Figure 1. Typical 5.1 surround to stereo downmix diagram. Click here to see an enlarged diagram.
Producing surround audio for broadcast is a process we all are either already facing or soon going to have to face. All HDTV broadcasts will be surround-sound capable, and most satellite and digital cable carriers already have the infrastructure in place to regularly transmit 5.1 surround sound. Further, all digital cable carriers use Dolby Digital to transmit both stereo and surround audio. The table-top channel box will either output this as an AC-3 data stream from its S/PDIF connector, or downmix the 5.1 channels into a stereo mix and send it out the analog stereo outputs.
With that said, even with broadcasts moving to 5.1 surround, most viewers will continue to hear your surround mix in stereo for many years to come. And maintaining the “do no harm” promise to the stereo listeners is difficult. It is a fairly straightforward process to create a surround mix that sounds good on the majority of surround playback systems; however, making surround mixes play back in stereo or even mono is more of a challenge. Producing a successful surround mix is not quite as easy as putting up three more speakers and throwing some ambient crowd noise into the back channels. The most important elements you'll need are a proper monitoring system and a way to check for downmix cancellations.
Let's discuss downmixing first. Here's what a downmix is and when it occurs. Every surround mix needs to go through a codec of some kind for transmission. Codec is an acronym for enCODe-DECode. It is the process where six or more channels of audio are squeezed into the space of two channels and then reconstructed back to the original channels. Both Dolby Digital and DTS are examples of surround codecs in common use, but you will only run into Dolby Digital (aka AC-3) for television broadcasting. These digital codecs really don't change the quality of your sound very much because they are totally discrete from channel to channel. That is, unlike matrix encoders such as Dolby Pro Logic and SRS Circle Surround, there's no steering interaction between the channels, so your mix will sound almost exactly like the original tracks after transmission and decoding.
All this would be perfect except that the majority of home listeners will be listening on a stereo or mono television. There is a backwards-compatible process where your six original channels are combined (downmixed) to stereo, and that is where most broadcast surround audio gets into trouble. Just like back in the day when we had to worry about stereo-to-mono compatibility and mixes that cancelled out when something was flipped in polarity, all surround mixes have to be checked for stereo (and mono) playability. To understand the potential pitfalls, let's check out exactly how a downmix works.
The default downmix ratios for Dolby Digital are as follows. The left rear channel is put into the front left speaker at -3dB down from its original level. The right rear channel is put into the front right speaker at -3dB down, the center channel is put into both the left and right front speakers at -3dB down, and the LFE (Low Frequency Effects) channel is eliminated from the downmix. These downmix ratios can be changed to values such as -4.5dB or -6dB within the Dolby Digital Metadata, but most broadcasters will use the Dolby defaults of -3dB all around. (See Figure 1.)
What this implies is that phase problems between each left front/rear and right front/rear channels will only be heard when those channels are combined in a downmix situation. The same thing goes for the center channel because it gets mixed to the left and right front speakers. For instance, if you introduce a static delay of some 30 or 40 milliseconds between the signal feeding the front channels and the rear channels, say to make the sound really big, those listeners hearing a downmix will have to contend with a comb filter altering the sound. This sort of sounds like a Wah-Wah pedal with echo (for those of us who are Old School).
Remember to monitor
Another thing that can cause cancellation failure during downmix is microphone placement during recording. For example, if you place your audience response microphones 12m apart from each other to cover the crowd, note that you essentially have introduced a 35-millisecond delay between them. If these mics are used for the front and rear speakers, then they will cancel as a comb filter during a downmix to stereo. And if they are used for the left and right rear speakers, they will cancel as a comb filter during downmix to mono. Either way sounds bad to the millions of listeners hearing your broadcast on a stereo or mono television.
How to hear if you are successfully transmitting your surround/stereo broadcast? You need to monitor for the bad effects. On every surround receiver on the planet, there's a little “stereo” button that will turn your 5.1 (or 6.1) surround mix into stereo, using the downmix ratios originally set within the Dolby AC-3 Metadata. Monitor your surround broadcast using a home-theater receiver both in surround and stereo (downmix) modes. Now you can hear what your final broadcast will sound like for both audiences. If the surround mix sounds big and full, while the stereo mix sounds out of phase or thin, then you have a downmix comb-filter issue. Moreover, note that a channel polarity can be accidentally flipped in broadcast, which can do things like completely cancel out your announcer's voice in the stereo downmix but still sound fine in surround. The only way to know if you have a polarity flip or comb-filter issue is to monitor both in surround and downmixed stereo. Anything less than that is asking for trouble.
Keep in mind that most of the surround speaker controllers from companies such as Martinsound, Studio Technologies and Blue Sky (among others) offer a downmix monitor switch as part of their controls. This downmix switch allows you to simulate the 5.1-to-stereo downmix in your monitor speakers while still broadcasting or mixing in surround. It is a great way to test your mixes for downmixing compatibility before they go to air.
Basic mic techniques to avoid downmix problems
Mike Sokol and Hector La Torre conduct more than 30 surround sound workshops each year at universities, AES and Grammy sections across the USA for Fits & Starts Productions. More surround information and a tour itinerary can be found at www.howtosound.com.
Avoid spacing microphones far apart for ambience pickup. Try a simple ORTF (Office de Radiodiffusion Television Francaise) coincident microphone technique, or X-Y stereo pair to pickup crowd response. (See Figures 1 and 2.)
The ORTF stereo is a common stereo recording mic array, so it is easy to purchase a simple spreader bar from, for instance, Atlas (among others) that will hold two cardioid mics and mount them in the proper position. The X-Y stereo pair technique is simply a pair of cardioid mics crossed in a V pattern at 130 degrees. This offers a bit less spaciousness in stereo when compared to ORTF, but does downmix to mono a little better because the capsules are spaced closer together.
If your budget permits, you can also use a dedicated surround microphone — such as a Holophone (www.holophone.com) — for crowd response, remembering to always fly your mics high enough above the crowd to avoid having any single person in the crowd stand out in the mix.