A/V 106: Surround Sound For Broadcast TV
January 5, 2005
This column will be devoted to the basics of surround sound for video, as part of my series for non-audio folks. As I noted last month, plain old stereo still isn't all that well understood in the video world.
Surround sound is in worse shape. It is both poorly understood and poorly implemented in our business. We've barely even begun to figure it out, much less use it well. For all that, surround sound can be quite effective and entertaining.
Stereo is a two-channel program that gives us a highly entertaining sense of place and "phantom images" (sound sources that appear to come from places other than the loudspeakers).
Surround sound is a considerably more complex playback modality that involves six channels: one center channel, stereo left and right, surround left and right, and a subwoofer for low-frequency effects (LFE).
At its best, surround sound can provide a sense of envelopment--a very pleasurable sensory quality of being "in" the place of the sonic sources in addition to the various benefits of stereo.
FOUR AUDIO FUNCTIONS
There are four primary sonic elements in a 5.1 surround audio production for video:
1. Mono on-screen--As in stereo and mono, the primary audio element for video is the mono on- or off-screen voice. In stereo, this appears as a phantom image between the two speakers. In surround, it is assigned to the center speaker. As far as I'm concerned, the quality and clarity of this element is most important in audio for video.
2. Stereo Music and FX--Music beds and some on-screen effects and ambience are sent to the left and right channels. These serve to spice up and support the mono center channel and to give the visual image a kind of "sonic depth."
3. Ambience and Reverb--In surround, ambience (short-term delays and room tone) and reverb are usually sent to the rear channels to fill in the sense of envelopment, which can be extremely effective in sports events and similar live performance productions. Occasionally, in dramatic shows, specific appropriate off-screen effects are sometimes sent to the rear channels.
In a cost-cutting strategy for consumer systems, many surround playback systems use small so-called "satellite" speakers for mid and high frequencies in the five-channel array, and a single subwoofer for low-frequency information. Meanwhile, a low-frequency effects (LFE) channel is created or derived to feed this subwoofer. It works moderately well, though five full-range channels work much better.
A lot of this is quite straightforward. The problems lie in our need to deal with an impossibly wide range of end-user setups and needs.
End users may be viewing video on a legacy mono or stereo TV, or on a more modern system, complete with a set-top box feeding either to stereo or multichannel surround playback. Finally, they may be viewing in a fully-configured home-theater setup, designed to mimic the quality of a commercial film screening room or small theater, with high resolution and wide dynamic range playback.
At the same time, broadcasters transmit mono, stereo and surround audio depending on the nature of the program and its source signals. Meanwhile, cable and DBS distributors often modify the original program.
The implications of this are a little daunting. A wide range of playback modalities must all cope with a wide and somewhat unpredictable range of signal transmission modalities. It is often impossible for even knowledgeable viewers using excellent equipment to obtain satisfactory playback that conforms to established specifications.
The first major problem is downmixing. If we transmit a surround program, it needs to play back with as little loss of audio information as possible on mono, stereo and surround systems. To do this, set-top boxes that receive surround sound need to downmix it to stereo and/or mono. This isn't difficult, but in my experience as a surround music mixer, it's impossible to have one static mix setting work for all surround audio programs.
We also want to give the owners of surround systems a little more bang for their buck, so we indulge in upmixing as well, wherein surround and LFE channels are derived from stereo signals to create "faux surround" playback.
Keep it simple. Concentrate on the quality of the mono center voice, then make sure the stereo music beds and effects are convincing and acceptable when summed to mono.
Finally, where a little ambience, crowd noise or reverberance can be fed to the surround channels, do it (making sure it's mono-compatible). Be conservative. My rule of thumb is that we should barely notice the surround channels, except when we mute them.
More importantly, make sure that both the stereo and surround channels can be summed to mono without ugly side effects. The LFE channel should never be hyped, and should extend conservatively the LF range of the program when downmixed. Interestingly, in much music mixing, we avoid the LFE channel except to put in some reverb wash.
Keep in mind there are legitimate differences between the surround production techniques in television compared with film and music, where substantial post-production time and resources are available. In film and music, we are trying for a considerably higher level of nuance and sensory engagement. In TV, we are simply trying to add a little additional spice. Keep it in mind.
Thanks for listening.