Features

As you will have gathered from Wes Phillips's comments, in addition to making a traditional two-channel recording sampled at 44.1kHz for release on CD, we also captured a surround-sound version sampled at 96kHz. This was the idea of Steven K. Lee of Canorus Inc., who distributes dCS and Focus Audio products in the US. With the cooperation of dCS Ltd. and Nagra USA, Steve assembled an array of four dCS 902D 96kHz/24-bit A/D converters and three Nagra-D, each fitted with the v2.0 version of the control software, which allows the machine to be switched to operate as a two-channel digital recorder running at twice the usual sampling rate. In effect, the 96kHz-sampled data for each channel is split into two 48kHz data streams.

The balanced preamplifier output for each of the four front-facing microphones was fed to a passive splitter via 1m lengths of AudioTruth Lapis x3. Additional lengths of Lapis fed both the 44.1kHz ADCs and two of the 96kHz ADCs, the latter being synchronized to a single 96kHz SDIF wordclock signal provided by the fourth dCS converter. 100' runs of Canare AES/EBU cable took the digital signals to the back room where the Nagra-D recorders had been set up.

What miking to use for the surround channels took some thought. I had not been very impressed by most of the surround-sound recordings demonstrated by the International Alliance for Multichannel Music at the 1996 AES Convention in Los Angeles last November. With the exception of a Decca recording of a choir and orchestra, the surround channels seemed to have little connection with the front information. Rather than help re-create a coherent sound picture of the original acoustic environment, the rear information tended to hang in clumps localized at the two surround speaker positions.

The surround recordings demonstrated by EgglestonWorks' Peter McGrath at the 1997 Winter Consumer Electronics Show last January were much more convincing. Peter had used his usual Schoeps KFM-6 "Sphere" stereo microphone, but had placed a forward- and backward-facing figure-8 microphone in front of each of the Sphere's sideways-facing omnidirectional capsules. In effect he was using a side-firing M-S array for the left and right halves of the surround soundstage. For playback over four speakers, he dematrixed the two pairs of M-S signals using Audio Engineering Associates decoders.

The important point to note about Peter's technique is that he is basically sampling the live sound at a single point in space, which should give a coherent reproduced soundfield when played back on a surround system. Inspired by Peter's results but not having either a Schoeps Sphere or a pair of figure-8 microphones, I decided to use a second pair of B&K cardioids, again configured as an ORTF pair but mounted beneath the front pair and pointing to the rear of the church. The outputs of these mikes were fed via balanced Cardas cables to a Forssell tube preamplifier, which in turn fed a third dCS 902D 96kHz converter and a third Nagra-D.

As Wes described, a problem with one of the Nagras meant that we could not record the omni mikes at 96kHz. I must admit that, at the time of writing, I have no idea how successful this surround miking has been. While I can listen to one pair of the 96kHz-sampled tracks at a time by feeding two AES/EBU cables from Stereophile's Nagra-D to a dCS Elgar D/A converter, I need a second Nagra-D and a second Elgar to simultaneously audition the four surround channels. I also need to update our Sonic Solutions digital audio workstation to run at the higher sample rate. And, as there is no release medium yet available that can carry multiple channels of 96kHz-sampled audio data, my editing of the surround-sound version of Rhapsody had to take a back seat to getting the conventional CD edited and mastered in time for HI-FI '97. But the 96kHz tapes are sitting there waiting. And, who knows, I might in the meantime put together a DTS-encoded surround CD, downsampling the surround-channel data to 44.1kHz. Watch this space!