Two Days in August: Stereophile's First Jazz Recording Multitrack Madness

Multitrack Madness
Recording Rendezvous, Stereophile's first nonclassical album at the deconsecrated downtown church Chad Kassem has transformed into Blue Heaven Studios was a challenge. It's a great-sounding space, but I was concerned that a purist microphone technique would result in balance problems with the particular ensemble Jerome Harris had chosen for his compositions. While the coincident-stereo microphone technique I've used for previous Stereophile classical recordings would faithfully capture the delicious acoustic of what everyone was starting to call the "First Church of Chad," both vibraphone and acoustic bass guitar would be overpowered by the drums and horns.

A paradigm shift was called for. Accordingly, the kick drum, snare drum, and bass-guitar speaker were all close-miked, as were Steve Nelson's Müsser vibes (these with a stereo ORTF pair of Shure SM-81 capacitor cardioids). A direct-injected feed was also taken from the bass-guitar preamp. A second ORTF pair of cardioids—B&K 4011s—was mounted high above the drums to capture both a good stereo picture and a sense of the church's ambience. The sax and trombone were miked with one B&K 4006 omni each, which could cope with the tremendous dynamic range of these acoustic instruments. Balanced mike cables used were from Cardas, Canare, and Beyerdynamic.

While close-miking and multitrack recording work best when you have complete acoustic isolation between the instruments, this turned out to be impossible, except for the snare and kick drums and the bass guitar's direct feed. Despite my best intentions to move the musicians as far apart as possible and isolate them with acoustic screens, they needed to hear each other to get a good ensemble. And the horn players also needed to get visual cues from Steve Nelson's mallets. This turned out to be a blessing in disguise: While there were no ambience mikes as such, leakage of the drums and horns into the vibraphone mikes and of the bass guitar cabinet into the drum mikes added natural reverberation that I could make use of when it came time for the mixdown.

The one DI and nine microphone feeds were all transformed to digital at a 44.1kHz sample rate using word-clock-synchronized dCS (vibes, drum overheads), Manley (spot drum mikes), and Nagra (horns and bass guitar) A/D converters, all operating at 20-bit resolution. The digital data were stored on a 4-track Nagra-D open-reel recorder and a Tascam DA-38 modular digital multitrack machine. The latter usually stores eight tracks at 16-bit resolution on Hi-8 video cassettes, but was converted to record six tracks at 20-bit resolution using a PrismSound MR-2024T "bit splitter." In order to maximize sonic purity, no mixing console was used—just low-noise, solid-state mike preamplifiers from Millennia Media and Bryston, placed out in the hall with 1m balanced AudioQuest interconnects connecting their outputs to the A/D converters. The AES/EBU links used to take the digital data to the control room at the back of the church were 50' lengths of Apogee Wyde-Eye and Canare 110 cables.

The fact that the Nagra and Tascam were not synchronized meant that playback of the multitrack data at the sessions was not possible. Accordingly, we used a small Mackie mixer at the sessions to produce a stereo feed for a Panasonic 3700 DAT recorder.

On my return to Santa Fe, I uploaded the 10 tracks of 20-bit digital data (some 26 gigabytes' worth) to a Sonic Solutions eight-channel digital audio workstation. As each track was recorded with the same word-sample clock and could be independently moved in time on the Sonic with respect to the others, I could align them all for synchronized playback. My plan was to do all the subsequent editing and mixing in the digital domain.

With four of the seven songs used in their entirety, choosing the takes for the CD was a relatively straightforward task. The mixdown, however, was more complicated. I had 10 tracks of data, but only an eight-track virtual mixing console on the computer screen. While I could assign two tracks to the same fader, this was hardly an elegant solution. In addition, the vibe tracks sounded too dry, and the trombone track had such an enormous crest factor that accommodating it within the CD's 16-bit dynamic-range window and balancing it with the other instruments depressed the overall level to an unacceptable extent.