One by One: the Cantus Recording Project Capturing Cantus

Capturing Cantus: John Atkinson

Cantus Artistic Director Erick Lichte was very clear about what he wanted: "A clear, focused sound, not a reverberant, cathedral sound. That would not be appropriate for the music." We were talking in the lobby of the Nassau Presbyterian Church in Princeton, New Jersey, after a astonishingly profound concert of unaccompanied choral music. Erick, a Stereophile reader and an audiophile who's served time at a Minneapolis specialty audio retailer, was floating the idea of me recording a CD for Cantus' eponymous label.

Five months later, I was listening to the choir rehearse in a small, modern hall in Northfield, Minnesota. The sound of the singers was indeed clear and bright, focused by an acoustic "cloud" that hung over the stage. The singers were arranged in an arc halfway between the front of the stage and the rear. To get a recorded sound that was true to the sound in the hall but that also benefited the music, a selection of traditional songs from around the was I going to mike them?

I opened my ever-expanding chest of microphones and took out the pair of B&K 4006 omnis that have provided the basic pickup for almost every Stereophile CD since 1994's Concert (STPH005-2). (The exceptions were the multimiked Rendezvous, STPH013-2, for which I'd used the B&Ks on the sax and trombone; and Rhapsody, STPH010-2, for which Millennia Media's John La Grou had loaned me a pair of 4003s, the high-voltage equivalents.) The 4006s have proved over the years to be superbly true to midrange tonalities, and their extended low frequencies are unrivaled by directional microphones.

In the past, I've favored using these mikes side-on to the soundsources, fitting their diffuse-field grilles to give a flat response. However, as this would have meant that the capsules, with their boosted on-axis top-octave response, would have been pointing straight at the reflective cloud and accentuating the reflected sound, I instead pointed the mikes at the center of the stage and used their flat-on-axis grilles. The omnis were spaced just over 6' apart on the front of the stage, with the capsules 9' high and 13' from the singers. This is too wide a spacing in absolute terms to give anything but a rough semblance of a recorded soundstage, but as usual I wasn't going to use just the omnis; I would be mixing their outputs in post-production with the outputs of a pair of coincident mikes.

This was a pair of Neumann M147 cardioids, arranged as an almost-coincident ORTF pair at the same height as the B&Ks—but, after some experimentation, 10' farther back from the edge of the stage. This was quite a bit farther back than the theoretically appropriate position on an arc based on the center of the soundstage, but what I was listening for was a similar balance between the direct and reflected sounds picked up by the two pairs of mikes. (Being directional, a cardioid mike gives an inherently drier sound when used at the same distance as an omni.) The tubed M147 is a re-creation of Neumann's classic U47. With a 1" capsule, its self-generated noise is low enough that it is usable at the 25' distance I'd settled on, while its on-axis response has a small rise in the presence region that would give good "reach."

I set up my gear in a small anteroom and tried some test recordings, but my intention of recording all four microphone channels at 88.2kHz with 24-bit resolution was thwarted—one of my dCS A/D converters (the older 902 I'd used on Duet, STPH012-2) stubbornly refused to output right-channel data. Working backward from the recorder, I checked cables and connections. Nothing. Examining the 902's AES/EBU output stream with the dCS 972 format converter I'd intended to use to produce DAT backup tapes indicated that there were no right-channel data. I e-mailed dCS's Robert Kelly in the UK, describing the symptoms. "It's terminal," he said. "You'll have to send the unit to us."