A job helping to provide the sounds and recording the music for an animated Christmas special had SOS contributor Paul Ward hastily upgrading his studio and exploring the deepest recesses of his synth collection, in search of sleigh bells and singing reindeer patches...
When I was first approached to record the music for a forthcoming feature animation my immediate reaction was "Great! When do we start?". It's not every day you get the chance to record music that will be heard on prime‑time Christmas Day television! And just how did I manage to get involved in such an exciting job? Well, it was all down to my brother‑in‑law, Richard Taylor Gmus PPRNCM (to give him his full professional title!). Richard is a highly talented, professional, classically trained musician and writer of such successful musicals as Once Upon a War and Whistle Down the Wind. With such a reputation, Richard (amongst others) had been approached by production company Cosgrove Hall [you may know the company especially from such perennial animated favourites as Dangermouse — Production Ed] to submit a piece of test music based on storyboard drawings of the opening title sequence for Father Christmas and the Missing Reindeer.
This animated feature is a roller‑coaster ride of mystery, comedy, action and adventure as jolly Mr Claus (given a voice by actor David Jason) attempts to find his absentee antlered chums in time to meet his December deadline. I won't give away too much of the plot here, but Santa has all kinds of help, from robins, talking books, magic trumpets and a small boy called Simon — in fact, even the sleigh manages to help by... Well, I suppose that would be telling. Check the ITV listings at Christmas and see for yourself!
Richard's music won out over the other submissions and he was asked to go ahead with the project. Not having a studio of his own (the test submission had been done on his Korg O1W/FD, using the internal sequencer), it was only a small logical step to his decision to 'keep it in the family' and use my facilities. Since we had never worked together before — musically or otherwise — it was in many ways a leap of faith on Richard's part to ask me to get involved, but I can happily report that we're still talking!
My first task was to determine whether my studio was actually up to Richard's requirements. After a preliminary chat with him I began to hear the distant (750ms pre‑delayed large hall reverb) alarm bells ringing. It was clear that a great deal of instrumentation was required for some of the more complex musical sections, which would undoubtedly leave my sound modules struggling for polyphony — particularly as most of my synths are analogue, and 'electronic' sounds would not be featuring heavily in the production.
Normally I'd get around polyphony problems by committing the most note‑hungry parts to my analogue 8‑track, but as soon as synchronisation to video was brought into the frame, it became clear that my usual working methods would not suffice. There was no way to synchronise video with the analogue 8‑track (at least not without a great deal of hassle) and I was not at all convinced that working with a traditional linear recording system would be the best way to get the job done in any case — waiting for tape rewinds has never been one of my favourite occupations.
I began to look at various methods of increasing my studio's available polyphony, perhaps by buying/hiring extra sound modules, or by using my Akai S1100 to sample some of the more complex musical parts. Eventually I decided that, rather than potentially compromising the smooth running of the recording sessions by resorting to any Heath‑Robinson tactics, I would be wisest to look at upgrading my system to do the job properly and make it better able to cope with similar projects (hopefully) in the future.
Paul Ward (left) and Richard Taylor (right) working on the score in Paul's studio.
Hard Times
Hard disk recording appeared to offer a very efficient solution, but I didn't particularly want more boxes cluttering up my studio, and a hardware approach seemed quite inflexible. Since much of my usual material involves very little 'real' instrumentation, a dedicated hard‑disk recorder would spend much of its time gathering dust. This approach would also fail to address my need for more MIDI outputs to handle the growing number of multitimbral sound modules in my collection. After taking a look at my gallant little Atari STe and staring glumly at my overworked/underpowered 386 PC, I came to the conclusion that the time was right to make the move to a full‑blown integrated MIDI and digital audio PC‑based system. At least I would be utilising the same machine for word‑processing, accounting, Internet access and all the other odds and sods that make a computer earn its keep. Playing the odd game of Formula One Grand Prix wouldn't go amiss either...
Before anyone starts writing to me and waxing lyrical about the delights of the Macintosh, let me state here and now that, although I did consider the possibility, the Mac was never an option I was very likely to take up. It was important to me that I was familiar with the new system in as little time as possible. I also have a great deal of time and money tied up in the PC, in terms of both software and experience. To begin the project by coming to terms with a new operating system was something I just could not justify. Most of my immediate colleagues and contacts are also PC users, giving me a ready supply of knowledge and support in an emergency — a Mac would have left me most conspicuously out on my own! A further problem is one of space. Given that I need to keep a PC around, a Mac would have to share studio space with it — and space in my studio is very much at a premium!
Carefully‑calculated budget in hand, I approached several PC suppliers with a view to buying the best‑specified machine I could afford. I was keen to get a machine which had at least a fighting chance of not becoming obsolete within two days of leaving of the shop. For this reason I thought it prudent to opt for an MMX processor — though it has subsequently been suggested that, for serious audio use, the software houses are highly unlikely to make use of the MMX's extra instruction capabilities, although the slight improvement due to caching differences is probably worth the minimal extra cost.
Making The Choices
After what seemed an eternity of comparing prices and specifications I opted for a Silica Shop Hurricane P200MMX. This machine appeared to exhibit all of the current improvements, including 512K pipeline cache and even a 16x speed CD‑ROM drive. I first upgraded the memory to 48Mb so that the PC would be better able to cope with the stresses of hard disk recording and running multiple applications. Steinberg's Cubase, my sequencer of choice, was duly ordered, installed and checked out with the existing soundcard to ensure that everything was working as it should. I would really have liked to try the new VST version, but I discovered that it would not be ready by the time we needed to start recording. I had to download a 'full‑duplex' driver for the 'cheap and nasty' soundcard that came as standard with the PC, and I also installed my old MIDIQuest ISA MIDI interface, which happily worked without a hitch. Once all this was running as it should, it was time to begin looking at the upgrades necessary to give me my full 'studio in a PC'.
Taking a look at the system interrupt usage gave me something of a shock: I only had one spare interrupt left. The soundcard was actually taking up three interrupts — one for the synth section and one each for the audio in and out! Removing the old card and plugging in a Turtle Beach Fiji certainly made a big difference to both the quality of sound coming from the PC and the interrupt resources available on the system! However, installing the Fiji was not without its problems. Initially, Cubase would only run for 10 or 20 seconds before the whole machine hung, requiring a hard reset to recover. Experiments with other audio playback programs achieved similar results, though for some reason the standard Windows 95 MediaPlayer seemed to work fine. Despite all indications to the contrary, the problem turned out to be a faulty Fiji card. The replacement worked perfectly first time, and (but for a problem with the joystick port which meant I had to install the Fiji in non‑Plug and Play mode) has given no other trouble since.
I bought a digital I/O option with the Fiji to enable me to transfer audio directly to DAT, and this also proved traumatic. For some reason, my Yamaha DTR2 DAT recorder refused point‑blank to recognise the digital output of the soundcard. My Akai S1100 sampler had similar problems, seemingly unable to lock onto the clock signal from the Fiji. My Kurzweil K2000 synth/sampler had no such difficulties, nor (fortunately) did a borrowed Tascam DA30 DAT recorder... I hadn't planned on buying a new DAT, but it now seemed unavoidable, and since I had proved a Tascam machine would work with my other equipment I went for a shiny new DA30 Mk2.
The Fiji digital I/O and the DA30 Mk2 also give me the capability to apply processing to previously recorded material by dumping to and from DAT without having to go through an analogue conversion process. This was to prove invaluable almost immediately when a close friend found that he needed a very old cassette cleaning up for inclusion on a CD compilation. A couple of passes through Steinberg's WaveLab and off went the recording, sounding as good as if it had been made yesterday.
I couldn't resist buying a Yamaha DB50XG daughterboard to provide sound for multimedia applications — not doing so would have seemed like an insolent waste of the Fiji's Waveblaster header socket! I was very impressed by the sounds of the DB50XG and would heartily recommend it to anyone. The depth and quality of the sounds are beyond reproach at the price, and it was far from being merely the add‑on for gaming that I had expected, clearly being capable of augmenting my studio's capabilities.
I added an Adaptec 2940 SCSI card and attached an Iomega Jaz drive to provide backup facilities. The 2940 interface is particularly suited to audio recording by virtue of its inherent high speed and use of 'buss‑mastering' technology, which ties up less of the main PC processor's time. The Jaz provides 1Gb of removable storage space, which would allow me plenty of room to shuffle an hour's worth of stereo audio around at the mastering stage. The advantage of the SCSI buss is that I can add on several storage options, such as CD‑R or a larger, faster AV drive, giving me some degree of future‑proofing. I can also share the Jaz simultaneously with my S1100 or K2000 for sample storage.
Windows did have one or two problems understanding that I didn't really have eight K2000s plugged in at the other end, but this never really posed much of a problem in normal use. The reason for the anomaly is apparently something to do with the way Windows interfaces with the SCSI network, and the K2000's internal hard drive being assigned device number zero. I was pointed in the direction of some new drivers, but they made little difference, so I decided to live with it.
As far as the odd game of F1GP goes... I hadn't realised that DOS‑based games wouldn't be able to make use of the Fiji or DB50XG cards, being incompatible with the usual Soundblaster standard. So I just plugged the old soundcard back in and all the DOS‑based games seemed to find it with no problem (or drivers), despite Windows appearing to know nothing of its existence! After several problems with a new set of Fiji drivers and the Fiji's joystick port I found that the old soundcard's joystick port was found by Windows too — more by luck than judgement. PCs are strange beasts indeed...
I'll spare everyone the story of the 'interesting times' I had setting up other hardware and software, such as my modem and Internet applications. Suffice it to say that I managed to muddle my way through with a little help from my friends (SOS contributors Janet Harniman‑Cook and Martin Walker, take a bow). I consider myself reasonably conversant with PC technology, but I was quite unprepared for the scale and scope of problems I had to deal with to get my new machine into an ideal state for proper use. On more than one occasion a Mac certainly seemed a very tempting alternative! I had every intention of adding a multi‑port MIDI interface, but as soon as I learned of the Windows 95 limit of 11 MIDI devices I decided to hold off for a while. As an alternative, I pressed all the available MIDI outputs from my current setup into service instead — including the COM1 serial port. For the moment this is adequate.
Were the results worth all this effort and expense? On the whole, yes. I overspent my budget (largely as a result of having to buy a new DAT machine), but I'll still turn in a profit, and I've invested in a very flexible recording system, with full backup facilities, that is eminently capable of the work I require of it — at least for the foreseeable future.
Key Questions
With the gear situation well and truly sorted out, it was time to make a start on the project. It began gently enough, with Richard and myself going through the sound sources in my studio and looking for a collection of sounds that would form the backbone of the cartoon arrangements. The K2000 was favoured for its bells, strings, brass and a particularly stunning oboe sample found on a distant corner of the Internet, whilst the S1100 was loaded up with harp, piano, acoustic guitar and percussion sounds, including the obvious sleigh bells! The Yamaha TX802 doubled most of the harp parts, to add extra 'bite' to the sound, and also supplied many of the twangy/tinkly patches that it does so well. Extra 'breathy' bell patches came from the beautifully gritty Kawai K1R. The Roland D550 played many of the horn parts, as well as some of the occasional synth‑type sounds, along with the Korg Wavestation SR. The Wavestation also provided ansuperbly realistic clarinet. Due to the nature of the project, my old analogue synths took a back seat, although my JX8P produced some of the string pads and basses to fill out the bottom end. Richard's Korg O1W/FD was drafted in during mixing to add sounds that we felt could not be improved upon with my own equipment, including harp and vibes.
After searching around for an age for the best pizzicato string patch, it occurred to me to check out some of the sounds available on the Yamaha DB50XG, and we were both very pleasantly suprised by the result! A couple of tweaks to the reverb settings, and the pizzicato strings were in place. The DB50XG's bassoon patch also put in an early and very welcome appearance.
Having compiled an extensive set of basic sounds, I ensured that all of the patch information was backed up into Cubase's Studio Module as a global system exclusive dump. As work progressed and further sounds were brought into play, I updated this dump so that all sounds were available by name in the Studio Module at all times. This allowed us to change between Cubase songs without having to fiddle around finding patches specific to a certain piece of music. The S1100 samples were stored in a single volume on the internal hard drive, which was loaded up at the start of each session by a program change command from Cubase. I similarly defined a boot macro on the K2000's hard drive to get the K2000 into a state of readiness at the start of each session. I always think it's important to let the machines do as much of the work as possible, leaving the humans free to deal with the creative process, relatively unhindered by technical considerations. If I had to hold Richard up for five minutes while I hunted for "that cello sound we used on the chase scene", I felt I wouldn't be doing my job properly.