tape, and you have to dig a bit to set them up to do surround. The primary distinction between the two approaches is that the final output for DVD consists of six mixed audio files on disk and not six tracks on tape.
My job was further complicated by the need to use MIDI to sweeten the music and do sound effects. Although Digital Performer 3 does MIDI and audio, and though its surround mixing works as advertised, basic MIDI tracks themselves cannot be panned in 5.1 — only in stereo. That's because MIDI synths and sound modules are stereo devices, and the MIDI Specification doesn't accommodate surround mixing.
So just how do you route, monitor, and mix a 5.1 soundtrack using MIDI and audio? The answer isn't immediately obvious, but here's the solution: you have to use the audio outputs of the 828 for monitoring, not for recording. When I first realized that, I unconsciously resisted the notion and wasted time looking for another approach. That may have been because the 828's outputs are of such high quality and are intended for final output. Quarter-inch TRS jacks with 24-bit D/A resolution seemed like overkill for monitoring. When I finally accepted that fact, though, things became clearer.
I then connected six of the 828 audio outputs to the AV receiver and hooked up the receiver's respective speaker outputs to the surround speakers. The receiver also had to be set up internally to accept six discrete line-level inputs and to route each to a separate surround speaker.
To set up surround monitoring in Digital Performer 3, you open the Audio Bundles window, where you can create various input and output configurations by moving little blue squares (tiles) around on the screen (see Fig. 2). For example, to set up the left-surround channel: in your Digital Performer 3 5.1 audio-output bundle, move the Left-Surround tile to the column that represents the physical 828 output connected to your left-surround speaker.
Follow that procedure for each of your six channels, and the worst is over. Double-check that the subwoofer output goes to the right place. If the Monitor amp has an output for a powered subwoofer, make sure you connect it to a powered sub. The other option you have is to send it to an unpowered sub with a passive crossover.
Assigning an audio track to a 5.1 output bundle equips the track with a surround panner in Digital Performer 3's Mixer window. That lets you send each Digital Performer 3 track to any combination of the surround speakers (see Fig. 3). That takes care of monitoring; the next step is setting up for MIDI.
THE MIDI QUESTION
As I mentioned earlier, I needed to compose, monitor, and mix my MIDI tracks to surround while combining them with the original music. Because MIDI tracks can't have surround panners (remember, they're triggering stereo devices), that posed a problem. Composing and recording would be simple; I could do that in mono or stereo. But how could I mix those tracks to surround? Fortunately, Digital Performer 3 offers a couple of options.
The approach I took was to bus all of the channels on my mixing board (the hardware, analog one) that were assigned to MIDI instruments to a single stereo output. I connected that stereo pair to a pair of 828 inputs (real, not virtual). I could have chosen any pair, because all of the unit's audio inputs were available; the rest of the audio was coming from the Macintosh through FireWire.
That done, the first option was to create an input Audio Bundle for the 828 inputs that were receiving the synth signals. I routed the inputs to a Digital Performer 3 aux track and assigned the output of the aux track to the surround output Audio Bundle. The aux track, like Digital Performer 3 audio tracks, does have a surround panner, which lets you put your synth sounds where you want them.
All that routing is not free, though; it creates a latency that delays the synth sound. Digital Performer 3 has a control panel to adjust that, and it may not be a problem in some situations. Because I like recording in real time, however, I didn't feel comfortable working with the latency. So I looked for another solution.
I solved the problem with a simple expedient. I recorded my mono and stereo MIDI tracks straight in as usual. I didn't hear them in surround, but I did hear them in time. When I had a performance I liked, I recorded the MIDI track as an audio file, and the problem was solved.
Recording the MIDI as audio is a no-brainer. Solo the MIDI track, add a new audio track, and set its input for Analog 1-2. Record-enable that track. Press Record, and you will write an audio version of the MIDI track to disk. The MIDI and audio tracks are even in sync. I heard no phasing, canceling, or other out-of-sync artifacts. I moved the MIDI track down out of the way and moved its twin audio track up to where it was accessible, and I was rocking.
MIXING TO DISK
When all of the tracks are right, your last step is to record the whole piece to disk. Digital Performer 3 has a great Bounce to Disk feature that can create six mixed Sound Designer II files from any portion of a project that you select. Those are the files that you encode as MPEG-2 for importation into DVD Studio Pro to author and burn the DVD. It's as simple as making a selection and choosing a menu item. All the audio, MIDI, and effects generated by Digital Performer 3 plug-ins are combined just the way you set up the project.
You pay a price for mutating a stereo project studio to do surround, but it's not too high. The toughest part is figuring out exactly what has to be done to get all of the players speaking the same language. Once my system was in place, I took a few small test projects from start to finish, and everything seemed to work.
At the moment, The Universe is still in production — on the video as well as the audio side. But it should be completed by the time you read this. The project offers new challenges every day, it looks and sounds great, and it presents another step forward in desktop production. If you'd like to see some samples of the project, check out www.universedvd.com.
Tim Tullyis a musician, composer, and video producer who has written hundreds of articles and cowritten two books on media production.