Ludwig van Beethoven's 32 Piano Sonatas Page 4

At sessions, I always record a few minutes of the room sound, to splice between individual movements so that the music doesn't fade disconcertingly to black. For interest's sake, fig.1 shows an FFT-derived spectrum of the 24-bit room sound. You can see that almost all the lower-frequency FFT "bins" are below -96dBFS, with all the noise components above 2kHz below -120dB.

Fig.1 FFT-derived spectrum of the background silence of the Maestro Foundation recital hall. Sample rate of original data, 88.2kHz; word length, 24 bits (12dB/vertical division).

However, you can also see two LF components at exactly 120Hz and 240Hz—power-supply frequencies—around 76dB down from peak level. Before I started taping, I went crazy trying to trace the source of these hum components. Because of their frequencies, I assumed they were electrical in nature. But when I recorded pure tones from a portable Neutrik signal generator feeding the mike preamps, the hum was absent.

It turned out that these tones were acoustic. They were due to mechanical vibrations from the Bösendorfer's actuator power supply, which was housed in the large tray under the piano keyboard. Removing the lacquered cover, which I suspected was amplifying the sound, and damping it with, say, Blu-Tack, would have been a step toward a solution. But if it wasn't possible to move the piano, it was hardly going to be possible to apply damping gunk to its innards.

We decided to live with the hum. I did try filtering out the 120Hz component during the mastering, using a narrow digital notch filter realized with the Z-Systems rdp-1 digital processor, but felt in the end that the sound quality was better with the acoustic hum intact. It was part of the piano's sound, after all. A purist, me!

Satisfied that the mikes and their positions would be serving the music, we started taping—but not before Murphy managed to strike. Packing of my wife's Land Cruiser the previous morning for the 1000-mile drive, I had forgotten in the kerfuffle to bring a DAC! We therefore monitored the recording from the balanced analog outputs of the Panasonic DAT recorder we were using as a backup. Not very elegant, I admit, but in defeating Murphy, elegance is not necessary. And we did have that Proceed AVP handy for speaker playback.

Nine months later...
I left Santa Monica the following Monday as planned, with 18 hours—78 gigabytes—of four-channel, 24-bit/88.2kHz open-reel and 8mm session tapes. Using a dCS 972, these tapes were downsampled to 39GB of four-channel, 24/44.1 hard-disk data, which in turn—using the Sonic Solutions Sonic Studio v.5.2 and my trusty Meridian 518—was edited, mixed, mastered, and noiseshaped to 10 CDs' worth of two-channel 16/44.1 music data.

As I write these words I am listening to the CD set, which I have just received from Bob Silverman. Yes, it is obviously a large piano in a small hall. But those early reflections help define the recording space without getting in the way of the music, and the dynamics are so wide, the sound so natural, the pianism so masterful, that I keep stopping pecking at the keyboard to listen, enthralled.

And oh, the music. I leave it to the record critics to have their way with Bob's performances and my sound. But for me, to have been able to contribute to one musician's lifetime of performing on the piano and one composer's lifetime of writing for the piano, both realized in this set, has been and will remain a highlight of my work as a recording engineer.

Robert Silverman's 10-CD cycle of Beethoven's piano sonatas costs US $65.00. You can order it from our secure "Recordings page.