Interviews

Boyk: Absolutely. The analog machine also makes some difference, and although one can say until blue in the face that the digital machine made a bigger difference, you can't quantify that. If you could, it would be all done, we wouldn't need to worry about this stuff. So the only criterion that one can meaningfully use is to examine which comes closest to the direct feed in allowing the listener to get into the emotion of the music. It is the only conceivable criterion, because that is the purpose of what you're doing; that's what music is for. To use any other criterion is intellectually dishonest, and pretends to an objectivity that, in fact, it cannot have.

Each individual has to make the assessment for himself on the basis of "Which one lets me forget about the technology the most?" And it will always be the case that a good direct feed will go farther in this direction than either analog or digital. But I'll tell you what does come the closest: it is the direct-to-disc process. When we recorded the Los Angeles Philharmonic, Doug Sax and I were in the control room all day listening to the direct feed, which, if I do say so myself, was gorgeous. It was really beautiful.

We had playback from five storage media: the direct-to-disc lacquer (an occasional lacquer that gets spoiled for some reason and doesn't get sent for metal-plating); the digital tape recorder (Sheffield's own highly-modified, very much improved machine); and three analog tape recorders. The machines were not in the control room, so this was an informal blind test with levels matched, polarity correct. The only medium that really stood up to the beauty of the direct feed—and it did so to an amazing degree—was the direct-to-disc lacquer. It was quite stunning. When you hear a lacquer, it's dead quiet; it's just been made, and the basic noise level is microscopic, especially when it's direct-to-disc and there's no tape hiss. It was close enough that I think that you could have taken a professional engineer in there, told him it was the direct feed, and he would have been more than satisfied that that was the sound we were getting from the mikes.

The next group down were the three analog tape recorders. They were all different machines—some experimentation was going on here—and I had my preferences among them, but all did honorably with the program material. Then, in a class by itself at the bottom, was the digital machine. Nobody could stand to listen to it, to the extent that one person in the control room spontaneously said, "Stop the comparison. Whatever that machine is, it's broken!" Doug said "No it's not broken; that's the digital!"

Now here's what's interesting: only one of those three media is pure analog, and that's the direct-to-disc. What we conventionally call analog tape is actually a sampled medium. The signal rides on the bias; in effect, the bias is turning the sampling on and off. This is at a very high rate, to be sure, and the turn-on/turn-off is very soft because it's a sinewave, but it's a sampled medium, nonetheless. So we have the true analog medium at the top (according to yours truly), then the "sampled" analog, with the sampled digital at the bottom. It was very clear-cut.

Atkinson: Where are the specific subjective areas in which you feel the digital machines fall down? Many writers, for example (myself included), have said that digital has fantastic bass definition.

Boyk: Completely false! I know digital is supposed to have good bass—it ought to have good bass—but the bass in all the digital I've ever heard has been horrible, truly awful. All I can assume is that a lot of people don't know what musical bass, from live acoustic instruments, actually sounds like.

Atkinson: Some writers have written that, whereas the best of the conventional media preserve the sense of space around instruments, the digital doesn't.

Boyk: The "room" does go away with digital, and the low-level information with it. The quietness of digital, the vaunted lack of background noise—which it certainly has, I'm happy to give that to it—in my opinion is a direct concomitant of the lack of detail and the lack of room. One side of the coin is no background noise; the other is no detail. And part of no detail is no instrumental timbre detail, no textural detail; the other part is no room.

Atkinson: When I did the piano recording for the Hi-Fi News Test CD, comparing the live feed with the PCM-F1 digital tape revealed that the digital tape did make the acoustic drier, as though I had been recording in a slightly smaller hall. Depending on the player, the CD goes farther in the same direction. As far as I'm concerned, my miking becomes less optimally balanced for the ratio between direct and reverberant sound. But to someone who wasn't at the original session, they would assume that, yes, it was a small hall, but it is still believable. Does this distortion of perspective matter? To many it doesn't seem to. Digital may have such problems but people still happily buy CDs.

Boyk: Well look, anything anyone is happy with, I'm happy for. Happiness is what we want more of in life, and if someone tells me "I love CD, it's a terrific medium, I get into the music," I'm not going to tell him that he doesn't, that he isn't happy. But he shouldn't tell me that I am happy.

You can't magically throw away the room without also losing all kinds of other information. The signal doesn't know which part of itself is room and which part is instrumental detail, and such fine detail in the sound is actually a crucial part of the performer's art. Instrumental textures become homogenized. The difference between a solo instrument playing and a whole choir of instruments answering gets homogenized. Timbral differences—the difference between an oboe and a clarinet, or a violin and a cello—also get homogenized by digital. Can you still tell it's a violin rather than an oboe? Probably. But does the difference have the full impact of that it should have? No, it doesn't. One of the great pleasures in music is simply the different voices of the instruments. Sometimes it can actually be quite difficult to tell.

There's often—not always—something in the background, almost as though there is a wind machine in the back of the orchestra. It's not loud enough to hear explicitly, but it intermodulates with everything else. The homogenization of timbres and textures comes from that intermodulation, from the fact that you have clarinet plus wind machine, oboe plus wind machine, violin plus wind machine. To the extent that they all have a wind machine in common, they all sound alike.

I said earlier that the first thing about audio is that music is beautiful. The second is that the beauty of music is the sensual bridge to its meaning. The raw sound of live acoustic music is a part of the meaning of the music. If the sound is different, the meaning is different. A thing I hear wrong in digital recording is the lack of what I've called "dynamic inflection," which is the natural rise and fall of dynamics. I'm not talking about gross dynamic range, but the thing that makes "Please pass the salt" different from "Please pass the salt." Dynamic inflection is what gives natural expressiveness to music: analog seems to preserve it well, whereas it seems to be damaged by digital.

The perceived dynamic range of digital recordings, to me, is inferior to that of analog. There simply is so much less dynamic range, perceptually, than analog that it's a joke to compare the two. For instance—I'll give an extreme example—I've never heard a CD or a digital master tape that has a perceived dynamic range equal to that of the Sheffield Firebird direct-to-disc recording (with the proviso that you have to have a system that can play back the dynamic range).