Bruce Emery

Beyond The Matrix: Dolby Digital Surround Sound Now

Published in Cinesonic - Experiencing The Soundtrack - AFTRS Publishing, Sydney © 2001
Transcribed by Andrew Sargeant & edited by Philip Brophy from the talk delivered at the conference in 2000

Dolby: A Brief Background

Dolby is a name that I'm sure you've all heard. It actually belongs to Dr. Ray Dolby, still active in the company. Dolby Laboratories is solely owned by Ray Dolby and has been in existence since May 1965. The term 'matrix' does not refer to the feature film partially made in Australia a couple of years ago; it is film sound parlance for the Dolby process of recording four channels into two and then playing them back as four channels in the cinema.

When the Dolby Matrix encoding system was first introduced, there was a lot of black magic associated with it. People were very, very concerned about how their films would sound, and there was a lot of myth - and reality - to some of the problems incurred by people. We've moved a long way forward since that time. Some remarks by Ray Dolby and other leading figures in Dolby Laboratories discuss the background of the company in a Dolby promotional DVD produced in 2000:

Ray Dolby

"People had been for many years - and by many years I mean at least three decades - aware of the need for some kind of noise reduction process. They'd been trying to create one and there'd been many attempts, and no one had succeeded. It was sort of a 'holy grail' of audio electronics."

"If I could create a noise reduction system that worked, it would be useful in a lot of different areas. The areas I had in mind were motion pictures, professional sound recordings, studio recording for master tapes, consumer recording at home and FM radio - and also land lines for transmitting the audio signals from city to city, from the studio to the transmitter."

"With my background in physics, I approached the problem from a slightly different angle, and I was able to solve the problem in a way that no one had before, and that's the reason for our success. It was a totally new approach, in which the high level signals were separated from the low level signals, and it was the quiet signals in which there's a noise problem and the loud signals when there is no noise problems. I fed these through two separate channels, no one had ever done that before."

Ioen Allen

"We were fortunate in that thirty years ago coincided with the arrival of the big multi-track tape recorders. Prior to that time, most professional tape recording had been done with two-track or four-track machines. Around 1969 and 1970, the 16-track recorder became very popular in commercial recording studios - but the 16-track analogue recorder is 'hissy'."

"Just like sixteen tracks of a tape recorder generate a lot of hiss, film technology means that sometimes you've got a hundred different tracks are played back together, and obviously there's an awful lot of noise there. When Ray Dolby introduced the A-type system, it was such a breakthrough, a totally different approach to the problem of noise, that our competition died instantly over night."

David Watts

"I joined Dolby Laboratories in 1977 - the day Star Wars first showed in London, so it was really a sort of milestone for Dolby. It wasn't the first Dolby stereo film but it was the one that really got noticed and the one that was really successful, and, if you like, it's the starting point of the whole film sound program taking off. Since then I think we've seen sound become so much more important in film theatres."

Ioen Allen

"I suppose that somewhere out there there's still somebody who believes that all the sound in a motion picture happens on location when the film is first shot. That behind the actor there's an orchestra and all the sound effects are done that moment. That there's one microphone hanging above the set and that's how the sound is done. Of course that's nonsense. The sound takes months to do in a motion picture, and about the only thing that is recorded on location will be the dialogue, but sometimes even that's replaced later. Before it actually turns up in the cinema, there are a lot of technical steps a film soundtrack goes through. The dialogue is all mixed together and there are numerous dialogue tracks, and during that process Dolby equipment will be widely used."

A brief chronology of important events in the development of Dolby

In November 1965, the Dolby A-Type noise reduction was demonstrated and then released to the professional recording industry. After a number of other developments, Dolby entered the film world:

* December 1971: A Clockwork Orange - the first film to use Dolby noise reduction on all pre-mixes and masters. It was released with a conventional mono optical soundtrack, but it was the first time Dolby noise reduction had been used.
* May 1974: Callen screened at the Cannes Film Festival - the first mono encoded film with noise reduction.
* November 1974: the Dolby stereo optical sound track was previewed to a SMPTE convention in Toronto using re-mixed sections of the film Star Dust.
* September 1975: Lisztomania - the first feature film for general release in a Dolby stereo optical sound track.
* June 1976: A Star is Born - the first 35mm Dolby stereo optical film with encoded surround sound effects.
* May 1977: the release of Star Wars.
* November 1979: Apocalypse Now - the first Dolby stereo 70mm film which was in the 6 track format, that we currently use for the major forms of digital surround playback: Dolby Digital, DTS (Dynamic Theatre Systems) and SDDS (Sony Dynamic Digital Sound). (Note: digital surround playback utilizes three screen channel; left and right surround channels; and a sub-bass channel in the front of the room - a total of 6 tracks).

Moving forward to 1982, Dolby initiated work on digital modulation techniques for satellite transmission and other technical processes. At the same time, Dolby SR - "Spectral Recording" - was being developed. This process was released in March 1986, and was a vast improvement on the old Dolby A-type noise reduction. In August 1986, the number of professional tracks of Dolby A-type noise reduction in the world was numbered at 100,000. In November 1986, the 1,000th Dolby stereo film - Heartbreak Ridge - was released to the world.

February 1991 saw the development and first previews of Dolby Spectral Recording in Digital (SRD), which was an extension of the work being put into Dolby transmission digital technology. This format was developed from a Dolby algorithm called AC3 (Audio Coder 3) and has been employed in many films:

* June 1992: Batman Returns - the release of the first Dolby Digital. The big breakthrough, I should point out, was that at that time the Dolby Digital sound track was printed on the side of the film: it was not carried separately.
* October 1993: the first Dolby Surround music mix - a re-release of Pictures At An Exhibition by Emerson Lake and Palmer.
* December 1995: Dolby estimated that there were 400,000 prints, Dolby Digital, made at that time with 40,000 in circulation in the world at any time.
* October 1996: Shine - the 500th Dolby Digital sound track recorded.
* May 1998: Dolby Digital theatres numbered 14,000 world-wide.
* In December 1998: Dolby announced the development with Lake DSP in Sydney of Dolby Surround headphones. This a new product initially being released on Singapore airlines.
* May 1999: Star Wars Episode 1: the Phantom Menace - the first production mixed in a new format Dolby Surround EX, which is the same format 5.1 Dolby Digital uses with an extra rear surround channel, so that you have a true 360 degree sound field to work in.

A couple of statistics to finish off before we go ahead:

* As of this year, Dolby have equipped 26,168 picture theatres with Dolby Digital re-play systems in the world.
* There are 2,500 Dolby Digital films released.
* There are now over 10,000 Dolby encoded films, that would include all formats analogue and digital.
* Since 1965 , Dolby have licensed 700,000,000 products in the world.
* 46,000,000 surround decoder products have been sold world-wide.
* There are now 4,216,000 5.1 Dolby decoders in home receivers.

We used to say: imagine if you had 50 cents for every Dolby product that Ray Dolby's ever sold. Well, he probably does.

Encoding & Decoding: The Dolby Process

Dolby SRD (Spectral Recording in Digital) has become almost standard production for film sound tracks today. But in SRD - itself a combination of the Spectral Recording noise-reduction with the Dolby Digital encoding of the data - we actually have two processes: (i) the production of the film soundtrack in a sound studio, and (ii) the encoding/decoding of that soundtrack in a movie theatre. The making of a film sound track in 5.1 channels has nothing to do with Dolby: it's simply 5.1 channels in an auditorium. As a sound designer or a sound mixer, you work with those channels. Dolby provides the technology to record those 5.1 channels and print them down the side of a piece of film and re-play it in a cinema.

I say "simply" because that's what it should be, and most times that's what it is. As with any photographic process, we occasionally run into problems - not to mention the problems you run into in other cinemas. The 5.1 format came out of the old 70mm era, which was when a time when 70mm prints were released with magnetically striped sound tracks and had 6 channels encoded on them. This was a very expensive process because the sound had to be striped onto the film after it had been printed and each print had to be individually recorded and the life of the print was very short. Although a number of cinemas around the world were equipped for that format, it fell into disuse because of the high cost of the prints and it wasn't until Dolby and then subsequently DTS and Sony were able to revive the format by providing it at a much lower cost and as an easier-to-manage solution.

AC3 encoding (the Dolby algorithm 'Audio Coder 3') is the basis of Dolby digital technology - not only in the film world but in DVDs and upcoming digital broadcasting. AC3 is described by Dolby as "an advanced perceptual coding technology for transmission and storage of up to five full range channels plus a supplemental bass only effects channel, referred to as a .1 channel." Along with the sound mix information, AC3 carries information such as 'meta data'. (Briefly, 'meta data' is information encoded with the sound track that tells the re-play system how to play that sound track based on the environment it's being played in.) Being a form of "perceptual coding", AC3 simply removes large numbers of digital bits out of the data stream in a way which makes it (the lost of bits or 'compression') inaudible to the listener in the theatre. This means that the 5.1 channels of PCM audio are actually reduced down to a single data stream which uses less digital space than one mono channel of PCM audio. It is a loss-based system: ie. there is no attempt to re-construct the audio back to its full digital bandwidth. So what you are hearing in the cinema is the effects of "lossy" compression. It's similar to what a MiniDisc would do, but the Dolby system also has in its perceptual coding technology a number of other strategies to conceal or mask any artefacts caused by that system. Dolby AC3 was developed from earlier AC1 and then AC2 algorithms, which were ways of transmitting audio using a minimum amount of data storage and a low speed transmission systems, even down to a low enough speed to be printed onto film. Out of Dolby AC3 we come into Dolby E technology, which has recently been released. Dolby E is Dolby AC3 modified for use in the video world. Dolby Digital started being transmitted in Australia in 2001.

Dolby E, Dolby Digital and Dolby DVD technologies are all licensed products - that is, you don't buy them outright; you simply license the use of the equipment to make your soundtrack. In the case of Dolby cinema, each production company is licensed by Dolby to use the recording equipment and the optical transfer equipment to produce a Dolby soundtrack. Included in that license is the use of the Dolby logo on promotional material, and part of that license agreement is that Dolby undertake to provide technical expertise to the production company and their sound mixing company to make the soundtrack and to take full advantage of the technical standard that's being offered.

That's where I come into the picture. I'm employed by Dolby to act as a technical consultant to a production company, so I'm available to answer questions at an early stage of production. At the final stage of production - ie. the mastering of the mixed soundtrack - I'll actually attend the mastering session and supervise the use of the Dolby recording equipment, and then I'm available to follow through with the film laboratory and release venues such as cinemas after the event.

In the process of creating Dolby Digital soundtracks there is an extensive amount of quality control applied after the soundtrack has been mastered onto film. Dolby are actually now able to predict the life of the Dolby Digital sound track on the film down to the nearest 50 re-plays before the soundtrack will no longer play due to print damage and/or mechanical damage. This figure outstrips the picture side of the print by 150 viewings at this point in time. The print format of Dolby Digital also carries an analogue soundtrack, so it makes it compatible with theatres that don't have Dolby Digital re-plays and - in a worst case scenario - a film projector on the back of a flatbed truck somewhere in South East Asia running off a small Honda petrol generator can play as a print's mono source and still be intelligible. This was very important in the development of any technology to do with film: ideal conditions are great, but so often picture theatres do not present ideal conditions - certainly not for sound in any case. So, Dolby Digital has generated an enormous amount of research and development and applied it to what you see and hear on the screen. It's not simply a case of Dolby Digital being a bit of software with some hardware that was developed which you use and in a couple of years you get some software updates to use it again.

Technical Aspects Of Surround-Replay In Cinemas

Why is there a centre channel in the screen? Why do we have to have left, centre and right channels? Can't we just have our stereo soundtrack and play that in the cinema?

The answer is quite easy. In a room that seats 250 or more, mono information in a straight stereo recording is generated by feeding it equally into either side of the recording left and right. Therefore, if you sit in between the two speakers, you will hear mono sound. If you are sitting off to one side, you won't hear a true mono sound. You'll hear it weighted to one side of the screen or the other. That becomes a huge distraction particularly if you're seeing people on the other side of the screen and the sound is coming out of the opposite side of the screen. You'll quite easily accept that their sound will come out of the centre of the screen even if they're on one extremity of another, but if you separate the sound and the visual enough, people will become interested in the process and lose track of what's taking place in the content. So, we have a centre channel for dialogue and some sound effects, and part of the mixing process involves the use of generating hard mono information, if necessary, out of stereo recordings. The advantage of using 5.1 - certainly the 3 screen channels of 5.1 - is that if you place dialogue coming out of the centre channel on the screen, you can place information such as music or sound effects on either side of that (in the left and right channels). This way, you get far better separation and you can increase levels of background effects and music against the dialogue because the listener will still have some physical separation of those sounds when they're listening in a cinema environment.

Along with the hard centre channel, Dolby Digital 5.1 has a sub-bass channel which carries low frequency effects. I'm sure we've all sat in a Cineplex and heard sub-bass thumping from the movie in the adjoining cinema (they always seem to be American films). The use of sub-bass has been around for a very long time. 20th Century Fox patented what they called "Sensurround" for Earthquake in 1976, where they actually tried to patent the use of low frequency sound information in a soundtrack. They weren't very successful with that but, again, the nature of replay systems playing back low frequency sound in variable environments meant that the format fell into disuse until a more reliable method of replay was guaranteed. Sub-bass certainly can create a number of interesting effects on people - particularly those who aren't involved in the sound industry or with making sound for movies - so you can certainly tell when it's been over-used. But we are getting a lot better at using sub-bass in our soundtracks now.

Why do they keep putting things in the surround speakers, because I was watching a movie and I was quite happy sitting there until I heard a noise from the surround and I turned around to look at it and I missed what was happening on the screen.

In terms of pass or fail achieving your outcomes with the soundtrack, I suppose you'd call that fail, completely. One of the features of 5.1 has been its ability to split surrounds, and in some cases there has been controversy in their use and abuse. But again, well-used surrounds aren't distracting and are part of the effect. We're now starting to see with the release of DVD and extended formats, surround music-mixing which a lot of people have been waiting for a very long time. Some people have been doing such music-mixing for a very long time, but we're now starting to see mainstream music being released in surround, and that is proving to be very interesting to hear how people are interpreting the use of ambient or rear or surround channels.

The 5.1 format in any form of digital technology also allows for an enormous amount of dynamic range to be recorded. Compared to current digital replay technology, the old mono optical soundtracks that we were using (and in some cases still are using) sound like telephone response. As well as the frequency considerations, the other important feature of Dolby Digital is the volume level that can be achieved in digital sound tracks. The industry is currently going through a stage of trying to compete on a volume basis. Certainly the Hollywood industry is, yet no one's actually realised that Dolby Digital is a system which can replay an enormously wide dynamic range - that is, from very loud to very soft sound. We're now just starting to see some people actually backing off on their loud sounds and starting to appreciate that 5.1 Digital technology can allow you to have a straight, clean replay.

This will also apply to television and DVD releases. Current practice for television is to allow for the worst case scenario: a small 3 inch speaker on the side of a television with an enormous amount of ambient noise in the room you're playing. If we consider those statistics I quoted earlier, the number of surround systems that are going into houses means that the general public's awareness and perception of soundtracks is going to increase. In 2000, Australian households had a DVD home theatre penetration of .4%, but by the year 2005 this figure is expected to have risen to 18.5%. This is something to think about for anyone who's involved in producing content at this point in time. It means a lot of the ideas we've traditionally had about mixing for television will have to be re-visited with the home theatre listener in mind. And the spin-off there, of course, is that people with home theatre sound systems are also ideally placed to listen to surround music-mixes and surround multimedia.

In terms of Dolby for cinema, we control the process very tightly. That is, productions are given a license to use the equipment for each production, and they re-new that for the next production that they make. By doing this, Dolby retains the ownership of all the encoding equipment so that it can be modified and improved at any point in time. It also means we can make sure that we can promote the best possible quality each time a film is made, rather than just have film studios purchase the equipment and then start hammering out soundtracks day in, day out. We maintain quality control over that side of things - in the film laboratories particularly - so that a Dolby production can be perceived as being the same high quality that is attached to the Dolby name in music re-production.

Why isn't Dolby so active in the cinema replay market? Why can't Dolby go in and clean up the cinemas and make them play the soundtracks louder, etc?

It isn't possible to do it - not for 26,000 screens across the world. We're slowly starting to see the introduction of Dolby Digital replay in more and more cinemas - certainly the newer cinemas that are being built - and we're getting better picture and sound quality than some of the cinemas that were made 10 to 15 years ago. It's a slow process and we are slowly improving awareness - particularly among cinema owners, who seem to feel that whatever they're happy to put up is what people are happy to listen to. Such cinema owners are now realising that they have to compete with DVD at home, and that people who actually want better quality from their cinematic experience can achieve that on their own now and will not be necessarily tempted out into the cinema environment to do that. So, it is a long drawn-out process that still continues, and in fact a recent development of Dolby's is the release of the first completely digital cinema processor in the world, announced in July 2000. This device is actually true digital from its input to its output to the amplifiers in the cinema, and it allows us to control the acoustics in rooms better, allowing us to give better quality replay. That's an important reason for people to actually produce their material in 5.1.

Granted that Dolby found it not feasible, or not reasonable, to police theatres and make sure that they are reproducing to an optimum level: THX have only got a few cinemas, yet they carefully police them. Why can't Dolby?

What it comes down to essentially is that THX allow you to put a plaque outside the door of your cinema and to advertise that you are a THX certified cinema. The THX program is a division of Lucasfilm and they are simply a company that polices quality. As a cinema owner, you can sign up to the THX program. You are inspected once a year; you have to achieve certain construction and electro-acoustic standards in your room; and you have to repeatedly meet those standards over each inspection period to retain the certification. The great thing is that if you go into a THX-certified cinema you can be more or less assured of getting the best possible quality of picture and sound. THX is concerned with sound. At the time Dolby started in the cinema business, there was no such awareness of quality in place, and Dolby elected to at least get people signed up with Dolby processors and Dolby equipment and slowly educate cinema owners into the fact that better quality will actually pay off in terms of audiences over time.

The biggest problem that we face with cinemas to date is the question of volume level. Films are getting louder and digital certainly has contributed to that. Studies that Dolby have made into the issue of loudness and how to monitor and control it show that film audiences are happy with sound tracks set typically between 3 and 6dB lower than film makers are happy with. Notice that I said audiences and film makers. Cinema owners and managers of course are concerned with complaints and exposure to legal action, so they are happy with much, much lower levels than that. Essentially, fewer people complain if a film is too low in level than if they feel a film is too loud. In the manufacture of the film soundtrack through the mixing process, Dolby control the monitor level in the mixing studio to a pre-determined standard: pink noise at 85dB on a centre channel, with 50% modulation of the soundtrack equalling a setting of Dolby 7 on their cinema processor. The theory says that if you go into a correctly aligned Dolby cinema and set its cinema processor to 7, you will hear the same sound level regardless of the disparity in size between the original monitoring room and the auditorium - but that's where the theory stops. The practice of cinema owners and managers is to place soundtracks much lower. There's not a lot we can do about that because essentially we only sell the processors. The people who can vote best are the audiences and also the film makers, if they still retain any control over their work at that point in time. The big problem you have is that you are trying to enforce a standard in an area where you can't directly influence the revenue stream.

You said that Dolby sell processors, but also that Dolby licenses them. Wouldn't there be an onus on Dolby to make sure they're being used the way they were designed to be used?

There's a possibility that you could do that, certainly. There are cinema chains in the world who take great pride in their presentation - both picture and sound - and who have no trouble spending money appropriately. The problems we have are with the people who don't quite spend that much money and who basically have an attitude of: well, if we spend $10,000 on this new wizzbang Dolby processor, how many extra people is it going to bring in to allow us to pay that money back?

In the cinema replay side, Dolby sell the product: there's no license there. It is an area that Dolby are concerned about, because obviously people are advertising that they use Dolby Surround or Dolby equipment, and there is an implication of quality being there. For the most part it is there, but we do tend to find - particularly in Australia - that there are also people who will spend a lot less than is desired for their sound systems. The interesting point to note is that if you watch a film in Australia and you're not happy with the sound quality, if you walk out afterwards and complain and ask for your money back, you won't get it. If you walk out while the film is taking place you can legally be entitled to a refund.

Designing & Mixing For Solby Surround

What are the levels that you set for left front, right front, left rear, right rear when mixing a soundtrack in the production studio?

In that situation, we actually set the same level out of each of the 5 screen channels at the mixing position in the room. So the rule of thumb is that the ideal listening position is roughly two thirds of the way back in the cinema, centre equally distant from either side of the room. The sub-bass level is actually a little trickier than that and, as I said, it is a comparative reading between that and the centre channel. So it's simply a case of sound pressure level.

When mixing, we also work to a frequency response called the "X Curve", which is actually a published frequency response curve, and expect that out of those channels as well. The X Curve basically states that you roll off the high frequency replay out of your speakers because high frequencies in a standard sized auditorium (250 plus) are actually built up over time by the reverberation in the room. Even a room acoustically dead of such size will still have a high frequency build up. So in effect we're actually dampening the high frequency replay so that the soundtrack is kept sounding even to the ear in the room, rather than looking absolutely flat on a computer analyser graph."

How do you set up your mix environment at the time of production in accordance with the bass management built into the AC3 processor?

In the cinema world, it's basically up to the user as to how they use sub bass channel (the '.1' in 5.1). Obviously, if they can still sit there at that excessive level, then there's probably still a few people that will want to sit there and listen to it as a replay in the replay environment. Other than that, there is no dynamic control over the sub-bass level over time. I've certainly been into cinemas in the past where managers have said: Oh, we actually turn our sub-bass down because our patrons don't like that. Another cinema said: Oh, yes, we always turn our surround speakers up much louder because they never put enough surround sound into the mixes. You sort of actually wonder if the director and the sound people and the musicians knew that they had that creative collaboration taking place. I'm sure they didn't.

Once we move away from cinema and into the DVD and home environment, we can actually start to incorporate some control over the replay, which is essentially user driven. Essentially all the surround sound home theatre processors will allow you to select how many speakers you have, whether they're big speakers or little speakers, whether you've got a sub-bass channel or not, whether you're listening in 5.1 or straight Dolby left, centre, right, surround or even in stereo. And based on the information you enter into the processor at that point, it will appropriately re-mix the sound track and, if necessary, compress the level to accommodate the listening environment that you've described.

Those parameters are the 'meta data' which I was talking about earlier. Meta data is basically information created by the people making the soundtrack which decides: if this sound track is being play through a single mono speaker, how tightly do you compress the dynamics? Do you feed surround information into that or not? What do you do with the sub-bass? Do you mix it in at a lower level or do you just throw it away completely? That meta data information travels impeded in the soundtrack through the replay system to the processor. You can essentially make optimum use of a variety of replay conditions in the one soundtrack, rather than the current practice which requires us to make separate sound tracks for home use, because obviously a full range cinema mix through a large sound system is a different type of soundtrack to what you would mix for a small speaker listening in a much closer physical proximity.

I'd like to know about the specs for the cut-off of the subs and what frequency range they carry and what frequency range goes to the other speakers. Are there set specifications?

The cut-offs for sub-bass activity usually goes down to about 40hz. The overlap in the system is particularly dictated by type of replay installation, the type of sub-bass speakers that are being used, and in some cases the processor in the cinema that actually controls that information. When we actually make the measurement to determine sub-bass levels, we compare sub-bass level in the low frequency end of the screen channel and the higher end of the sub-bass channel to look for overlaps and that frequency could be anywhere between 60 and 100hz. There aren't really any published specifications; it's more a standard that has developed by default.

And as I said earlier, sub-bass is one of the most difficult things to control acoustically. Sub-bass sounds travel a very, very long way, so building sub-bass cabinets and speakers systems that are effective and give a wide coverage in an auditorium is a practiced art. Not all systems are equal. We've seen systems that have long tube type cannons; other that have speaker arrays. One of the great problems you have with sub-bass is that to actually achieve reasonable sub-bass level in a room, you might actually find that 150 feet further down the back of the building, there's an enormous peak of sub-bass energy which is shaking someone's office to pieces. Something like that is difficult to control, and the low frequency of the waveform in length can measure metres. So a lot of power is also required to deal with it.

In your experience as a sound designer and also a technical consultant for Dolby, what do you consider to be the implications for sound design by working towards this sort of Dolby encoding? There must be some sort of aesthetic implications placed on the sound designer to work to the 5.1 system.

By choosing to work in the 5.1 environment you certainly now can get a much greater audience if you like, by way of consideration. We have a competing system which is the Sony SDDS system. It is a digital film sound technology which Sony promote for their productions. It is an 8 channel system - not a 6 channel system like Dolby's - but not many people use it because it's just not widely available in cinemas. Village cinemas in 2000 decided against implementing any SDDS in Australia. The advantage of going to 5.1 is that as a concept, it's been around for a very long time, so there's a base of experience to draw on, particularly with film for vision. People who have listened to 5.1 and then have to work with it understand that basically they'll be dealing with three front screen speakers combined with some rear surround speakers. I suppose the downside of that is they have a certain expectation of what they're going to want to hear, so that any kind of extreme use of the format - such as putting dialogue into the back of the cinema - is always going to be a creatively risky thing to do.

But then, as I said earlier, the increase in home systems opens up a field for sound design and music mixing that's never really been there before, so there's no real limit to what you can do because there aren't any rules. At this point in time, using Dolby equipment for 5.1 sound design and surround music recording for DVD and digital broadcasting has no restrictions placed on it - unlike cinema. As far as Dolby is concerned, people who physically make those soundtracks and master those recordings have the Dolby equipment to use as they see fit. We at Dolby have no say in how you actually use the system to do that.

As a sound designer, how careful do you have to be if you're designing for a 5.1 system where you may be dealing with replay situations in straight stereo or even mono? How aware do you have to be of phase cancellation issues? Is constant checking involved?

Essentially: yes. Like you would deal with phase in straight stereo recordings, to a certain extent being out-of-phase within the Dolby matrix (or 424 - the old analog standard) was still an issue. Excessive amounts of out-of-phase material in the front channels would be directed to the rear channels. It's just a by product of the matrix decoding system, where any out of phase material will go into those channels. I guess the beauty of using a relatively cheap piece of software for ProTools is that you can quickly encode it and listen to it. For more elaborate hardware solutions Dolby do have a product, and I believe there are a couple of companies building monitoring systems which will allow you to quickly switch between different formats to listen to how the down mixes are going to sound. So making mixes for different environments is certainly something that professionals in the industry have been aware of. It's simply a matter of making sure that you've got the tools to do that - as well as a bit of experimentation to understand what to listen for.

Is it wise to try and keep, say, your surround stereo mix and your front stereo mix as separate as possible in order to avoid those phase cancellation problems? When you're cross-checking your surround to mono as well as your front stereo to mono, should they not have common elements within them?

You can certainly do that in any of the intermediate stages - ie. straight stereo down mixes or Dolby Pro Logic down mixes. But if you've kept your surrounds very separated, they may in fact disappear. So you should consider: is the surround information only viable or valid when you have a surround environment? Or if you don't have a surround listening environment, do you want to try and incorporate some of that information into the rest of your mix? It becomes a creative decision really.

Is that part of what you do when consulting with sound designers in terms of what they want to achieve with their surround mixes as well as their straight front stereo mixes?

We're very fortunate with Dolby because you can hear the effects quite easily. Not to detract from our competitors, but if you actually mix a film in DTS you simply send a D88 tape of the sound track away and it comes back as a stack of DTS CD-ROMs which play with the film print. But, you have no ability to actually listen to the effect of the encoding process in the mixing environment. In straight Dolby stereo analog or Dolby Digital, you can monitor the effects of what you're hearing, and of how the final replay will be.

Dolby, DVD & Future Developments

With the advent of DVD technology and the acceptance of Dolby 5.1 soundtracks on DVDs, there's now a whole host of tools coming out - from the Dolby 5.1 encoder for ProTools to many others. For the sum of US$900 you can make your own Dolby soundtracks in your lounge room if you really want, so long as you've got that many speakers. There's no restriction on the use of it. You buy it; you use it; and with a suitable amount of DVD writing technology you can make your own AC3 soundtracks and release them out to the world.

Conversely, if you're looking at re-releasing or re-mastering an old title, you may need to actually go to a mastering or mixing studio to re-mix earlier versions of film sound to suit the new format. Dolby also manufacture a number of products for DVD creation and monitoring particularly. They produce the same thing: a file that plays roughly at 320 to 380 kbs a second data transfer rate. Essentially, that is very small by comparison to what the full bandwidth full linear PCM recordings of the same soundtrack would make.

Looking forward to other developments in the technology, people are already proposing 24bit 96khz recording systems. These of course use up far more processing and far more storage. The sound quality is arguably better, and at the moment there are such systems available as stereo systems. There will come a time, I suppose, when people will want them as 5.1 systems, and there are even moves afoot to look at 192khz sampling rate for the increased fidelity. At this point in time we would envisage that AC3 and Dolby 5.1 still is a viable alternative in those mediums, purely because the cost of creating and storing information at those data rates is still very high. Going to 24bit recording on disk, you double the amount of disk storage that you're currently using at the sampling rates of 44.1 or 48khz. There's not really enough 96khz hardware out there to know what the implications will be, but certainly it means all the gear you purchased last year will probably need to replaced by new versions or updated versions of it next year, if that's the way you want to go.

Dolby AC3 is also now starting to move Dolby Digital into the world of computer games and multimedia. As of July 2000, there were approximately 369 computer and video games using Dolby surround technology. That will either be straight Dolby Surround encoded material or, in some cases, people are replaying Dolby AC3 out of their computers with the kind of fidelity you would normally get in a cinema. There are now a number of computer sound cards that are equipped to decode Dolby Digital, or to simply output Dolby Digital into your home receiver. And there is a growing market for 5.1 music mixing in both the DTS format and Dolby format. To date there have been now at least 30 Dolby Digital music recordings made in various styles of music, from orchestral through to pop and rock music, and certainly record companies are experimenting with those kinds of mixes.

Is Dolby is looking at standards coming out from 3D audio spatialization - that kind of surround sound where you have the ability to move beyond the X/Y dimensions and add a vertical dimension in re-play?

I imagine Dolby are looking at those kind of technologies. There is a Dolby "Virtual Surround" which is, I imagine, a similar kind of thing to 3D or the Aureal systems which basically produce surround field experiences out of a minimal number of speakers.

That's also part of the work Dolby are doing with surround headphones. A company called Lake DSP in Sydney have developed a processing system where, listening on a pair of headphones, it is like listening in a cinema: the sound is not in your ears and very close; it's actually distant and you get hues that will give you a feeling of having sound coming from behind you. There's certainly been systems like that over the years, but the problem really is that there's just not been the uptake, and people don't want to go and buy another new system. But there's a lot of people doing work in a number of areas. It'll be interesting to see what does come up.

You mentioned software and plug-ins, but I'm wondering what I actually need with DVDs or possibly games to get full use out of the meta information in terms of telling, say a DVD that I've got 4 speakers and this is the size of them and this is how I've set them up?

Essentially the software tools and the connected hardware you use to master your 5.1 soundtrack will have the ability to enter the meta data into it. Typically, the meta data will have two main functions in Dolby AC3. One is "dynamic range control" which effectively applies compression to the signal (reducing the loud signals, bringing the low level signals up in level). The other main function is called "dialogue normalisation" which allows you to adjust the overall volume of your material, which you can do over time as well, so that instead of just sitting there and making lower level signals come up in level and restricting high level signals, you can also control the average volume of that replay.

Any of the Dolby-licensed products have an interface to set those numbers; then it is simply a question of management of sub-bass and surround information. Most of that information is interpreted by the decoders, which is either built into the hardware you'll use to decode the data stream, or you'll set it up as a user option. Such options would allow the user/listener to: (a) play a full range mix with no treatment of any kind in order to get the full sound field; (b) hear the mix at a lower level; (c) listen to a straight stereo down mix of the 5.1; (d) decide on whether to have the sub-bass and/or surround information. You can then control whether that becomes a full Dolby Digital soundtrack in replay, or a Dolby Pro Logic soundtrack that you can feed into a normal analog Dolby decoder to receive the surround information. But essentially, whether it's a Dolby "black box" or a piece of software inside a Macintosh with ProTools or any of the other systems that around, those meta data parameters will be there to be set.

From your experience of supervising the mastering for theatre release prints using Dolby Digital, is there a standard or a norm to adhere to for the meta data, or have you had experiences where you've been asked to change any standard or norm there?

I haven't had a great amount of experience in that regard specifically, but in talking to some of the Dolby people who have been dealing with this in the UK and Europe, meta data settings have really started to fall into a similar set of numbers. Over-setting of the dynamic range control will just make a mix sound unpleasant: it'll sound over-compressed, so it becomes an aesthetic setting. People who are involved in the making of the soundtrack aren't likely to go and over-compress their mix. Similarly with dialogue normalization: the dialogue norm setting can be over-ridden by the end-user, and the idea of setting meta data is really to not have that happen - to not have people sitting there trying to adjust their volume controls. The numbers are simple just 1 to 10 scale numbers and I believe that particularly the range control is the one that everyone's been most concerned with, and that seems to be settling into a certain band of numbers.

With Dolby 5.1 being used in DVD, will a 6 channel mix will be reduced to the same amount of data as a mono PCM recording for DVD?

In terms of just looking at numbers: yes, it's a big data reduction. In terms of just listening to the difference between linear PCM recording and a Dolby Digital, you would be reasonably hard pressed to tell the difference. Now that's what they said about mp3 encoding, but remember that you've been listening to Dolby Digital movies since 1992-93, and there aren't many people who go: I'm not going to go and see a Dolby film because it's a lossy compression system. The fact of the matter is that for professionals who've created a piece of sound, a listening test between the a Dolby encoded recording and a linear PCM recording will invariably point out differences in the sounds, but for the average listener, those things aren't there to hear.

So, for example, does the ProTools software actually do the data reduction itself, so you store a file?

Yes - you create a file that you can then hand to a DVD authoring house. I haven't seen the ProTools software in operation (it's made by a company called Kind of Loud) but I do know that it has some control over the amount of data reduction you do within a certain range. But ultimately the DVD system will determine the limit as to what can be stored. DVD is by no means perfect, but a lot of people are very interested already, so the development continues. There are other proposed disk formats that offer much greater storage times than DVD - in theory - but how far into the future we don't know. I suppose the question is: where does it all stop? As well, once we get to providing 24bit 192khz full range 5.1 or 6.1 un-encoded audio, how far further are people going to go before they'll settle upon the ultimate in quality?

Could you outline the major differences between the Dolby Digital and the DTS formats? Particularly in relation to perhaps the DVD environment?

The DVD environment is a different coding system from Dolby, though both deal with a data reduction system. One of DTS's marketing claims is that a lesser amount of data reduction - ie. a higher data rate - is used. The trade-off is that you do get larger files with Dolby, but in the DVD environment that's not really an argument.

In the cinema environment, the difference essentially is that DTS requires a separate CD-ROM to travel with the piece of film, and you go from cinema to cinema with that CD-ROM. If you don't have the CD-ROM, you don't have a digital soundtrack. With a Dolby film print, it's all there on the one medium and you don't have to worry about separate CDs. The DTS system simply prints a time code on the side of the film which their hardware locks up to and runs with.

To a certain extent, you can only do a listening test between the two systems if you've got the CD-ROM. Village cinemas seem to have a ratio where out of a 14-16 multiplex they will have probably two DTS replays in the building, whereas all of them will be Dolby Digital. So if you thought that one movie sounded better than other movies you've heard, you'd have to find out whether in fact they were playing Dolby Digital or DTS. It's not enough to just look at the credits and see that it was a DTS movie, because it's becoming standard practice now to release films with an analog soundtrack, a Dolby Digital soundtrack and a DTS soundtrack, and sometimes they'll even squeeze a Sony soundtrack on there just for good measure. But you can pretty much guarantee it won't be mixed down in 8 channel.

In the DVD world, you need to sit down and listen to the titles and look at the choices that are available. Typically, DTS have fewer DVD titles released than Dolby Digital, but it seems from the reading I've done on DVD players there seem to be a lot of players out there that will offer both formats as a pass-through, and there are certainly a lot of amplifiers that'll offer both decoding systems. So maybe you'd buy both.

One hears criticisms about the loss of quality induced by noise reductions systems like Dolby. What are your views about that?

I suppose you could make an argument to say that if there was a greater concern about the compression induced in the AC3 encoding, the format would die out, because it would just be too hard to manage and people wouldn't want to use it. I suppose the real argument is : is the replay standard acceptable for the vast majority of people who are listening to it as opposed to any alternative? In the 5.1 environment, if we wait for 5.1 channels of PCM on DVD, we're going to be here for a very long time. And already people are upping the ante by saying: well, in fact, we need even higher quality than that. So we wait another technical generation or two until we get an even better standard. What we're talking about - irrespective of whose sound tracks they are - is that there is an enormous amount of program material in the world already that is in the 5.1 format that can be readily adapted to DVD and played. It's a bit 'donkey and carrot': the manufacturer holds out a certain technology and says: if you do this, this many people will buy it. I think, ultimately, if you hold out a bit of technology that nobody wants to bite, then you go and look at another alternative.

There will always be concerns about the quality of replay by industry professionals, undoubtedly, and there are a lot of people who say that vinyl LPs are still the best sounding replay system. I could be persuaded by those kind of arguments, but the fact of the matter is that CD came in and the market embraced it. Why? Because on average CD will always sound pretty much the same, whereas vinyl deteriorated quickly and became expensive to set up and maintain. I think DVD involves the same sort of argument, both in picture and sound quality: yes, it's not ideal, but the alternative is what? Stay with VHS HiFi? It's a tricky question.

But really, from the first product that Dolby made, they were actually bending the rules. They were saying: we will take this sound; we won't record it and store it in its purest form; we will actually modify it, save it on tape and then modify it further when we play it back. This was done in the name of achieving an outcome: to reduce tape hiss and generated noise. So I suppose Dolby is 'guilty' of actually providing replays with systems and stretching the boundaries of the technology to achieve results through sacrificing X amount of quality. I don't know how you get around that. For sound people, it comes down to deciding how to make people fully appreciate what you've put into your work.


Text © Bruce Emery 2001.