A Day in the Life of a MIDI Studio
Ever wondered what goes on in a commercial 16-track studio once the red light comes on? Roger Jackson of West Coast Music reveals all...
Ever wondered what goes on inside a commercial 16-track studio? Roger Jackson of West Coast Music lets you in on a few secrets...
Back in early 1984, we made the momentous decision to introduce MIDI into our studio on the Isle of Skye - it was a choice between MIDI or a new tape machine and we decided to gamble on the new technology. During the ensuing three years, we have seen MIDI progress from controlling keyboards, to live drums, to effects switching and now to the mixdown, and although the old tape machine is still in residence, it's days must surely be numbered.
No two studio sessions are exactly the same, but the following account is typical of our working day and may help bands entering a MIDI studio for the first time, as well as provide an insight for other studios into our sometimes unorthodox approach to recording.
9.00am Maintenance. Our engineer cleans the tape heads and dusts everything with a vacuum cleaner and paint brush. Dust in electronics is bad enough, but cigarette tar is even worse and so the studio is a no-smoking area. I shovel out the empty beer cans from the previous session and make a note to get a bigger dustbin.
The diary shows a first time, two-day booking for a Glasgow band. With all new bands, we suggest that they send a rough cassette - either from a gig or practice - of the tracks to be recorded, but this seems to have been lost in the post. Our food questionnaire reveals that there are no vegetarians in the band, that the drummer doesn't eat tomatoes, but that they all like chips. Good.
10.00am The band arrives. They have already received our brochure and demo tape before making the booking, but they are still surprised at the absence of a separate control room, or studio room, whichever way you look at it. In other words, everything takes place in one large open space.
The band consists of guitar, bass, keyboards, drums and vocalist and their previous recording experience is one session in a 16-track studio. Over coffee, I explain our basic recording philosophy - everything DI (direct injected) into the mixing desk and an electronic kit recorded on computer via MIDI. The reasons for this are improved noise performance and better isolation between sound sources. I point out that once a microphone is placed in front of an acoustic drum and the sound is processed through a mixer, the drum becomes electronic anyway. Similarly, the guitarist is happy to plug into a stage amp, so why not into the desk, which is simply a much better amp than he could ever afford.
The band are not entirely convinced and so I suggest that if they are not happy with our methods by the end of the session, they won't have to pay. This convinces them.
10.30am The instruments are tuned, plugged into the desk and the band run through the number to be recorded, with the drummer playing our electronic kit MIDI'd to a Roland TR707 drum machine and the keyboard player using the Yamaha DX21. All the sound is relayed through the studio monitors at a comfortable listening level. Everything sounds good when it's loud, but if it sounds good when quiet, then it really is good. This track sounds terrible! The problem is not the song itself, which is a strong disco-based number, but the band's timing which is very sloppy.
We start the recording process by establishing a tempo with a basic drum pattern on the trusty old Drumatix (our system runs both MIDI and Sync 24) and by recording FSK code onto track 8 of our 8-track tape [the studio has since taken delivery of a Fostex B16 tape recorder -Ed], The band run through the song again, using the Drumatix timing track in place of the drummer, and a rough vocal line is recorded live from the control room onto tape track 1. This track is to act simply as a guide for the other musicians and to save the singer's voice whilst the recording is pieced together. The timing is already improved and the bass, which is playing eight to the bar throughout, is the first proper track to be recorded.
"A bright, clean bass sound is selected, remembering that it is easier to tone down tape tracks than to sharpen them."
A bright, clean bass sound is selected, remembering that it is easier to tone down tape tracks than to sharpen them. Compression is used to give the bass more power and a noise gate is patched into the audio chain to keep the silences clean - a little thought at this stage can save the engineer a lot of time later. The bass is recorded onto track 2, with the monitors relaying only the voice, the Drumatix and the bass itself. In this way, the bass must take its timing from the drums and will not be confused by any other instruments. In this particular song, the backing guitar is also strict tempo and so it's recorded onto track 3, using the same technique as was used for the bass. We now have a fairly tight rhythm section.
11.30am More coffee arrives, the vocalist and the drummer decide to take out the rowing boat (one of the perks of our seaside location) and we turn our attention to the keyboards.
The song has a simple structure with an intro, three verses of 16 bars each, a 16-bar instrumental break, and a fade to finish. The keyboard player has found a nice 'Fender Rhodes' piano sound (remember those?) on our DX9, which he thinks will fit in well with the backing track and so the 8-bar intro is recorded as the first computer sequence, or block. (We use a Steinberg Pro-16/Commodore 64 sequencing system.) We always reserve MIDI channel 1 for recording only, and re-assign channels on playback. This means that we only hear what we choose to hear.
The verse is recorded as sequence two and this is then overdubbed with a 'Prophet 5' voice from the Korg Poly 800. The sequence is then copied several times to form the other verses, the instrumental break, and an extra verse for the fade. Sparing bits of synthesized vibes are added to verse two, some brass stabs to verse three and some synth-sax to the fade, all using the DX21. Each verse has a DX21 patch change loaded into a separate computer track to select the correct voice and a further patch change is loaded into the instrumental section to change the Poly 800 to an analogue string sound, whilst retaining the original notes. The sequences are pasted up in the correct running order and we have the makings of a good song.
1.00pm Lunch arrives and it is discovered that the boating half of the band have lost an oar and are disappearing on an off-shore wind. The other boat is launched in a daring rescue bid.
2.00pm With the excitement of lunch and the rescue behind us, we listen to the results so far. The band is happy, but both the engineer and I think that the bass is still sloppy, despite our previous efforts. To demonstrate our point, I quickly programme a step-time bass on one of the micro-composers and the engineer substitutes this for the original tape track. The difference is spectacular and the band decide to keep the synth bass. We improve things even further by adding a fat synth riff on the other micro-composer in harmony with the bass line, but now the backing guitar is starting to sound loose and is detracting from the otherwise tight sound. The guitar is also dropped.
"Choral backing voices are needed to reinforce the strings... and the Vesta sampler/delay is chosen as the quickest and easiest method of achieving this."
Choral backing voices are needed to reinforce the strings during the instrumental break and the Vesta sampler/delay is chosen as the quickest and easiest method of achieving this. The machine is set to sample for one second and then to immediately overdub. A five-second note from the vocalist produces a one second sample overdubbed four times and this is overdubbed again an octave higher, giving a 10-voice effect. The result is looped and replayed from the keyboard via a MIDI/CV convertor. The band get excited.
3.00pm Even more coffee arrives. The Drumatix is removed from the mix and the band's drummer plays through the song on the full electronic kit, which is recorded directly onto the computer. It sounds awful! An hour later, it sounds even worse and the drummer wishes he'd gone down with the boat earlier in the day!
For us, the problem is a common one. In most situations, a band will take its tempo from the drummer and most musicians become practiced at this, whether consciously or subconsciously. In the studio, it's not difficult for the band to switch to playing along with a timing track, but for the drummer it's a whole new world.
Whilst the drummer goes to lie down in a darkened room, our engineer listens to the live drum track and duplicates the bass drum and hi-hats, in step-time, on the TR707. Then, without waiting for me to suggest it (I must be getting predictable), he uses the Drumatix and TR707 together to put parts of the hi-hats into stereo, bouncing alternate beats left and right - a trick we've been using longer than Stevie Nicks. The drummer is recalled and placed in front of our Roland Octopad, which is triggering the sounds in a TR505. The song is replayed with the step-time bass drum and hi-hats and we suggest that the drummer experiments with just snare and toms.
Ten minutes later, his confidence has returned and we realise that he's actually an excellent musician. Being freed from the responsibility of keeping tempo, the drummer's breaks and fills are superb and the whole song is given a lift by the inventiveness of the live drums. I signal to the engineer, who records the next run through and we eventually have to drag the drummer away from the Octopad. We listen to the playback and even our engineer gets excited.
5.00pm Yet more coffee arrives, with sandwiches. We run a rough mono mix onto track 6 of the tape, to serve as a monitor mix whilst the final tape tracks are recorded. This saves having to start at the beginning each time and is cheaper than going for SMPTE. The guitar records easily on the third take and it's the break we needed.
Up to this stage, none of the band has had to wear headphones and we've even developed a system which frees the vocalist from this unnatural contraption. Whilst the singer looks puzzled, the engineer and I damp down the wall of the control room opposite the monitors, using drapes and acoustic screens. A directional microphone is placed in the centre of the room, with the singer facing the monitors. The sound travelling from the monitors is ignored by the microphone and as there is no reflected sound from behind, overspill is minimal. It sounds hard to believe, I know, but it actually works. And because the singer can perform in a much more natural environment, almost as though he was on stage, the results are much better than with conventional recording methods.
"...he uses the Drumatix and TR707 together to put parts of the hi-hats into stereo, bouncing alternate beats left and right..."
In this instance, the singer is delighted with the new arrangement and, as he is already hyped-up from the day's recording, we have no difficulty in capturing enough excitement on the vocal track. Finally, a short repetitive vocal line is sampled into the Akai S612 and the computer is instructed to drop this in, in the required places. By this time, even I'm getting excited.
6.00pm Cans of beer arrive and we start the mix. Our first task is deciding which instruments to route through the new Simmons SPM8:2 MIDI mixer, which is on loan to the studio for evaluation. For those of you who missed the review of this device [SOS April '87], it's an 8-channel rack-mount mixer which permits 64 complete front panel settings to be selected via MIDI, and, by means of a programmable cross-fade, also allows a form of dynamic automation. We select the DX21 for level changes, the Poly 800 for EQ changes, and the Vesta sampler/delay and second micro-composer for auto-panning (another of the Simmons' handy features). Running through the song several times, we arrive at six different settings of the Simmons mixer and programme the patch changes into the computer. We continue to play the song, with the Simmons taking charge of the hands-on parts, and begin to adjust the other tracks.
The 'Fender Rhodes' patch, which sounded so good on its own, is now getting lost in the mix and something sharper is needed. The keyboard player manually selects various DX9 voices during playback and, suddenly, a beautiful electric harpsichord cuts through the mix. Now that the DX9 can be heard clearly, we notice that the timing would benefit from quantising and this is attended to.
The drummer now feels that the snare sound is wrong, so by patching a Yamaha RX21 drum machine into the system on the same receive channel as the Roland TR505 and juggling with MIDI note numbers, the snare is re-directed. The final drum arrangement is as follows:
Bass Drum - TR707
Hi-Hats - TR707 + TR606
Snare - RX21
Toms + timbale - TR505
Cymbals - RX21 + TR505
We now introduce several reverb units into the mix, but with so many 'live tracks' there are never enough devices to go round. We economise by using the same MIDIVERB for vocals and lead guitar. As these two tracks never sound at the same time, we use a different program setting for each one and instruct the Steinberg sequencer to switch between them, again using MIDI.
There are now 26 desk channels in use, of which only two are tape tracks and so the mix should be as good as any digital recording. It does not, however, have that exaggerated top end which all digital recordings seem to possess and which has become so fashionable - so we add the necessary 'sparkle' with an Aural Exciter across the main outputs.
Finally, the mix is fed through a pair of noise gates to maintain total silence on the lead-in and is digitally mastered onto a Sony PCM701. It sounds great.
8.00pm Ten hours after we started, we have three and a half minutes of recorded music. Four years ago, it would have taken twice as long and sounded half as good. We adjourn to the pub for food and drink and the self-congratulation that always follows a good session. The singer hands me a pint of Guinness and asks how they can reproduce the sound live. I pretend not to hear.
Contact West Coast Music, (Contact Details).
Feature by Roger Jackson
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!