Dire Straits’ ‘Brothers In Arms’ needs no introduction. It was arguably the Compact Disc’s ‘killer-app’ album which almost single-handedly launched the new digital format over two decades ago into households across the world, not to mention bringing digital recording technology into studios. Two decades later at Mark Knopfler’s own recording facility at British Grove Studios, the album gets the high-resolution surround-sound treatment we’ve all waited for. Renowned audio engineer, Chuck Ainlay — known to some as “Mr Digital” — explains to High Fidelity Review’s Martin Fendt how he remixed this project to DVD-Audio and SACD formats.
Before going into all the ‘nitty-gritty’, it is worth noting that the honour of following in Neil Dorfsman’s footsteps to remix Dire Straits ‘Brothers In Arms’ (BIA) onto high resolution DVD-Audio and SACD came to Chuck Ainlay mainly by circumstance, since he has been the default producer and recording engineer for all the recent Mark Knopfler solo albums as well as the last couple of Dire Straits ones — ‘On Every Street’ and the live album ‘On the Night’. Once he received the call and agreed to take on the project, the most pressing task was to actually locate the original master tapes — which as luck would have it, were scattered far and wide.
“When we started out with the search for the tapes, we assumed that for an album of that magnitude these things would be highly protected in a vault and everyone would know precisely where all the assets were. But as it turned out, some of the reels were in LA, some were in Europe in the Mercury vault, and some were found in Mark’s own vault. In fact I’d been asked to come over to British Grove Studios in London a couple of times to mix the album, only to realise that we didn’t have all the assets we needed. It was not until the second week of January 2005, when I went through all the notation, that I was pretty certain we had everything necessary to mix the album.”
Transferring from 16-bit DASH Format
It should be noted that the original transfers were recorded digitally to the then new and pioneering Sony 16-bit 44.1kHz DASH format, which at the time only supported 24-tracks. Ainlay also faced the fact that pre-emphasis had been applied to the recordings on the DASH tapes. In essence, this was intended to help reduce PCM quantisation noise by basically boosting the high-end on record, and then decreasing it again, with the inverse slope to flatten the response back out on playback. Even today, the problem is there is no practical way to strip the emphasis whilst staying in the digital domain. In the end, the solution Ainlay came-up with was to use the machine’s analogue-outs which would output with the frequency response re-corrected on playback. Moreover, to achieve the best possible sound quality, the team used the latest model of Sony DASH machine they could lay their hands on: the 3348HR which used much better converters than those of the original 3324.
The next stage was to directly feed the discrete multi-track output, along with the 24-track analogue slaves, into Apogee 16X converters and save everything onto hard disc using Steinberg’s latest Nuendo software running at 96kHz and 24-bit. The digital-audio-workstation (DAW) PC used was supplied by AMD and was equipped with dual Opteron 64-bit processors. While this process was underway, everything was locked to time-code to make sure it was all perfectly synchronised.
Subsequently Ainlay A/B compared the captured tracks with the original master to make certain that the tracks he was using corresponded to the ones used on the original production.
“This stage was necessary because in many cases there was no documentation on the tape boxes, or in some cases, the track sheets were missing altogether,” he notes. “I just had to do much of the checking by ear to make sure that every element was present and correct. It was quite a ‘needle in a haystack’ search to try and find all the relevant masters, and so it was therefore a somewhat tedious process just to get to the stage where we could say that we had everything, and were in a position to actually mix the record.”
The digital tracking then came directly out of the DAW via 48 discrete channels of Apogee 16X D-to-A conversion and mixing was subsequently performed on a Neve 88R console. In short, Ainlay was able to bus-out the appropriate tracks all individually from Nuendo and directly into the Neve console. Moreover, there were additional channel outputs from Nuendo comprising a complementary five-channel surround-effects mix which was created in Nuendo , and which he could then bring up on the console to mix in together with all the other analogue and digital effects on the Neve. SSL’s new surround compressor was used on the Neve’s mix buss insert, before the master fader output feeding the mixdown machine.
The resultant 5.1 mixdown — which did not involve any analogue tape intermediate stage — was then saved to another Nuendo system via a set of Prism A-to-D converters, which was once again performed at 96kHz. Finally, when all of that was complete, the mix went to Bob Ludwig at Gateway Mastering where he performed the final mastering, EQ and compression.
“I recall spending about 12 days actually mixing the album at British Grove Studios in London,”
“There are nine songs, so it was not a long project in that respect, and I pretty much mixed one song each day. However, adding in the amount of time it took to complete all the transfers, we are probably looking at 20-25 days to produce the entire album.”
Analogue versus Digital “In the Box”
Rupert Coulson assisting with the remix of Dire Straits’ ‘Brothers in Arms’.
Some people naturally question why — given all the high-resolution digital technology available today — did BIA’s 24-bit, 48 channel tracking actually come out of the digital domain and be mixed in analogue before going back into digital again? Ainlay explains:
“In essence, it should be appreciated that to accurately model a compressor or an equaliser digitally would have taken huge amounts of horsepower from the computer, especially if you are talking lots of channels of EQ like we require when we are mixing a record — practically everything has to have an equaliser on it and some degree of compression associated with it. In short, to try and do all that ‘inside the box’ — with EQs and compressors which really model the analogue equivalent — is still something which is not quite possible, even with today’s massive computers.”
“Of course there are Nuendo plug-ins which I do find particularly useful, but I still find to a large degree that I just can’t realise the same sort of high-quality result from a digital equaliser unless I was able to do it on some highly oversampled one. However, once again, I stress that it eats up so much horsepower that I can only use it on a few channels. So for these reasons I still prefer to mix on an analogue console for the sonics as well as the ergonomics. Therefore my approach is still basically a hybrid one — until things progress further. Granted, things are getting better, but we are still not quite there yet.”
He points out that Steinberg’s most recent Nuendo release [used on ‘Brothers in Arms’] already benefits from the new Opteron Dual core processors which provides the computer with nearly twice the power. Looking ahead, he hopes that future versions of Nuendo like Cakewalk’s Sonar 5 will be written with native 64-bit floating-point engines [rather than the 32-bit of today’s applications] to take full advantage of 64-bit operating systems and processors.
“Only when this does eventually happen, might we get to the point where I can actually mix entirely ‘in the box’ with results as good — or better than — I can presently only achieve in the analogue domain on a quality console such as the Neve 88R.”
As an aside, readers will certainly be intrigued to know that this “hybrid” and arguably convoluted D-to-A-to-D-to-A-to-D signal routing, by coincidence, actually mirrors what Neil Dorfsman undertook in 1985 with the original stereo album. The main ‘architectural’ difference when comparing the two respective approaches being the intermediate tracking into Nuendo at 24-bit 96kHz resolution. “The original album, which I consider to be a masterpiece, was mixed in a SSL4000 analogue console from the analogue outputs of the DASH 3324,” says Ainlay.
“However, I would still consider that to be an ‘all-digital’ album. To be honest, at the time, there was no way of doing a pure DDD album, so the analogue stage of mixing through a console was never differentiated on the CD jewelbox. Anyway, apart from a few analogue slave reels, there was never any analogue tape storage stage where you would ‘lose’ it, so to speak.”
Preserving the Original Intent
In essence, for the 20th Anniversary edition of ‘Brothers in Arms’, Ainlay stresses how he certainly didn’t want to deviate to any degree with regards to the creative approach which Neil Dorfsman had applied all those years ago. “I referenced continually back and forth since I was creating a mix to be as true to his work as I possibly could,” he recalls. “Of course he didn’t have available to him the same sort of quality converters as I do now. Back then Neil was mixing to something like a Sony 1630 which were pretty awful converters by today’s standards, and moreover, he was mixing from one of the original 3324 DASH machines with inferior converters to those in the HR.”
“This time around I definitely feel I had some advantages as far as being able to ‘warm up’ the album, and take advantage of new technology. But I also tried where possible to use the same sort of equipment which Neil would have used — i.e. space stations and plates and chambers which are the real analogue devices, rather than some digital emulation of them. Having said that, I also took advantage of the digital domain, where it made sense, by using some of the cool hip things which are available now as Nuendo plug-ins. For example, to make the surround more interesting, I used fair bit of Universal Audio’s UAD1 card which has a really effective EMT plate emulation, not to mention Fairchild, Pultec, LA2As and 1176s which are some of the best plug-ins available which use dedicated processors on the PCI board, as opposed draining the processor of the PC itself. But at the same time I also used some host based plug-ins in the PC itself where necessary. Of course the analogue Neve console has great compressors and equalisers as well as allowing me to use all my other vintage outboard equipment.”
Readers of High Fidelity Review are no doubt aware that some so-called audiophile purists become alarmed whenever the terms ‘compression’ and ‘equalisation’ are mentioned. When this point was put to Ainlay, he acknowledged this widespread perception.
“Many well-meaning observers simply don’t understand what it takes to make a record such as this. For a start, the output from the microphones — depending on where they are placed — invariably doesn’t sound like the source or the instrument. In many cases there is ambience which needs to be dealt with, such as when mic-ing a drum kit, where the microphone is unnaturally close to it. In such a situation, what that microphone hears is not what you would hear in the room.”
Furthermore, he stresses that recording and mixing is an ‘art’ which is all about allowing a sound which is actually quite loud to still sound good at a lower volumes. As for the dreaded ‘EQ’, what he also points out is when one sound has the same tonal spectrum as another, it would tend to mask that other sound. For example, if one would put these two full range sounds up against each other, one might well cancel-out the other.
“I therefore need to use EQ to ‘carve-out’ a place for an instrument to ‘live’ in. I believe that the only way you could even conceive of doing an album without any EQ or compression is if you just put up a pair of stereo mics in the hall and leave it at that. But making a pop album — where you have to make it exciting to hear, with many discrete sounds from several speakers — is a wholly different creative process which takes many years to learn how to do. Moreover, compression and EQ are an essential part of that.”
Bass Management — Avoiding the Pitfalls
Now to the vexed question of bass management — a stage which is so often badly implemented from disc to disc, and from consumer player to player. Ainlay insists that studios should never apply bass management onto the disc itself, i.e. applying a crossover filter to the main channels and redirecting all the bass below a certain frequency slope into the subwoofer channel.
“The problem which arises is that when this is reproduced in the home on a machine which already has 5.1 bass management, the end result coming out of the speakers would be dual-bass managed, leading to all kinds of phasing problems with the different filter slopes [i.e. one in the studio, and another in the player/HT system]. This issue is especially critical today, given that more and more people now have bass management on their home system.”
To this end, the approach with the BIA 5.1 remix was to assume that the listeners would have either a full-range centre, left, right, and surround speakers, or will have their own bass management. In short, would be listening to full-range reproduction on each channel.
“My feeling is not to use the sub as an accompaniment to what I have in the front and rear speakers, but rather, to use it as what it is indicated as — i.e. an LFE channel which was originally intended to be only used for effects in theatres.”
In the end result, Ainlay acknowledges that a discrete sub channel does have a useful purpose for multichannel music:
“I do use it, for example, to accentuate the bass drums, but that is very sparingly applied. On BIA there is also some thunder and some low-frequency percussion elements for which I logically use the sub channel to do what it does best, without taking anything away from the main speakers.”
The Mix — From the Listener’s Perspective
Unlike many so called surround-sound mixes which just use the rears sparingly and for ambience, Ainlay intended that for this project, the rear speakers would be highly utilised. Indeed, there is information placed in the rears which is as loud as that in the fronts. And we are not just talking about effects and ‘jingle-jangles’ in the back. To this end, and since this is not a live performance, Ainlay takes the artistic liberty of utilising the rears to a large extent thus surrounding the listener by the music as if he/she is ‘centre-stage’. He also makes the soundstage even larger by using effects to produce dimension ‘beyond’ the speakers themselves. He was of course mindful of the fact that the two ‘featured’ parts of Dire Straits were Mark Knopfler’s lead singing and also his solo guitar playing. Consequently, Ainlay feels that for the most part, Mark’s guitar playing needs to be in front of the listener.
“However, there are some deviations from that idea,” he notes, “such as when some of Marks’ acoustic guitars and rhythm instruments are placed left and right rear, complemented by various ethereal ‘off-in-the-distance’ effects.”
For example on one of the songs the saxophone is placed in the right rear while a guitar comes from a far-off place using the left rear, and all this whilst not distracting the listener from focusing on the actual lead subjects in the front and centre. Regarding the lead vocals, Ainlay carefully keeps these in the centre channel primarily, which also spreads out to the adjacent left and right fronts, being about 6dB down in these. The result is that even if the listener is off-centre, he would still find the vocal imaging to be anchored straight up in the front centre.
However, so as to make the vocals ‘float’ a bit more, Ainlay pulls them very subtly away the front speakers and towards the back by putting a sparing amount also in the rear speakers. This effect is also aided by the judicious application of delay effects, harmonisers and reverb etc. to create dimensions beyond the front speakers and give that same sort of depth which one would be used to hearing in a stereo presentation.
Now for drums and percussion: Where possible, and with the help of ambient microphones, Ainlay draws the cymbals back slightly and by putting them mostly in the front but with some also in the rear. The toms also float back into the room somewhat by doing the same thing, whereas the snare and the kick-drum are more obviously anchored on the front wall, whilst the piano is spread with effects so it fills the room entirely. There are also appearances of a venerable Hammond B3 organ, which, in some cases has been doubled-tracked such that there are moments when one of the Hammond B3 tracks is in the front left/right, while the other emanates from the rear left/right, to result in a huge swirling sound all round the listener.
Synth pads are also used to effect in both the fronts and the rears simultaneously which ‘fill-up’ the room. But besides these more ethereal ‘paddy’ elements, Ainlay is deliberately more discrete about placement of other cornerstone elements including accordions, saxes or percussion instruments which usually feature in the front left/right or surrounds.
“It is worth noting that listeners generally become pretty used to the speakers in the rear, and in terms of our human auditory perception, they start ‘going away’ — unless there is activity which moves around, such as percussive hits,”
“This is illustrated when you take a synth pad and place it in a rear speaker, and pretty soon you just don’t hear it in the rear anymore. This phenomenon occurs in real life where we are just so used to hearing things in front of us and turning our head to find out where that sound is coming from, so your mind just tends to start ignoring sounds that are continuous from the rear. So what I have accomplished, hopefully, makes for a lot of entertaining movement.”
Thoughts on DVD-Audio and SACD Formats…
And now to the very end product: As touched-upon earlier, Bob Ludwig performed all the EQ in the digital domain from the PCM 96kHz 24-bit 5.1 Nuendo tracks at his extensive facility in Portland, Maine. This was basically just a digital transfer through his EQ and compression at 96/24 for the DVD-Audio version, but then he had to resample it for SACD.
“I think both the SACD and DVD-Audio products sound great,” enthuses Ainlay. “However, from the purist point of view, I do prefer the 24-bit Meridian Lossless Packing (MLP) DVD-Audio (on the DualDisc) over the DSD hybrid-SACD version. On a previous Mark Knopfler album, ‘Shangri-La’, we mixed to lots of different formats: We had a Nuendo session running 96/24; We also mixed to a DSD Sonoma system. Interestingly, as much as we had heard about how great DSD recording technology was, we nevertheless felt that it altered the sound to some degree, whereas the 96/24 LPCM was more representative of the original mix straight out of the console. In addition, we also mixed it to half-inch analogue tape at 30 IPS, as well as to 15 IPS, one-inch tape in two-track. And everyone in the room had the same opinion that the 96kHz 24-bit PCM was the closest representation of the console bus, and coming second was the 15 IPS one-inch analogue tape. Thereafter, it was a toss-up of which came in third and fourth — i.e. was it the Sonoma DSD, or the half inch 30 IPS tape? So going forward now, we have basically determined that we prefer 96kHz 24-bit PCM for recording over the Sonoma DSD. I was really surprised at that. I thought that I would like the Sonoma DSD recording the best, but after speaking with other people and Bob Ludwig himself — who is highly knowledgeable of DSD — we feel there is nevertheless is an inescapable ‘softening’ effect which DSD imparts. To me this seems to emulate analogue to some degree, but is not necessarily the closest representation of the console output.”
High Fidelity Review would like to thank Chuck Ainlay for all the time he spent talking to us and for providing the extensive background information found within this article. We’d also like to thank him for allowing to use some of his photographs.