Positioning the choir in the new chapel, with the organ behind; we found the best position for the Soundfield mic to be some way above the choir!
Last month, Hugh Robjohns described the planning and setting up he undertook in preparation for recording the choir of Cheltenham College. This month, he turns his attention to the recording itself — and the problems of moving between two venues in one day.
The location recording project I introduced last month was, from the beginning, intended to incorporate recordings made in two different venues — although the same temporary control room, a lecture theatre, would be used in both cases. The plan was to record the choir both unaccompanied in the 'old' chapel of Cheltenham College, and with the organ in the 'new' chapel, with its rather different acoustics. This is the last article in a two‑part series. Read /www.soundonsound.com/sos/sep99/a....
Since we were planning to record in the 'new' chapel in the afternoon, with an organ accompaniment, we had to ensure consistent tuning between recordings made in the two halls, since they would be interposed on the final CD. Consequently, when recording in the 'old' hall, the choirmaster used a portable keyboard tuned to match the pipe organ, to establish the pitch of the choir for the unaccompanied pieces. Although we were planning to do only minimal editing, we knew there could be several takes of the more complex pieces, and keeping the tempo accurate can sometimes be a problem in long works. Fortunately, however, most of these choral pieces were fairly short and the choirmaster was confident of his ability to keep the timing consistent, so we didn't bother with a metronome and had no significant problems in the editing.
It is important to record a couple of minutes of silence at the start of a DAT tape to ensure the music recording is well past the region most prone to dropouts. This was done while we rigged the mics in the hall and we also checked that roughly the same A‑time information (used for the logs) was being recorded on both machines. During a run‑through of the first musical item, we optimised the signal levels at each stage of the chain starting with the Soundfield ST250 stereo mic, then the GML mic preamp (gain‑switched in 5dB steps, leaving plenty of headroom) and, finally, the Yamaha O2R mixer's input. Assuming sufficient headroom at the analogue stages, the crucial element is setting the input to the A‑D converter (which in this particular signal chain is located on the mixer input) for the highest possible output level without risking overload — we aimed to peak between ‑2 and ‑6dBFS.
With this kind of music, it is very useful as a recording engineer to be able to follow the score: in this context, the benefit is mainly in identifying cues for re‑takes and edit points, but in more complex multi‑mic rigs, it is important to know when different sections of the choir or orchestra are performing, which is when the score becomes vital. On this occasion, we had the luxury of a 'producer' to follow the score, look after the cue lights and make the recording logs for us, but this is normally a very important job for the location recording engineer. When you are doing it all yourself, the challenge is to listen critically at the same time as taking the notes and following the dots!
I find that once the first recording is in the can, it is well worth getting the musical director in to the control room for a replay — just to make sure the balance and 'acoustic style' match the expectations — so we invited the MD in. Half a minute later he arrived... along with the entire choir, which was not quite what we expected, and definitely not something to be encouraged in normal circumstances! However, since we were using a lecture theatre as our control room, everyone was able to find a seat and, thanks to the raking, they could all hear a reasonable stereo image. The little PMC monitors were perfectly able to fill the room with a scale and quality of sound which compared very well to the original, and the beaming smiles of our audience confirmed that we had hit the mark with this first track of the day. In fact, they sang with much greater confidence and enthusiasm from that point on, and we had to turn the GML mic preamp inputs down by 5dB!
The New Chapel
The scaled‑down, mixerless control‑room setup used to record the additional sessions (see text).
The rest of the morning's recordings went extremely well, and we were able to break for lunch on schedule having achieved acceptable recordings of all the required songs. While the choir relaxed for an hour and enjoyed the school's inimitable version of a traditional Sunday roast, we de‑rigged in the old chapel, ran out even more multicore, and re‑rigged the mics, talkback and cue lights in the new chapel.
The unaccompanied recordings made in the old chapel had used a single Soundfield ST250 stereo mic, a few metres in front of and above the choir (see last month); a satisfactory mix of source sound and room ambience was achieved simply by positioning the mic and adjusting its polar pattern. In the new chapel, things were more complicated, as we had the additional factor of the organ to consider. Our plan had been to combine the Soundfield mic (and outriggers) on the choir with a separate spaced stereo pair for the organ, which is mounted high up at the rear of the chapel. The only practical place for the organ mics was on a balcony directly in front of the display pipework. However, the stalls for the choir were quite a distance from the organ, right in the middle of the long axis of the chapel and we were concerned that timing differences between the mics could cause all sorts of coloration and phasing problems. Consequently, we decided to arrange the choir in the stalls at the entrance to the chapel, directly in front of the balcony, and rigged the mics in suitable positions. After another stereo lineup to recalibrate the channel gains, we were able to monitor the organist practising and it became apparent that our rough positioning of the Soundfield mic was capturing the organ with a lovely perspective. With the mic's polar pattern set to crossed figure‑of‑eights, the balance of the direct organ sound to its reverberation was almost perfect, which was a reassuring start! As a result, we dispersed with the separate stereo pair idea.
I find that once the first recording is in the can, it is well worth getting the musical director in to the control room for a replay...
After lunch the choristers took up positions on each side of the stalls facing inwards, as they would for the normal services held in the chapel. However, this was not exactly an ideal arrangement for recording with a stereo pair at one end, and they were reorganised to form a wedge shape in front of the Soundfield mic. The idea was to have them almost fill the 90‑degree front acceptance angle of the (crossed figure‑eights) mic so that the stereo image matched that of the morning's recordings, with the organ positioned centrally behind them. The rear pickup angle of the 'eights' was capturing the reverberation in the room, to balance the relatively close frontal pickup of the choir.
During the warmups and rehearsals we moved some of the choir soloists around to optimise the stereo imaging, and played with the height and relative distance of the Soundfield mic, to find the desired balance between choir, organ and hall acoustic. We eventually settled on the figure‑of‑eight pattern, with the mic about three metres in front and three and a half metres above the choir. Once again, the Soundfield was delivering a fabulous sound all on its own, so there was no need for separate organ mics.
The afternoon session went fairly smoothly, apart from having to do a couple of retakes because of ambulance sirens and traffic noises! However, it was a long and tiring day for the choir and their performance began to fall off towards the end of the end of the afternoon. There were also a couple of extremely heavy downpours, which imposed a lot of background noise on a couple of tracks, which we put up with at the time because the performances were so good.
By the end of the day we had well over an hour of material in the can, but after the MD had auditioned all the recordings and discarded a couple of the weaker tracks, we were left with only 47 minutes of edited music. Since the aim was to provide a full CD, a second recording session was organised a few weeks later, again in the new chapel. However, the only time available for this extra session was in the evening after school, and the lecture theatre we had used as the original control room was unavailable. A small office behind the chapel was chosen as the new control room, and although we had to match the recording setup exactly, so the new material would fit in with the other music on the CD, we wanted to simplify the control room setup to make the process of rigging and de‑rigging faster and easier.
We dispensed with the backup DAT — a risky option, but at least we could call on the CD‑R if the DAT suffered a problem — and the Yamaha desk, which was not needed just to handle two mic channels! The Soundfield mic was connected to the GML mic pre as before and patched directly into the Sony PCM7050 DAT machine, using its own excellent 16‑bit A‑D converters. The CD‑R recorded the signal from the DAT's AES‑EBU output, while the DAT's analogue outputs fed my stereo metering box. The analogue outputs of the CD‑R were connected via a passive level control to the PMC monitoring, which made playbacks fast and easy. A switched mic in the control room was connected to a spare channel of the GML in the chapel, and then to a powered speaker for talkback. All in all, a small and very convenient rig — but only made possible because we already knew exactly what was required to match the original recordings.
The recording log is a vital element of any recording session, particularly if the post‑production is delayed.
Post Production
Co‑engineer Paul Hedges adjusts the Soundfield mic.
The additional two‑hour session was more than sufficient to record all the extra material we needed and, after some nifty editing on a SADiE system, we ended up with 72 minutes and 22 eclectic tracks of lovely choral music. The two tracks marred by the noise of thunderous rain of the chapel roof were retained and salvaged, courtesy of the fabulous CEDAR de‑hiss system. This was operated by the company's resident engineer, who managed to reduce and re‑shape the noise spectrum to match the ambient noise of the other tracks of the CD — a task done so well that the CEDAR tracks are virtually indistinguishable.
After agreeing the edits and a track running order with the MD we produced a final master on CD‑R, complete with all the track start IDs in the right place, for the CD pressing plant. Producing the artwork for a 12‑page booklet, tray card, and the CD label was as much of a challenge as recording the music in the first place, but eventually the whole lot was approved and dispatched to the pressing plant a couple of weeks before the end of the term. A week or so later five large boxes of finished CDs, all sealed in cellophane, were delivered to an ecstatic musical director, 40 happy choristers and, hopefully, a lot of satisfied new choral CD owners!
Recording Log Sheets
Cheltenham College, showing the old (left) and new chapels.The recording log is a vital element of any recording session, particularly if the post‑production is delayed. It is easy to make log sheets on any word processor; at the top of my sheets I have a table with spaces for the project, names of personnel, dates, locations, sample rates, bit resolution, dither type, and a diagram of the instrument and mic positions. It is also a good idea to leave space for any other notes that you might want to make about the session.
In the main body of the sheet I provide a separate row for each take, with spaces marked out for the title of the track, the take number, the start and end times, the Start ID and CD‑R track number, recording status, and a space for more notes. The 'status' column is used for shorthand notes about the recording: for example, a complete usable take would be marked as 'OK', a take which had a false start would be 'FS', and one which collapsed part‑way through would be 'BD' (for breakdown).
A recording which could not be used would be 'NG' (no good) and one which would form all or part of the master take would be 'M' — usually made more obvious by drawing a circle around the M. Where there is to be an edit to join two or more 'M's together, I would write something like 'M insert bars 23‑37 T5', which translates as 'this is the master take but with an insert of bars 23 to 37 from take 5', or 'M to bar 12 then T4'. All quite logical really, but fast to write and extremely helpful when you come to post‑produce it all weeks later. I also try to write notes with each take to explain why it was recorded or stopped — again, it makes all the difference when no one can remember why take 6 sounds exactly the same as take 5!
The same basic system works just as well for other kinds of music other than classical works with full scores to refer to — just replace the bar numbers with things like verse 2, chorus 3, intro, bridge and so forth.
