Features

Before the first concert, I'd done what I always do: speak into each microphone capsule in turn so that we could check that the left channel was indeed the left, the right channel was indeed the right, and that they were indeed different from one another. Apparently, however, a capacitor in the Manley's pattern-switching circuit had failed early in the first concert, selecting that most useless of stereo microphone techniques: a pair of coincident omnidirectionals. Two-channel mono. Sodding mono.

Although we had a pair of Wilson WATTs/Puppies set up in the control room, the acoustics were so poor and the conditions so tense that all they told us was that we were capturing something on tape—not whether it was what we were actually expecting. Yes, we did hear the piano; yes, we did hear the crackling cellophane; no, we didn't hear that it was not exactly stereo. Mono.

It took some weeks before I was able to bring myself to listen to the Revox tapes; before I did so, I fed the two channels into the 'scope. This time, there was the familiar circular, ball-of-wool trace indicating a stereo source. Reassured, I listened. Stereo. And not just stereo, but a magnificent piano sound. A diffuse soundstage, yes; but the sound of the Steinway in the Albuquerque hall was deliciously accurate. A shiver went down my spine as I listened to the shimmering arpeggios cascading down behind the big hymn tune in the third Chopin Scherzo. I was back in the church.

David Smith of Sony Classics, a man as fascinated as I am by the art of recording the piano, once told me that the trick to getting musically natural, reverberant piano recordings was to arrange for that reverberation to sound warm. Probably by accident more than design, the reverberation captured by the B&Ks was warm—no cold, "bathroomy" sound here.

And, because the cellophane-crinkling woman was quite a distance from the omni mikes (footnote 3), you can't hear her crackling! Yes, you can hear the occasional intake of breath from Bob—you might mistake it for tape-modulation noise, as it happens most at big climaxes—and after a while you become aware of a pleasant baritone crooning emanating from our artiste. But this is part of what makes live recordings sound real. I can't stand sterile classical recordings that sound as if they were made by MIDI-programmed robots; I want my musicians to breathe, even sing if they can't control the emotion of the moment. Yes—we could put out an excellent-sounding two-CD set from the Revox tapes. There were a couple of passages that weren't quite right at either concert, but since the purpose of the album was to re-create the live experience, Bob agreed that it was worth releasing anyway.

It's All Down to Dither
For the editing, the Revox tapes were converted to 20-bit digital with the Manley A/D converter, the data being stored on the 2.6 gigabyte hard-disk drive that came with our Sonic Solutions editing system. All the editing and master assembly was done preserving the 20-bit resolution, not so much because we were sure that the final four of the Manley's 20 bits were valid—and the analog tape hiss was well above the 16-bit LSB level—but because this meant that any artifacts resulting from the digital manipulation would remain below the CD's 16-bit noise floor.

Fig.2 Meridian 618, Type B noise-shaping curve (10dB/vertical div., linear frequency scale).

Once we had assembled a master edit list for each CD, we had to decide how to transfer that 20-bit data to a 16-bit master. Simply dumping the output of the Sonic Solutions to DAT, thereby truncating each digital word from 20 to 16 bits, reduced the air around the instrument. This was easily audible on the soundstage-mapping track mentioned above. Then Meridian's Bob Stuart serendipitously offered to loan us an early sample of the Meridian 618 Mastering Converter (later marketed as the Meridian 518).

The 618 uses a single Motorola DSP56001 digital signal processing chip (the 518 uses a DSP56002) to manipulate digital data on the fly. The input and output word lengths can be independently selected; pre-emphasis can be added if the operator so wishes; and, most important, the 618 redithers the data with a choice of noise-shaping curves. By shifting quantizing noise up to the inaudible 20kHz region as it reduces the output word length (fig.2), it preserves as much as possible of the original's resolution in the midrange.

The interesting thing is that our starting point for the mastering process was an analog original with a noise floor not much more than 60dB down from the peak signal levels. Any changes occurring in the digital domain would be well below the recording's floor of analog tape hiss. Yet, using the Meridian's noise-shaping algorithm when we downloaded the master to CD-R gave a noticeable improvement in air, space, and overall palpability.

1996
"You know," said Bob Silverman as we finished listening to the tapes, "I wouldn't mind having a crack at the Liszt Sonata for my next Stereophile recording." Who were we to disagree with a musician who already has one Grand Prix du Disque for his performances of Liszt's piano music?

Oh, and the cellophane-scrunching lady. We asked her in the interval if she would mind not crumpling her wrappers. Outraged at our temerity, she stormed from the hall, never to return. (Robert Silverman sent her a check for the ticket price in gratitude.) For the Liszt recording, however, we have foregone the audience. Madame Fate had a heap of fun with us on this project. Why tempt her a second time?—John Atkinson

---

Footnote 3: Because their uniform pickup pattern emphasizes the contribution of ambience, you have to get in closer to the sound source with omnidirectional mikes than with other types to get the optimum balance between the direct and reverberant soundfields. This will be illustrated on Stereophile's Test CD 3, due for release in early 1995.—JA