FREQUENTLY ASKED QUESTIONS:
Surround mixing is becoming increasingly important to all of us, rather than just those who work in audio-for-picture applications.Paul White and Hugh Robjohns explain the techniques and the technology.
Nobody can have failed to notice the recent onslaught of media hype about surround sound -- all the manufacturers seem to be trying to sell us surround monitoring equipment, surround processors and both software and hardware mixing facilities. But what's the truth behind the hype? And if you assume that consumer demand for music in surround is, or soon will be, significant, what does that mean for the project studio owner?
The number of questions on this subject from music technology students and project studio owners during the recent Pro Audio Focus recording techniques seminars only served to reinforce the impression that this topic is the source of much confusion. So here are some of the most common questions we've encountered, along with answers to help clear things up.
Q Home theatre already has Dolby surround, but how is this different from the surround systems being proposed for music?
Dolby Stereo was a system initially developed for the cinema and it encodes four audio channels through a relatively simple matrix in such a way that it can be delivered as a two-channel audio signal that is nominally stereo (and mono) compatible. When decoded via a Dolby decoder (the domestic version is called Pro Logic), the signal is expanded back into four channels, which are used to feed separate left, centre and right channel speakers at the front, plus a single surround channel which feeds two or more surround speakers usually wrapped around the back and sides of the cinema or listening room. A sub-bass speaker may also be used to augment low-frequency effects such as earthquakes and explosions, but as bass is more or less omnidirectional, a single sub-bass speaker behind the screen is often adequate. There is no mandate for a separate sub-bass channel in Dolby Stereo, but it is a requirement in the recent discrete 5.1, 7.1 and 8.1 systems. The centre speaker generally carries the dialogue and other sounds that need to be keyed to the centre of the visual image (the screen) while the left and right speakers carry the stereo elements of the soundtrack.
What many people don't realise is that the surround part of a matrixed Dolby mix is actually bandwidth-limited mono -- the same signal is fed to all the surround speakers. Low frequencies below 100Hz are filtered off so that small surround speakers can be used and high frequencies above 7kHz are removed to provide a psychoacoustic sense of greater distance and also to help disguise crosstalk anomalies that result from the matrixing system. In a home cinema system, this mono surround signal feeds two or more speakers positioned behind and/or to the sides of the listeners. In conjunction with the sound from the front speakers, this produces a convincing 'wrap-around' soundstage, but the rear speakers provide no real spatial information other than the general impression that something is going on behind you. Figure 1 shows a typical Dolby Pro Logic surround playback system.
Q What is 5.1?
Today, when we talk about surround, what we usually mean is 5.1 ('five-point-one' or often just 'five-one'). A 5.1 system comprises five full-bandwidth discrete main audio channels (hence the '5') plus a restricted-bandwidth sub-bass signal (the '.1') called the LFE channel (LFE stands for Low Frequency Effects). Again, we have left, right and centre speakers at the front and two speakers at the rear, but this time the rear speakers (SL and SR) carry separate full-range signals that can be used in any way the mix engineer sees fit. What this means is that sounds can be positioned much more accurately in the surround environment.
Q How do I arrange my speakers for surround monitoring?
According to the ITU recommendation (number 775, should you wish to check it out), the left and right front speakers should be 30° either side of centre, with SL and SR at 110° (±10°) from the centre. It is normal for the speakers to be located on the periphery of a circle, all angled towards the centre, which will be the listening 'sweet spot'. All five main speakers should be identical. The sub-bass speaker normally goes front centre, but some flexibility is possible in difficult environments, as sub-bass is not as directional as full-bandwidth audio. For 5.1 to sound as it should, the monitor system speakers need to be set up as shown in Figure 2, and all adjusted to generate equal sound levels for identical input signals.
Given that most people pay scant regard to the positioning of stereo speakers in a domestic environment, it's likely that the majority of consumers will have a less than optimum setup, but when you come to mix in surround, it's vital that you have five accurately matched loudspeakers set up in the right position in an acoustically suitable room. This means that your mixing position is likely to be closer to the centre of the room than with a stereo mixing system and that, of course, the centre speaker will probably sit right in between the left and right speakers behind the mixer. It is also much more important that the acoustic treatment of the room is distributed evenly around the space to control reflections from the rear speakers.
Q Why go to all this expense when record shops are still selling stereo CDs?
CD players are already being superseded by DVD players, and all but the cheaper models are capable of playing back 5.1 audio as well as your old CDs. There is also the new Sony/Philips Super Audio CD (SACD) format which boasts dual-layer discs with a conventional Red-Book layer (allowing stereo replay on existing CD players) as well as a high-density layer containing high-resolution stereo and surround tracks using the 1-bit, 2.8MHz sampled DSD format. All new Sony DVD players from Christmas will be able to replay SACD discs.
As yet, the final standard for the format of DVD audio is still 'before committee', but some surround material is expected in the shops by Christmas 2000 -- the story is that certain companies will go ahead and produce material using the audio element of the video format whether or not the DVD-Audio format gets the official blessing, though some issues have yet to be resolved over copy protection. SACD has very clever mechanical copy protection systems which avoid modifying the audio in any way; a significant advantage over the schemes proposed for DVD-Audio.
Q DVD videos use a data-compressed audio format to allow the sound and picture to fit on one disk. Does this mean that DVD-Audio will sound worse than CD?
While most DVD video discs do use data-reduced audio to make room for all that video data on the disk, DVD-Audio is a different format and will use 24-bit, 96kHz sound quality. It also includes options for 192 and even 384kHz sample rates, which is technically superior to CD and broadly equivalent to the SACD format (discussed earlier) as well as having more channels than conventional CD. While such technical excellence is laudable, the vast majority of pro audio equipment currently works at 24-bit 48kHz, with some working at 96kHz, so upgrading everything to work at 192kHz, or even twice that, is going to be horrendously expensive -- a tough decision when the format may or may not offer any audible improvement. How the transition to high-resolution audio pans out remains to be seen. Will SACD appeal only to an audiophile minority or will the Sony/Philips marketing machine win the format war against the late-starting DVD-Audio?
Q If the DVD-Audio format isn't even approved yet, and the DVD-Audio versus SACD battle is still undecided, is there any point in worrying about surround now?
It depends on what kind of work you're doing. If you're working on material for commercial release, then you should be aware that many record companies are insisting on having surround mixes made of all new material so that they can archive it until the time is right to release it. There are techniques that can be used to 'fake' surround from existing stereo audio, but the success of the results depends very much on the type of material being processed.
Q How do I mix for surround?
Sounds are positioned in surround in much the same way as they are in stereo mixes -- by 'power panning' (the difference in level between the same signal in different channels), except that you are balancing across five speakers, rather than just a stereo pair. Very often this is bodged on a regular mixer by using six separate channels for each source (representing the five surround channels plus the sub-bass) and feeding them to separate mix busses. These in turn feed the monitoring system and a multitrack recorder in order to capture the six separate signals that comprise the 5.1 mix. While this is not difficult to set up, it makes mixing difficult, and it makes dynamic panning extremely complicated as you have to adjust several different level controls at once.
A better solution is to use a joystick or a similar surround pan controller, and this is how many professional sound mixers work. However, in the project studio environment, a piece of audio recording software with virtual joysticks is a much more cost-effective solution.
Q How do we use these surround channels?
The whole point of surround audio is to give the listener a more interesting and involving listening experience. Nobody has all the answers yet, as surround music is still a relatively new art form, but new mixing techniques are sure to evolve.
The LFE channel is not mandatory in all formats and may not be implemented in some home systems, so you cannot guarantee anything you create specifically for the LFE channel will be replayed. All five main channels are full-range and should carry bass in the normal way, though some consumer systems may use small satellite speakers and split off the lower frequencies to a separate bass speaker just to make a compact system practical. For mixing, though, it's probably safest to make your decisions based on a set of full-range speakers.
Then there's the centre channel. In film work, it makes sense to pin the dialogue to the centre of the screen, but in music we're used to hearing stereo with no centre speaker. Mixing with all the mono components coming from a centre speaker sounds different to what we're used to and, because of the axial response of the human ear, it also makes the signal sound tonally different. So some producers and engineers may put the voice in the centre channel while others may want to stick with the phantom image we've all got used to in two-speaker stereo. There is also the issue of incorrectly aligned domestic playback systems. If, on playback, the centre channel is a few dB adrift of its correct level balance, a carefully crafted mix could end up with the vocals, bass guitar, kick drum and any other centre-panned sounds either too high or low in the mix. This is another reason why an increasing number of engineers and producers are avoiding the centre speaker and creating conventional balances across the left/right pair, which not only sound more familiar, but are far more tolerant of system misalignment. The only drawback is that the image placement will not be as solid as with a true three-way discrete frontal format.
Q What should I feed into the surround channels?
This is the subject that causes the most heated discussions, because it's a matter of art rather than science. In classical or acoustic recordings, the room acoustics (or simulated room acoustics) can be used to provide the listener with a 'best seat in the house' listening experience -- the music still comes from the front while the reverberant field (and audience noise, if a live event) come from all around. On the other hand, if you fancy yourself as a rival to Pink Floyd, or as a cutting-edge dance producer, you can put instruments all around the room. With some types of music this might be too distracting, so you may decide to play the game a little more conventionally while still moving some of the effects or incidental instrument parts out to the sides or even to the rear. The point is that there are no rules, so you can do whatever is most artistically pleasing -- keep in mind, though, that it may be prudent to help compensate for bad setup in consumer playback systems by keeping all the important elements of the mix near the front. Subtlety is the best approach, just as when mixing in stereo. Don't go mad by spinning sound sources around 360° all the time. Use special effects as exactly that -- as special, and therefore rare, effects. That way they become far more powerful and effective.
Q Once I've set up my surround mix, what do I record it on?
At the moment there are no dedicated surround mastering formats, so most people record to six channels of an ADAT or DA88, (using the other two tracks for a dedicated stereo mix). And, yes, we know that ADAT doesn't record at 24-bit/96kHz, so there will probably be a lot of surround releases starting life as 44.1kHz or 48kHz masters. There are a couple of different standard track layouts, but the most common (and that consistent with the ITU/SMPTE recommendation) is: L, R, C, LFE, SL, SR on channels one to six respectively.
One of the reasons this is recommended is because the 5.1 signal will typically be transferred to and from the eight-track machine as AES-EBU pairs. In the unlikely event that these three pairs are routed quite differently, perhaps incurring timing offsets of a sample or two, pairs that clearly need to be coherent will remain so: L with R, and SL with SR. This will also apply to the secondary stereo pair on tracks seven and eight. Under these kinds of error conditions a sample or two difference will make little or no difference: The LFE track can survive quite large timing errors without disturbing the image and, while the coherence of the centre channel to L and R remains a bit of an issue, at least any artefacts will be symmetrical.
If you want to save your work in a 24-bit, 96kHz format, it's probably more convenient to work on a computer workstation, and then to save the audio files to CD-R (in a data format rather than as a Red Book audio disc), recordable DVD or other data storage format. Note that 5.1 data at 24/96 takes up a lot of space compared with regular stereo -- you may only be able to save one or two tracks onto a CD-R.
Q I've already got a lot of stereo material that I don't want to remix. Can this be processed to work in surround?
There are various ways to process existing back catalogue other than remixing it. However, as mentioned earlier, the results will be better with some material than with other types. It also depends on what you're trying to achieve -- if you just want to create a wrap-around ambience, this will be easier than attempting to simulate ambitious surround panning.
Mastering engineers often split a stereo signal into its middle and side (sum and difference) components so that they can process the mono (centre-panned) components of a mix differently to those panned left and right. For example, they may wish to EQ or compress the middle of a mix without changing the way the edges of the mix sound. Once the processing is done, they recombine the signals to reproduce conventional stereo signals. In the context of surround, this technique allows the 'side' component to be artificially enhanced in order to add artificially generated 'rear of room' reflections that create a more convincing sense of space. High-end processors, such as the TC Electronic M6000, Lexicon 960L and Eventide Orville, have specialised algorithms to handle this type of work, but a lot can be done using regular stereo studio processors.
Q At the moment I can burn CDs of my own music very cheaply, but how can I get audio DVDs that will play on commercial DVD players?
Currently, the only DVD recorders on the market are designed as computer data storage devices -- the discs they produce can't be played back in a consumer DVD player. At the moment, DVD mastering is a very expensive and complex process, and is therefore beyond the capabilities of most home studios. We can only hope that, as the format becomes more popular, a DVD equivalent to a CD burner becomes available.