Seminar 1989: Exploring the Current Best Thinking on Audio, Part II of the Three-Part Transcript (Issue No. 14 Summer through Winter 1989-90)

Seminar 1989: Exploring the Current Best Thinking on Audio, Part II of the Three-Part Transcript

The complete transcript takes up so many pages that we have decided to publish it in three parts, instead of two as originally announced. To get the most out of Part II, you need to have read Part I. Have you?

All the introductions, opening remarks, and prefatory throat clearings were taken care of in the last issue, at the beginning of Part I. We refer you to the capsule biographies of the six panelists published there; we certainly want you to be aware who is who when you read who said what. You will recall that the transcript ended where the participants adjourned for lunch; it resumes here where they returned to the conference table. The juicy gossip in the dining room was unfortunately not captured on tape, probably a blessing in disguise from the legal point of view...

EDITOR: Let's start discussing the pro gram material in general and recording techniques in particular. I would like to ask each of you-three of you being very active recordists and all of you having strong opinions on the subject-I'd like to ask you severally and together what you think the right recording technique is to achieve the kind of results that we've been talking about, that we all want. Who would like to start?

EARGLE: Okay, the thumb is pointing in my direction. First off, I think we have to put this in some perspective by saying that, in this country at least, less than 5% of the recorded material sold is classical or based on a natural model, a natural acoustical model of some kind that we want to trans late over, the rest of it being entirely multi track and open to a new set of laws or lack of laws. It's purely whatever is created over the loudspeakers in the control room.

The healthiest thing to happen in classical recording, which is really what we're talking about here, happened about ten years ago, and it was the fact that going to digital recording forced you more or less back to two tracks-because there were no digital multitrack machines at that time that were practical. The Soundstream machine could always lay down 8 tracks at once but with only 20 minutes playing time for the whole pass. So that encouraged a lot of people to throw away their multitrack machines for classical and go back to even earlier than where they had been in the early '60's, when I began in this business. They were using three and four tracks in those days.

And TI might say parenthetically that the worst thing to try to figure out a use for was a three-track recorder back in 1962. It was an invitation to put the soloist on the center track and the violins on the left and the cellos on the right, without any real sense. In other words, the only stereo was really whatever leakage there was between all three of these. So if you ever tried to play with the level of the soloist, you were really playing with the level of the middle of the orchestra at the same time, so it really wasn't a satisfactory solution.

LIPSHITZ: You put the Mercury recordings in that category?

EARGLE: The Mercury stereo recordings were basically the two outside tracks, with just enough of the center channel, the mono channel, to give you a little bit of center fill, and it was usually run about 6 dB down.

LIPSHITZ: All right, but it was three widely spaced microphones, and each went to one track.

EARGLE: They weren't that widely spaced. There was a group of microphones that might have been panned together; they were sent to the two outside tracks. The middle track was always used for the mono disc alone. I mean nine times out of ten; that was a single mike that somebody had figured a good place to put. But I really think that clearing the house, so to speak, was the best thing that ever happened in classical, and a lot of people are still recording two-track classical, even now with multitrack available. Not every body, however; in fact, Deutsche Grammophon regularly uses a Sony 32-track for recording classical music. They have their own editor that will edit that many tracks at once, and after the Aufnahmeleiter has done the editing, it's then given to the Tonmeister to do a mixdown, and it's entirely in his hands from that point. But in this country, many of the smaller companies have gone to a basic two-track format and as such have sort of wanted to simplify the microphone array, maybe limiting it to no more than 8 to 10 microphones, maximum.

LIPSHITZ: But the fact that the recording might be done on two-channel machines doesn't mean it will be only a few mikes.

EARGLE: No, it means that you've got to make the right decisions.

LIPSHITZ: You can't change your decisions.

EARGLE: That's right, you can't change them later, and it will make you go for a safer perspective on the ensemble because you know you can't fudge it later.

LIPSHITZ: Now I believe Decca/London historically always have-except for their big opera recordings-mixed down live to two channels.

EARGLE: Well, the few Decca sessions that I've been to have been basically direct to two but with a multitrack backup. And the way they do it is really very, very sensible. The have a console, a homemade console that has several sections in it. Each section is really a two-channel board. One section will be devoted to the main array across the front; another section will be devoted to house microphones; all of the accent mikes will be routed to another section--and these will all be fed to their own left/right outputs, which individually go to pairs of tracks on the multitrack. But at the same time all the L's and R's are sound and feed directly to a two-track machine.

Now, if they find that they made a mistake in the balancing-in the on-the-fly balance of the two-track, whatever that mistake might have been-they always have the option of pulling out the multitrack tape and setting it up again with individual calibration points and duplicating the two track output, except they can now go in and ride gain on these various sections.

Now, that's limiting because you can't go in and change one microphone at that point, but you can change the entire section, the entire subgroup-which seems to me to be the most sensible way if you really feel you need backup to cover yourself for a potential artist problem or balance problem in the eyes of the conductor.

That's a very good way to do it. But over all I would say that we're seeing a lot more sanity in the miking of all kinds of classical music, and we're seeing, more than anything else, a new generation of micro phones. Remember that you never heard of B&K instrumentation mikes being used for this purpose until the first digital recorders came along-because they had flat power bandwidth whereas many analog machines of the day would actually have high frequency saturation.

LIPSHITZ: But the B&K mikes would not always be used correctly.

EARGLE: Well, when you say correctly...

LIPSHITZ: Well, I am thinking of a specific example. Some of Marc Aubort's recordings, for example, where he used B&K 4134's, which is the pressure mike, not the free-field mike.

EARGLE: Okay, yeah. A lot of people didn't understand the difference between a flat-on-axis and a flat random-incidence microphone. You're right; many times the wrong mikes were used. They were also quite a bit noisier than most of the mikes, but they at least didn't have a high frequency peak.

LIPSHITZ: Well, wait a minute. Wait, wait, wait.

McGRATH: Depends on the axis, how you point it.

LIPSHITZ: I asked him about that; I asked him how he had used his 4134's-were they pointing at the musicians or were they pointing up at the ceiling?

EARGLE: Well, granted, he was misusing them in terms of...

LIPSHITZ: And they were pointing at the musicians, which meant they were rising above 5 kHz.

EARGLE: But if they were properly used, they didn't have the peak that the current Neumanns and AKG's had at that time.

That's the point I'm trying to make.

LIPSHITZ: Right. That's smooth, but not flat the way he used them.

EARGLE: But if properly used, they were smooth and flat.

LIPSHITZ: Yes.

EARGLE: And I think that today we see everybody using all redesigned microphones, things that are the product of the last ten years. Schoeps, the Sennheisers, the newer Neumanns, AKG's-and I think that that's made a drastic difference in the texture of our recordings, of everybody's recordings.

LIPSHITZ: For example, Decca/London, at least according to what I've read in the magazines, shortly after they started coming in for some criticism about a rather harsh, strident high end-I think this was actually before the CD was out but when some of the digitally recorded tapes were put on LP-bought a whole batch of Schoeps microphones to replace their omnis with the rising high end; they bought a batch of omnis with flat high ends.

EARGLE: Yes, they had been using the M-50's, which have the rising high end.

That's a hybrid microphone that is omni at low frequencies and becomes directional at high frequencies; it's got a mike mounted on a little spear inside, about that big, if you've ever seen the internal structure of it. That was a product of around 1950 or so and was designed by Neumann for rather distant pickup in concert halls, for broad cast, where they couldn't maybe have the mikes right down on top. And Decca adapted that, and that was their mainstay, and with the old analog recorders it didn't sound so bad.

CARVER: What did you mean when you said, change "the texture of your recordings"?

EARGLE: Well, okay. There are two words that I use, structure and texture in the recording. Structure, to me, is being able to determine quite unambiguously where something on the soundstage is. I feel the two are of almost equal importance, and one without the other is a very unsatisfactory recording. Structure is what tells you, you know, where the violins are, where the winds are, and the brass...

CARVER: So that would be our, or my, "imaging," perhaps.

EARGLE: Well, imaging in that sense, in the sense of left/right, front/back. Texture, to me, implies an overall coloration or glow to the sound. In my own miking, for the most part, there's a near-coincident pair in the middle, an ORTF pair usually, which provides all the structure that you're

"But overall I would say that we're seeing a lot more sanity in the miking of all kinds of classical music, and we're seeing...a new generation of microphones."

going to need in a recording. Flanking that, and operating anywhere from 6 to 8 dB down, might be a pair of omnis.

CARVER: Those give you the texture.

EARGLE: Well, they help in that regard, and they also give you a little bit more of the string sound up front, across the front of the orchestra.

CARVER: They supply the bloom.

EARGLE: That’s right. You're using the word "bloom," you see, and I could ask what that means, except I think I probably already know what it means.

LIPSHITZ: An ambient field.

EARGLE: Yes. Then there's also a pair of mikes in the back of the hall, if I'm operating on the stage. If you're operating well out into a room that has a lot of reverberation, then you may not need those. And all these subsidiary microphones are really operating lower enough in level so as not to dominate. I'm sure that you operate in similar ways.

McGRATH: Well, simpler. It depends on what I'm recording. Probably two thirds or three quarters of what I do is orchestral recording for broadcast or for just concert recording, which are live musical events, where what I can do in terms of the positioning of the orchestra, or indeed the microphones, is very limited. So I have to go with the best that I can do, and having worked now for about 12 years in one particular hall that I enjoy working with, I've found what is a very hot "sweet spot" for my spaced omnis, which gives me the kind of perspective that I'm very comfortable with.

EDITOR: Which ones do you use?

McGRATH: I use the Schoeps. I've been using them for the last six years.

LIPSHITZ: These are the Colette Series capsules?

McGRATH: No. Well, they could be on Colettes, but they're CMC-5's with the MK-2S capsule.

LIPSHITZ: 2S?

McGRATH: It's called the 2S. They also have the MK-2, and then there's a new MK-2 which I'm not sure I like. I've got them all.

LIPSHITZ: What's the difference between the 2 and the 2S, Peter?

McGRATH: The 2S probably has a little bit more bite.

LIPSHITZ: It's up at the high end?

McGRATH: I don't know if it's brighter; it just sems to have a little bit more focus.

Perhaps it is simply a rise; I don't know.

LIPSHITZ: Yes, I think it is.

McGRATH: Probably is; it's probably a bit peaky, but it sounds very, very good.

EDITOR: You used to be into B&K's, the instrumentation mikes.

McGRATH: I used to use the 4133, not the 4134. 1 didn't like the 4134. But the 4133 did sound good.

EDITOR: And what about the new ones that are for music, expressly?

McGRATH: I bought the 4003, and I didn't like that microphone. I found that it was appreciably quieter than the 4133 that I had used, and that in itself was a plus, but then...

EDITOR: That's why they came out with that series, right?

McGRATH: Yes, but then the Schoeps MK.-2S was brought to my attention. I kind of glommed on to that mike, so to speak, and I have really not moved from it. I like it a lot.

LIPSHITZ: Peter, when you used the 4133's, you used them with the little grids on?

McGRATH: I used it with the regular grid on.

LIPSHITZ: Pointing at the musicians?

McGRATH: Yes. So that was rising on axis? By how much?

LIPSHITZ: No, no. The pressure mike, the 4134, would have been rising. The 4134 is flat, pressure, if you remove the grid...

McGRATH: I tried both capsules.

LIPSHITZ: ...and if you were using the 4134, as Marc Aubort did, you should have pointed them at the ceiling, so that the sound came at grazing incidence. That gives them flat direct and flat random incidence pickup; that would have, in fact, been better, in that sense, than you could get with the 4133, which would be flat, direct, but rolling off, random incidence.

EDITOR: That's right in the instruction manual, for that mike, if you use it for measuring.

EARGLE: But who ever reads the instruction sheet on a microphone?

EDITOR: John, you use a Japanese con denser mike, don't you?

EARGLE: Yes, the Sanken CU-41.

LIPSHITZ: That's a two-way microphone.

EARGLE: Yes, it's a woofer-tweeter microphone. And what I like about it, in particular, is the frontal response of a cardioid--this microphone is zero on axis and exactly -6 dB down 90° off, all the way out to 12.5 kHz. And the pattern is absolutely uniform over that entire frequency range in the front half; the back half is pretty choppy, as it is on all cardioid mikes. The reason I do that is that, when you have a near-coincident pair splayed out-and the splay angle may be well beyond 110°, which ORTF recommends- the fact is that the middle of the orchestra comes in with lack of coloration, of off axis coloration, and I find that to be very, very important.

LIPSHITZ: I think the B&K cardioids are probably also very good in that respect; at least according to the curves they look very good.

EARGLE: Yes, they are.

McGRATH: I've not found cardioids that I like, for that very reason. I haven't tried the Sanken, nor have I tried the B&K, but the ragged quality of the middle is horrendous.

LIPSHITZ: I quite honestly don't like cardioids.

McGRATH: I don't, either.

LIPSHITZ: You know, I'm not involved in making commercial recordings, and essentially everything I record is live concerts.

EARGLE: You're more or less in a documentary mode.

LIPSHITZ: Well, it's the local chamber music society. We put on annually about 50 concerts, and they're all recorded for broadcast on the university radio station.

We have a weekly radio program, and with 50 concerts a year we have the material to do that. In fact, we have been recording digitally for about six years now, I guess, and broadcasting live from the digital master tapes. I suspect we may have been the first North American station to broadcast live digital recordings over the radio. We must have 250 or 300 master tapes.

EDITOR: Stanley, I understand you believe in single-point miking.

LIPSHITZ: Well, I believe in single-point miking, and we can say more about those ideas later, but single-point cardioids-the ORTF is spaced, slightly spaced, quasi coincident cardioids-I don't like single point cardioids.

EARGLE: Well, you have way, way too much middle, unless you splay them way out.

LIPSHITZ: The reverb is all bunched up in the middle of the soundstage between the speakers. Edward Tatnall Canby in Audio magazine a few months ago was talking about this; he had a number of articles on the Denon Mahler cycle with Inbal, in Europe...

EDITOR: I reviewed one or two of those.

LIPSHITZ: ...and they used the B&K cardioids even before they were released. He made some comments about imaging and ambience and recording techniques, and so on-some things I disagreed with, and in fact they just recently published a letter of mine commenting on some of these things

-but he made one interesting point that I hadn't really appreciated that clearly till he made it, and that is that coincident cardioids are in-phase for the full 360° pickup.

There's no out-of-phase rear lobe, and the result is, when you reproduce that over speakers, that all the reverberant images are between the speakers.

CARVER: That's right.

LIPSHITZ: And if you analyze it, the reverberation sort of bunches in the center, so you have a center-weighted reverb. It's closed-in sound.

CARVER: The reverb should have a lot of out-of-phase components.

LIPSHITZ: Yes.

CARVER: It drives me nuts when I hear recordings that are made, and the reverb is in-phase.

EARGLE: I agree. You mean predominantly in-phase?

CARVER: Predominantly in-phase. And that bloom that I like a lot just evaporates.

LIPSHITZ: So you see-just to complete this little thing, then I'm sure you'll all want to comment-I don't like coincident cardioids. 1 far prefer coincident figure eights or coincident hypercardioids, both of which have a full-level or -6 dB out-of-"I believe in single-point miking...but I don't like single-point cardioids. The reverb is all bunched up in the middle of the soundstage between the speakers." a ST EO Sere] phase rear lobe, both of which give a significant amount of incoherent reverb pickup-pickup of directions other than the direct sound. In fact, the figure eights, of course, because you've got a sin" + cos = 1, they have equal pickup in all directions in the horizontal plane, so that they're totally neutral for direct and reverb pickup, and both side quadrants are out of phase.

EARGLE: I can give you a problem in using a single pair of microphones for picking up both reverberation and direct sound. A lot of times... well, it depends on the hall you're working in, and it depends on your ability to reseat players. If you're on the floor of a large English church that's been deconsecrated and converted to a recording studio, then you can pretty much move microphones...

EDITOR: In that order? (General hilarity.)

EARGLE: Yeah, yeah. The players will be consecrated at the first break. If you know English players, they'll find the nearest bar. Anyhow, the thing is that, when you're at the beginning of a session, you've got to make a lot of very, very quick decisions, and having separate quantities in the recording available on separate faders is a godsend. Because if you wanted more reverberation, for example, or less reverberation, with a crossed pair of figure eights, you'd have to move in or out of the orchestra. So in changing one perspective, you're changing the other one, too. And that gets to be a problem because, some times, the right texture on strings and their balance with the rest of the orchestra depend upon being in a very narrow front to-back zone of-and I'm not kidding you-around two feet.

LIPSHITZ: I think you're right. There are trade-offs. What I'm saying has nothing to do with the commercial realities of having to make a recording. My point was simply that-given that one was making a quasi coincident-type, ORTF-type recording- the reason that ORTF is preferred to coincident cardioids, I believe, is the slight phasiness due to the spacing...

McGRATH: It gives that open quality.

LIPSHITZ: ...because you have no phasiness with coincident cardioids, since everything is inphase, the direct sound and the reverb. Now the direct sound should be coherent, should be inphase, but you want the rest to have the kind of relationship it has live, which is incoherent. And you do get that with figure eights; you get it with coincident mikes with out-of-phase rear lobes. I think a lot of people who have tried coincident or quasi-coincident recording have done one of two things: they've either used mikes with very poor polar pat terns or not good enough polar patterns- because, as you pointed out, John, you've got to have very good off-axis performance if you want to be free of coloration in the pickup-or they have used cardioids and didn't like the sound and ruled out the technique. There are all these other reasons why one might not be able to manage with that sort of configuration, but what worries me about some of the techniques, for ex ample Dave Griesinger's equalization of coincident mike recordings, where he boosts the L - R signal below, say, 500 Hz by 6 dB or more...

EARGLE: Or more--sometimes more.

LIPSHITZ: ...now boosting L - R below 500 is not the same as cutting L - R above 500. When you boost L - R relative to the L + R reference, what was a coherent recording becomes phasy in the low frequency regime because, when the L - R is added to the L + R to get back left and right, the L - R can dominate the L + R, and you can figure out that with a coinci dent pickup what was inphase direct sound then becomes antiphase direct sound for sources more than a certain amount from the center. So he is adding phasiness on di rect sound, and that is a great pity in order to produce some kind of ambient, airy, phasy sound quality altogether. Whereas if you cut L - R above 500 Hz-of course you just make things more mono-you don't add the antiphase. So I don't see that as a solution to what is perceived by some people as a problem. I guess that's enough for this moment.

EDITOR: What I hear so far is that there is no substantial disagreement among you about the kind of sound you would like to achieve; there is some minor disagreement as to the tools that you would use to achieve that. Do I misunderstand you; are you after different kinds of sound?

McGRATH: Well, I think that's an assumption that may or may not stand up. If Stanley were to play a recording that he liked on a system that he approved, and then I would play one that I like on a similar system, we might then get into some very serious disagreements as to what we like and what we don't like.

LIPSHITZ: I think we would find we disagree because your playback system sounds-on account of the speakers-very similar to mine. We're talking Quad ESL 63's, which are speakers where, because of the polar pattern of the speaker, you've got less room effect on the playback, so that you're listening analytically to what's in the recording-and I think we would probably disagree.

McGRATH: We'd probably disagree on what we like. But what figure eight do you use? What coincident figure eight capsules?

LIPSHITZ: Well, we had no good micro phones available initially. We had a pair of U-87's at the university radio station that we used and built a jig to hold them. The only polar pattern by the way for the U-87 that's at all good is the figure eight. The most difficult polar pattern to make by the way is a cardioid. Most mikes that are omnis or figure eights will be better than the corresponding cardioid. So we used those and experimented with them for a number of years, and then in 1978 Anton Kuerti, Toronto pianist, one of Canada's finest pianists, played a complete Beethoven sonata cycle for the chamber music society- eight concerts. And we broadcast him live from the Theater of the Arts straight to the FM transmitter and bought some custom equipment to do it. We applied for an Ontario government grant for some micro phones, and we got it. So halfway through that cycle, after a great deal of soul searching and investigating, we went for Schoeps-a couple of bodies and a set of capsules. The capsules we bought were the MK-41 hypercardioids, the MK-40 cardioids, and the MK-8 figure eights, which have only recently become available. The second half of those concerts were broad cast using the MK-8's; the Theater of the Arts is fairly dead. So the figure eights I'm talking about that are good are the Schoeps MK-8's. The Schoeps cardioids are also quite good; the Schoeps hypercardioids are very consistent; but more recently we use a Calrec Soundfield microphone as a stereo mike, and there the polar patterns are all extremely accurate because they're synthesized from the outputs of four capsules in the head, the array, of the microphone, which are individually equalized for frequency response, so you're talking figure eights that do not roll off below 100 Hz but are equalized flat to 30 Hz.

McGRATH: That's what I was going to ask. On the MK-8's you had to use some kind of EQ, didn't you?

LIPSHITZ: Not for chamber music, usually. No, we didn't.

EDITOR: Dave, you wanted to say some thing?

CLARK: I'll make a statement about using recordings. The only recordings I've been involved with are studio-type recordings.

But in using recordings, the work that I do at this time is in automobile sound systems, which I guess aren't up to the level of what we're talking about here...

EDITOR: Conceivably they are.

CLARK: ...however, getting back to sit ting off center in a living-room situation, you have the same problem--the biggest problem of listening in a car is that you're not in the center. I feel that a lot of these techniques are evaluated by sitting on the center line between the speakers, head in a vise so to speak, and you say I've got imaging from here to there, and I can pick things out-for me that's not good enough.

In a way it's better than it needs to be; I can get better imaging-or structure, as John says-of the soundstage, sitting in the center between two speakers, than I need...

EARGLE: Sometimes it's too good, yes...

CLARK: ...but as soon as I move off center, it degrades to a level that is not up to my standards. Namely, everything piles into the near speaker. In a simplistic way I tend to blame that on coincident miking because it almost always has elements of the sound, of any sound on the soundstage, coming from both left and right speakers.

LIPSHITZ: Which does? Sorry?

"...I can get better imaging sitting in the center between two speakers than I need...

but as soon as I move off center...everything piles into the near speaker."

CLARK: Coincident.

LIPSHITZ: I'm not quite sure I under stand.

CLARK: Without a time delay.

LIPSHITZ: You're saying that those speakers carry all the instruments.

CLARK: Right.

EARGLE: But not in the same amount, except along the median plane.

CLARK: Not in the same amount.

LIPSHITZ: Of course not. If you're using figure eights, any speaker on the right channel axis would be in the null of the left microphone-so that would not be true for hard right and hard left, which would be the direct sound, single channel only.

CLARK: That's right. You can find them, just as with a pan pot in a studio you can pan all the way from one side to the other.

LIPSHITZ: And such a source would stay in the right loudspeaker no matter where you listened from because there would be no left-loudspeaker direct sound for such a source on a coincident recording.

EARGLE: Dave, I think what you are addressing here is a tendency of anything that is absolutely common mode between the two channels-in other words, equal information between channels-to really flop over to the nearest loudspeaker when you're off axis.

CLARK: That's exactly what I've been saying.

EARGLE: And for that very reason, I find that in a recording made with three spaced omnis-Ieft, center and right-I find that this hard center, which is absolutely in phase between channels, just collapses over. And it's one reason why I never pan anything in the middle-even if I have a soloist. I always have a stereo pair on the soloist. And those mikes are fed hard left and hard right, and it's just the fact that the soloist is off axis on both that creates a center image that's not pinpoint-it has a little bit of area to it.

McGRATH: But it feels real. I found that technique works very well.

EARGLE: It feels very real. Also-a very, very critical thing for me--if I'm seated right on axis, you can postulate that that's the best place in the world to sit, but if you're a little bit off, that common-mode information is going to get into some pretty broad comb filtering.

CLARK: Yes, it does.

EARGLE: Until you get maybe a couple of feet off axis. One way you can diffuse that problem is to avoid it by having quasi coincident microphones which don't pick up that tight a center-pan-potted middle.

CARVER: So that the vocalist has a little stereo spread.

LIPSHITZ: A little phasiness.

CARVER: A little L-R.

LIPSHITZ: Let me put a contrary view on this, to what Dave was saying. With coincident recording methods, the direct sound is essentially in-phase in the two channels.

EARGLE: It's like pan pot.

LIPSHITZ: Yes. And the reverberant sound done properly should be in random phase-that's why I was making the case against using coincident cardioids. That being the case, as you move away from the center the image moves over and fairly rapidly moves into the nearer loudspeaker.

That's true. So central images will move over; any hard left image would stay in the left-hand loudspeaker, and any hard right would stay in the right-hand loudspeaker when you move. So what happens is that the image stretches. The left-hand end stays where it was; almost all the rest moves over to the nearer loudspeaker.

Now let's look at an example of, say, spaced omnis, the sort of thing that Peter uses.

EDITOR: And John does.

LIPSHITZ: John adds them to a quasi coincident central thing at a slightly reduced level.

EDITOR: Oh, I see.

EARGLE: An organ recording might have a pair of spaced omnis because of the special ambience that you're trying to pick up, but that's not an image problem.

LIPSHITZ: Well, it's not clear what the image of an organ is because different organs have different images depending on where they put the pipes.

EARGLE: That's right.

LIPSHITZ: But anyhow, the point was this-that with the spaced omni recording there isn't any clear image. You know, I can go into more detail about that, but my great objection to spaced omnis is that the theory of imaging doesn't exist. For classical-type recordings, the microphones being relatively distant from the musicians, the main difference for sources near the center-musicians near the center, woodwinds and so on-is time difference, time-of arrival difference at the microphones. With coincident recordings, of course, there are no time-of-arrival differences; it's all level differences in the two channels, the two mikes. Time-of-arrival differences at the mikes produce level differences at the listener's ears. Level differences at the mikes produce time-of-arrival differences at the listener's ears, the reason being that each ear hears both loudspeakers. This is a very misunderstood thing, and I can't prove it for you here, but...

EARGLE: Well, it can be proved.

LIPSHITZ: It can be shown that over the region from 1 kHz down, or at least 500 Hz down, which is half the audio band width in terms of octaves, that is the case.

Now the dominant hearing mechanism is time-of-arrival differences at the ears, and coincident recording produces recordings which match that-in other words, which are natural in that sense. Spaced-mike recording produces level and polarity differences at the ears. And what you actually find is that-take a slightly off-center source, spaced mikes, and let that instrument play up the scale-and what you find is, as the instrument plays up the scale, you have...

EARGLE: Rotating vectors...

LIPSHITZ: ...yes, at your one ear the signal level will change, go through a null, reverse polarity, change, go through a null, come up again. And the rate at which it changes with frequency is different at the two ears. So what you're getting is in-phase and antiphase signals at your two ears. In other words, total phasiness. You never, live, get signals that remain in-phase over a band of frequencies and then antiphase over a band of frequencies.

EARGLE: On the other hand, if you take a coincident recording and sit off axis...

LIPSHITZ: You add some of that to the other things. Yes. So my point is this. You see, time-of-arrival differences at the ears produce phase shifts, but a phase that varies linearly with frequency, so they are delays. You don't get a phase that stays 180° over this whole band of frequencies and suddenly flips to zero and then back to 180°-that kind of thing. That is what I think a lot of people are crediting with ambient, airy warmth. It's not ambience.

It's phasiness on direct sound. That would be fine on ambience, but it's there on the direct sound. Now, just to complete my point-because you're all going to want to comment at length on what I just said- when you move away from the center of a coincident-mike recording, you add, as John has just indicated, some of this phasiness problem, which you don't get at the center. The image moves over, collapses into the one speaker, and phasiness is introduced. When you start with a spaced mike recording, you have that phasiness already at the center listening position; you don't have any kind of precise imaging;

things near the center are really quite anomalous and move around with frequency. And when you move away from the center, you don't get the same impression of collapse because you didn't have any thing really to collapse in the first place.

But things are still happening the same way. It's just that the aural impression is of less of a degradation because you've got less to degrade; it's already degraded in the middle. There's real things to attack now.

McGRATH: Well, the response I would give to that is simply this: I'm not disputing the validity of anything that you've said. The only thing that I can offer is that I own the Schoeps MK-8's and I've used them in maybe 30 different halls, and I've used the MK-2S's in a spaced configuration and I've used them in a variety of different musical situations ranging from solo piano to Mahler symphonies. And in many cases I've not only done that but I've also hung a pair of spaced omnis and then critically positioned a pair of MK-8's and have run parallel recordings-and without injecting my personal bias on it, I've played both back, and in every situation that I've been able to conjure up, I've gotten a more generally preferred result with the spaced omnis. That's it, simply stated. [ don't dispute you. And if I could get an MK-8 or a

---

"That is what I think a lot of people are crediting with ambient, airy warmth.

It's not ambience. It's phasiness on direct sound."

---

coincident mike or indeed a figure eight that gave me the tonal characteristics that I achieve on a consistent basis with my omnis, I probably wouldn't hesitate to use it for all the reasons that you cited.

LIPSHITZ: Can you tell me what the criteria are? You see, I think we probably wouldn't agree on what we mean by stereo.

McGRATH: Well, I think that the criteria are, you know, very simplistically speaking, that one sounds more like what most people's perception of beauty is. One sounds pretty thin, one sounds pretty hard, one sounds pretty unlistenable, sounds pretty constricted; the other one sounds maybe very vague and phasy, but it has palpable beauty in it, and people generally tend to like it.

LIPSHITZ: Now the MK-8's do roll off below 100, 150 Hz gradually, whereas the omnis are much flatter-but that's not what you're talking about.

McGRATH: Also, the string tone sounds always very harsh to me; it's always very gritty and aggressive.

LIPSHITZ: Are these good acoustics that you're talking about? Because, of course, the omni has a worse direct reverb characteristic than the figure eights.

McGRATH: Well, we're dealing with the reality of going in and making a recording in a hall, and in different hall situations I keep coming up with this consistent result.

That's all I can argue. I mean, the results speak for themselves.

EDITOR: Maybe we could bring this down to the level of most audiophiles if each of you could state what you like among commercially available recordings and what you don't like, and use those as an illustration of what you're talking about. Because this is highly theoretical; in fact, Dave just showed me your article here, Stanley, which is in the September 1986 Journal of the Audio Engineering Society, where you provide mathematical proof of what you just discussed.

LIPSHITZ: Well, I mean, it's not original, please! There are at least seven or eight references there. You'll find some of the proofs elsewhere. It's just-I was getting frustrated that people would apparently misunderstand that thesis.

EDITOR: But could you provide some commercially available examples of these various characteristics that we're talking about?

McGRATH: You mean in terms of recordings that substantiate one point of view versus the other?

EDITOR: Yes. Right.

CLARK: I have a hard time finding a recording I think was made with two spaced omnis.

McGRATH: I've got 27 of them that I brought with me.

EARGLE: Listen. I can tell all of you gentlemen here-and anybody who disagrees will have to substantiate the disagreement--on a typically good stereo recording I'm hard put to tell exactly how it was made. On a bad stereo recording I can cite chapter and verse, probably, on how it was made.

McGRATH: That's a very profound point.

I would agree with that a hundred percent.

EDITOR: That's a striking thought.

EARGLE: So the thing is that there are many, many nice recordings about which I'm shocked when I find out how they were made. There's a recording of Arleen Auger that's a Grammy nominee for best female vocal, classical.

EDITOR: This is your recording?

EARGLE: Yes. And you will all agree when you hear it that this is a very natural, very nice concert setting of a singer.

Horse-feathers. It was not. It was done in a chapel, where I used the natural ambience of the room, but because of the music making conditions I could not mike her standing in front of the piano. First off, you all realize that when you have a vocal recital, the piano is usually on half-stick, and the singer nests herself in the crook of the piano. And it's because of half-stick that you get the proper balance from instrument and singer going out into the hall. Now, when you're trying to document this as a recording, a piano doesn't sound good on half-stick.

McGRATH: Sounds hard.

EARGLE: Sounds terrible.

LIPSHITZ: Loses its brilliance.

EARGLE: Sounds terrible. Okay, so if you raise the piano cover all the way up...

McGRATH: She's lost.

EARGLE: ...she's lost. So you have to move her out a little bit; they lose communication...

McGRATH: Then they can't play together.

Always a problem.

EARGLE: ...and then they can't play together. Okay. So the way we solve the problem is very simple.

LIPSHITZ: Stand to the right of the key board.

EARGLE: No, not even that. You want her in the middle. You want her in the middle because that is an expectation for many listeners. It's an expectation for broadcast. So what you do, you mike the piano as an event, and you put the singer about eight feet away facing the piano and put a mike in front of her, in this case a cardioid, about four, five feet away, and balance to that way.

LIPSHITZ: If you'd used figure eights, you could have put them one in the front and one in the back, except one batch would have been out of phase, out of polarity.

EARGLE: That's right. I'm not sure how audible...

McGRATH: How do you correct that?

EARGLE: You really can't.

CLARK: She could sing into a reflector...

LIPSHITZ: She could breathe in while she sings.

MCGRATH: She could sing backwards! (General merriment.)

EARGLE: The point is-and I also had a pair of mikes in the house to pick up a little bit more room sound and mix in-that I defy you to spot any of these ingredients when you hear the recording. Anyhow, the best recordings-of course, as in the case of anybody who's in it professionally and has product, you can all go out and listen to my records if you want to-I would say that among the work of others some of the more recent English Decca things have been masterpieces of multi-miking. A re cording that I've liked rather recently is one that Jimmy Locke did with the San Francisco Orchestra. It sounds to me like they've gone away from the M-50's and the KM-84's that they've been using for years.

EDITOR: On what label is that, John?

EARGLE: English Decca. London in this country.

McGRATH: Some of the Dutoit recordings are rather exquisite, and I don't know what they do, but they're thoroughly enjoyable.

EDITOR: Would you make that statement about the Dutoit recordings from day one or just the very recent ones?

McGRATH: I haven't listened to them all.

LIPSHITZ: They've all been done in the same church near Montreal. I haven't visit ed it but I did read somewhere that the ambient sound you hear is not entirely the church; it's been assisted electronically. I believe that's the case; I may be wrong.

EARGLE: A little bit of that is all right. If you have a natural base to work from, you can tail out just a little bit more with the Lexicon and have it sound very, very natural. You wouldn't want to do that in a dead room.

LIPSHITZ: No. I believe they take foam things and put them in some of the seats. I think they actually erect a new platform in the body of the church, where the musicians sit. You know, it's not what most people probably visualize when they think of a recording being made in a church.

EARGLE: Well, Decca always-and they do it in San Francisco, they did it in L.A.-where the orchestra is seated on a stage, they will build a platform out from the stage, extending it maybe 20 feet out into the house. And one thing that you get when you do that, you find that the entire orchestra excites the same acoustical space. When the orchestra is back inside a shell, the back of the orchestra is in one environment and the strings are in another, and that's a problem.

LIPSHITZ: What's best for listening is not necessarily best for recording.

EARGLE: Absolutely. Because, in con cert-hall listening, you may have 2500 people listening to one watt peak power coming out of the orchestra, and you need all the assistance you can get from a very sympathetic and reflective shell. And that plays havoc when you're trying to record.

It gives you very unnatural, dry sounds.

CARVER: Dry sounds?

EARGLE: Yes. Well, dry in the sense of

"[As for] the best recordings

...I would say that among the work of others some of the more recent English Decca

[ London] things have been masterpieces of multi-miking." having a very high direct reverberant ratio.

EDITOR: John, among your own recordings, which are you the proudest of?

EARGLE: Well, I would say that the Petrouchka recording that came out recently from the Seattle Symphony, with the

"Scherzo Fantastique" and "Fireworks" and the "Russian Easter Overture"'-it's a real bargain, you know, approximately 70 minutes' worth of music-that one would certainly be an example.

LIPSHITZ: That's on Delos?

EARGLE: Yes. There's a new recording that's coming out on the Pro Arte label that I did; it was a recording of the Strauss Le bourgeois gentilhomme, done in Studio A at RCA with the New York Chamber Orchestra. I haven't heard the recording yet in its edited form, but I suspect it's going to be good. An orchestra of that size fits beautifully in a room that size, without artificial reverberation, just using the room itself.

LIPSHITZ: But you're not suggesting that the sound would be different in the mixed two-channel digital master you prepared and in the edited version of the CD surely?

EARGLE: No. Absolutely, the sound will be the same; it's just that I haven't had the chance to hear it at home.

LIPSHITZ: Okay.

McGRATH: There are people who would suggest that's a possibility, though.

LIPSHITZ: Yes, but I was hoping that John presumably believes that they're not going to modify his work.

EARGLE: That's true.

LIPSHITZ: Peter, you probably have examples of recordings you want to bring up.

McGRATH: I would have to divide them into two camps: recordings that I've made that I'm most satisfied with in terms of the music and, indeed, in terms of the sound. I think the Leonard Shure recordings that I've done of the Schubert sonatas are the most satisfying musically. I've never witnessed anything more profoundly moving to me in my life musically than the time I spent with him, doing those things. The sound is very, very good but it's certainly not the most up-to-date.

EDITOR: Those are strictly analog, right?

McGRATH: Those are LP's now. We're in the process of bringing out CD's of those.

LIPSHITZ: And that is on...?

McGRATH: Audiofon. And then, probably, sonically-and also musically it's not a bad performance at all-our recent Water Musick on Harmonia Mundi with the Philharmonia Baroque Orchestra.

EARGLE: And that's the San-Francisco based group?

McGRATH: That's correct. Yes. And that's a lot of fun, and I'm very, very happy with the sound of that.

LIPSHITZ: I don't have any commercial recordings out, so there's nothing people can get hold of.

EDITOR: Well, you must be aware of some things that you like.

LIPSHITZ: I'll make some comments. If people want to experiment, there are certain labels which have a house recording policy-or had. And one can therefore be fairly certain-if one picks out recordings of a certain vintage-one can be pretty sure how they were made. For example, a lot of the early Delos recordings that were recorded by Marc Aubort have two widely space omnis-that's it. If you, for example, take almost any recording-I believe any recording-on the Nimbus label, they are ambisonically two-channel matrixed coincident-mike recordings. So they are coincident. There are a few other labels that have recordings of this specific kind. I think the Opus 3's are usually coincident recordings; I'm not positive about that.

There are some others. But if somebody wants to try one thing versus another, I would suggest for example the following comparison. I've used it in a few lectures I've given. The Equale Brass Quintet CD on Nimbus, a coincident-mike recording with the Soundfield ambisonic mike, in a stereo reduction. It is ambisonically encoded but doesn't have the slight phasiness that is introduced by the two-channel matrixing of the three-channel original.

But you are immediately aware when you listen to it of the depth-goodness, the trumpet is really much closer than the trombone, or whatever, in the recording.

Take a totally contrary recording, which is the American Brass Quintet on Delos.

EDITOR: That's an early Delos, isn't it?

LIPSHITZ: Oh yes. It's a much more reverberant acoustic as the recording was made, and you're immediately aware of the fact that it's very reverberant and ambient. But think about the question of how you can tell on the Nimbus recording that the one instrument is much further away from you than the other one. It is also very ambient, but it's not blatantly ambient. The Delos is a widely spaced omni recording with a great deal of phasiness and no real precision in imaging.

McGRATH: What you're saying is, one is a blurred image in an ambient space and the other is a sharp image in an ambient space.

LIPSHITZ: I'm saying, one is a blurred image in an ambient space, with the blur ring giving the impression of heightened ambience, and the other is a precise image in an ambient space, which doesn't sound ambient because it doesn't have the blur ring. And I would suspect that if we sat down and passed judgment on these, we would disagree as to which was the more desirable.

McGRATH: I'm not sure, but I probably would.

EARGLE: On the other hand, we would all probably agree that we're hearing the same thing. But the judgment that we attach to what we hear would be different, depending on where we all come from and what our expectations might be-and what we just happen to like.

LIPSHITZ: That Nimbus is quite spectacular, by the way, from the point of view of dynamic peaks. They never released it on LP.

EARGLE: On the other hand, there are some Nimbus piano recordings, again using ambisonics, that are so far removed from the piano that they're hard to listen to.

McGRATH: I would agree with that.

LIPSHITZ: I'll agree. Some of them sound too reverberant.

McGRATH: Empty classroom sound, as they say.

LIPSHITZ: They may be fine-I don't know-they may be fine on surround sound playback.

EARGLE: It could be. But the ears will help you lock in your...

LIPSHITZ: But I think they're not really right for stereo.

EDITOR: You haven't heard these decoded and played back that way?

LIPSHITZ: I have. I have decoded some of them ambisonically, but I have not when the system was set up properly; I have not assessed these.

McGRATH: You know, it's interesting.

Without the prejudice of the discussion that just preceded this, I would wager that if you took my most recent recording, which is of the Mozart horn concertos with the baroque horn, and you took the Nimbus recording of the same music also with the baroque horn, and if you were to present both of these recordings, let's say, to another "august" group having the same discussion, and they listened to these two recordings, they would draw the opposite conclusions as to what microphones were used. I mean, I would just wager that would be the case.

LIPSHITZ: Yeah. I don't know.

McGRATH: Theirs is very vague and very, very ambient, and it's swimmy, and you have a hard time hearing the horn, particularly when he goes up and down the scales, and it's very disjointed from the orchestra and swimmy, and mine I think, if anything, it almost-well, I wouldn't say it suffers-but it's analytically right there.

Absolutely crystal clear in focus. They used the Soundfield; I used a pair of spaced omnis.

LIPSHITZ: Of course, the tools are only tools in the hands of the user. It doesn't guarantee anything.

EARGLE: And the room is so important.

McGRATH: It's everything.

CARVER: And as John says, he challenges you upon listening to the finished product to determine which system was used.

LIPSHITZ: And you can't tell easily.

Sometimes you just can't tell.

CARVER: I have a question. The kind of recordings I like paint a picture that extends slightly beyond the speakers, with nice pinpoint imaging-I can close my eyes, and it's there, it's there, it's there, it's there... ;

CLARK: It's too wide.

"True coincident [miking] would mean there's no time-of-arrival difference in those two channels, and that I know I don't like. When I hear that, it sounds flat."

CARVER: Huh? No, no, no. I put my speakers close together.

LIPSHITZ: What angle do your speakers subtend?

CARVER: I like the soundfield to subtend about this angle. (Bob must think he is on TV.-Ed.)

LIPSHITZ: Well, that's about 60 degrees.

(Ah! Stanley to the reader's rescue.-Ed.)

CARVER: My speakers are about this angle.

EARGLE: Is this with your sonic thing in there-the hologram working and that business-or without it?

CARVER: I'm just describing the soundfield I like. Never mind how I get there.

EARGLE: Okay.

CARVER: Because I'd like to know what microphone...

McGRATH: ...best enhances that image?

CARVER: That's half of it. And then the other half is the ambient soundfield, the bloom, all of the out-of-phase stuff that I respond to so, you know, in my heart; I just love that, I love it, I love it. But it's everywhere and it goes "shhht" and it goes higher than the speakers, it goes deeper than the speakers, lower than the speakers, wider than the speakers-and I love it when it happens. And I search, and I search, and I search, and so many...

EDITOR: You like ketchup on your sound.

CARVER: No, that is not ketchup on my salad. That's just wonderful, it's wonderful...

EDITOR: Sound, I said, not salad.

McGRATH: No, that's a wonderful dressing on a good lettuce. That's what it is.

LIPSHITZ: The only way you get direct sound images that extend beyond the speakers is antiphase signals...

CARVER: No, I know how... I understand that.

LIPSHITZ: ...so you would not get that with a coincident mike arrangement unless you go beyond the in-phase pickup angle or you process it to increase the L -R signal.

You can get that with other mike arrangements.

CARVER: Too often, when I bring home a new CD and I put it on my system, I hear a flat curtain of sound strung between the two speakers. And it's distressing when I know that I've heard recordings that have this palpable depth that you talked about, where you knew that-was it the trombone?-was behind...

EARGLE: Or the horns, especially. And the horns really come out of back there because they're playing in that direction.

CARVER: Now, I understand exactly what the sound vector field has to be in my ears so that I hear that. But what's not clear to me is how you get it with microphone technique. I mean, I could map the sound field that's required to generate that impression in my mind, but if I were to make a recording, I wouldn't have a clue.

How do you get the imaging, and then superimpose on that the bloom-and make it work?

LIPSHITZ: You can't get the imaging without having a coherent pickup in the two mikes-assuming you're using two mikes. So that essentially forces coincidence, or near coincidence, if you want the imaging. If you want the bloom...

CARVER: Now near coincidence means- this is really the important part to me...

LIPSHITZ: Spaced no more than your ears are apart.

CARVER: True coincident would mean there's no time-of-arrival difference in those two channels, and that I know I don't like. When I hear that, it sounds flat.

EARGLE: I don't like it, either.

CARVER: Okay. So near coincident-we move them apart a little bit, at least a good fraction of an interaural time delay-is that right?

McGRATH: Like a little bit pregnant, right?

CARVER: Huh? Is that right? We move them apart a little bit.

LIPSHITZ: Yes. It would have to remain less than eight inches apart.

CARVER: Ah! Okay. Now we have some thing that approximates an interaural time delay, and that probably will sound very good, since...

LIPSHITZ: No.

CARVER: Why not? He says yes; you say no.

LIPSHITZ: Because you're assuming that moving the microphones apart introduces an interaural delay. It introduces an inter loudspeaker delay.

CARVER: It introduces the distance... no, no, no.

LIPSHITZ: It introduces a delay between the signals at the two reproducing loud speakers, but that doesn't mean it introduces a delay between the signals at the listener's two ears.

CARVER: No. That's right.

LIPSHITZ: And that is precisely the point I was trying to make before. It introduces phasiness at the listener's two ears.

CARVER: Well, maybe phasiness isn't so bad. Maybe we'll interpret it as some thing...

LIPSHITZ: I didn't say it's bad. I said phasiness on the direct sound is bad. I said phasiness on the ambience is nice. You see, you're caught in the following dilemma-and this is a dilemma. The only way to prevent phasiness on the direct sound to get imaging is not to do this...

CARVER: I know.

LIPSHITZ: ...the only way-not the only way, one way-of introducing an ambient airiness is to add phasiness to the direct sound and lose the imaging. What you really want is no phasiness on the direct sound and plenty of ambience.

CARVER: That's right.

LIPSHITZ: The only way to get that in a live recording is to use coincident mikes which pick up a lot of ambience and have ambience to pick up.

CARVER: Okay.

LIPSHITZ: But it must be nice ambience to pick up, which means you've got to record in a good venue that's suitable.

EARGLE: There are a lot of those, right?

McGRATH: Oh, they're common.

LIPSHITZ: There are two. (Laughter.)

CARVER: To get this straight: so, to do it, we've taken a coincident mike and made it near coincident, so they're spaced about eight inches or so apart...

EARGLE: Well, I have, but he hasn't.

CARVER: You have. Okay.

EARGLE: Because there's one aspect of having it absolutely coincident that I don't like, and that's when I wander a little bit off the median plane, there's that sudden shift.

LIPSHITZ: I'm never very aware of it. I know what you mean. Yes, I know the effect.

EARGLE: I mean, it has to be there.

CARVER: It also makes the center image too hard; it nails it too hard in the center.

LIPSHITZ: Not much, though.

CARVER: And real life isn't like that.

LIPSHITZ: But, Bob, careful! You move your mikes six inches, eight inches apart;

that's the distance we're talking, no more than that.

CARVER: Yes. That's a lot.

LIPSHITZ: It adds phasiness only in the frequency range above 1 kHz or so. About 500 Hz.

EARGLE: And the ear doesn't really hear phasiness per se up there.

CARVER: No.

LIPSHITZ: You're only sensitive to phasiness below 1 kHz, really. So you're adding it in the region where it doesn't degrade the low-frequency imaging; the sound is still essentially coincident below 500 Hz, SO you retain the low-frequency imaging;

you add some phasiness at high frequencies, which broadens and diffuses the image a little bit...

CARVER: I understand.

LIPSHITZ: ...and, well, I'm not quite sure what it does, but it makes a slight difference...

CARVER: It changes the time relation ships a little bit.

LIPSHITZ: But it's nothing at all like what you get when you move your microphones three feet or six feet or nine feet apart.

CARVER: Oh, sure. Yeah.

LIPSHITZ: That is phasiness over the whole audible spectrum range, you see.

CARVER: Now, with that, when we've taken the coincident microphones and moved them apart, have we built the complete soundstage picture, or are we only halfway there?

EARGLE: It depends on the room you're working in.

CARVER: Do we have to add the flanking microphones? Do the flanking micro phones make the...

EARGLE: The flanking microphones come under the heading of local seasoning at the hands of the master chef.

CARVER: So, you use them for two purposes: to enhance the pinpoint imaging that I was talking about...

"...the left-center-right spaced-omni approach has a lovely texture to it. I give ita B minus on imaging -not as analytical as what I like- and an A plus on texture."

McGRATH: No.

EARGLE: No.

McGRATH: No. On the contrary.

CARVER: So they’re used strictly to make the bloom?

McGRATH: To get the color of the hall.

EARGLE: To pick up a little bit more hall sound and to give a little bit of an assist to a string section that may otherwise be swamped out by high-pressure brass and percussion.

CARVER: Okay. All right. So the...

CLARK: What's the timing? You bring them back a little farther?

McGRATH: You put them on the same plane.

EARGLE: I usually set them up on the same plane.

McGRATH: I do the same. Yes.

EARGLE: And they're always a good 6 dB down, often 8 dB down.

CARVER: Uh-huh. I understand.

LIPSITZ: I still disagree with your basic statement.

CARVER: I asked a question; I didn't make a statement.

LIPSHITZ: No, no, no. You did.

CARVER: I stated what I liked.

LIPSHITZ: You said that you maintain the imaging, and now this adds bloom. I say, as soon as you've moved your micro phones a few feet apart, you have lost the imaging.

CARVER: No, no, no. Either you misunderstood me or I misspoke. I said what I like in my soundstage, and then I described it. And the question was, how do I get it?

LIPSHITZ: You called those images, though. By image, do you mean something whose location isn't a function of its pitch?

CARVER: That's right. Whose location is a function of its physical position in space.

LIPSHITZ: Physical position. All right. I don't believe you can get that with spaced microphones.

CARVER: I think that's right. I mean, I don't see how you could. You have to have the timing cues just right to produce the illusion of that sound being ten feet back or less... But how do you do it?

EARGLE: There is one area of recording that we ought not to leave unspoken about here, and that's Jack Renner's approach of three spaced omnis. And the spacing we're talking about is roughly maybe nine feet between microphones.

McGRATH: Collectively?

EARGLE: No. Eighteen feet between the outside ones. That's roughly about it.

CARVER: You say he puts the micro phones 18 feet apart?

EARGLE: That's right. But there's one in the middle.

LIPSHITZ: Mixed.

CARVER: Okay. Mixed.

EARGLE: Now, the middle mike normally is a little bit down. It's normally 4 to 6 dB down, and the way Jack normally sets that is to pull it out and then bring it in until this vast hole in the middle now has a...

McGRATH: A fill.

EARGLE: ...a fill to it. So, when you do that, you have avoided the whole middle because you've just filled it in. Okay, in terms of stereo, you're now talking about two stereo pairs. One left and center, an other one center and right, being subtended by the listener over narrower angles. And the net result of that is that left center information in the orchestra does sort of come out on the left center; right center information comes out on the right center, and you have very good localization of the middle of the orchestra, the left of the orchestra, and the right of the orchestra.

And a lot of sins of spaced microphones are alleviated by that. Now, I can normally spot that in a recording because I can shut my eyes and really lock in-the winds, for example, all sound like they're right in the middle. Normally, in my recordings, or those that are done with a coincident or quasi-coincident pair, you hear the flutes over here, oboes over here, clarinets over here, and bassoons over here-because the winds, depending on how many there are, can be seated the length of this room here.

Sixteen winds would occupy this space- and two rows deep. Anyhow, the left center-right spaced-omni approach has a lovely texture to it. The imaging is not as analytical as what I like, but it's good. I give it a B minus on imaging and an A plus on texture.

EDITOR: When I sit in the concert hall, I don't get anything better than that B-minus imaging, as you call it.

CARVER: That's right.

EDITOR: In a typical seat in the concert hall, I have a sense of left, center, right, front, back, and under the very best conditions one point in between these.

LIPSHITZ: Right. But the microphones, of course, are not as far back as you would be in the concert hall.

EDITOR: That's right.

EARGLE: They wouldn't work that far back.

LIPSHITZ: They would not work that far back. And this is, of course, stereo versus surround sound. There is no way you can put the microphones in there for stereo re cording where you'd prefer to sit in a con cert hall. If you could do a true surround sound type recording, then in principle at least you'd want to put the microphones where you like sitting, and then when you reproduce that sound you would have the same statements that you could make about it. ] can't tell exactly where anybody is in the orchestra, and there's the hail ambience all around me.

EDITOR: But why aren't you happy with that kind of presentation in ordinary stereo?

LIPSHITZ: Because we're not reproducing the ambience around you. It's all coming from the front.

CARVER: That's right. Now...

EDITOR: Wait a minute. Do you mean to say that under those circumstances...

LIPSHITZ: You've got to move the micro phones closer.

EDITOR: ...you need more precise information in order to get an illusion of reality?

LIPSHITZ: Yes. The reason why I reject the argument that imaging is an overrated aspect-because you don't normally hear the location with that kind of precision from where you normally sit in a con cert-is that the recording is made from a much closer location, from which there would be no doubt, if you sat there, where the instruments were.

EARGLE: Well, the conductor himself normally...

LIPSHITZ: Has no trouble.

EARGLE: ...has no trouble telling where everybody is.

LIPSHITZ: Yes. You know, in the third, so-and-so was off key on such and such.

CLARK: Why should you be trying to duplicate what the mikes hear?

LIPSHITZ: Because, if you can recreate accurately a replica of the original perception that you would have gotten from there...

CARVER: If you were close...

LIPSHITZ: ...then you can do other things.

CARVER: Yes, that's right.

LIPSHITZ: Then you can do other things.

Then you can make it wrong-in many ways. But if you can't make it right in any way, you're worse off.

McGRATH: I agree with that. That's a profound thing for me, and I agree with that a hundred percent. Yes. That cuts to the core of the illusion.

CARVER: Let me give you an example, an example of what Stanley just said, in actual practice.

LIPSHITZ: You can control the illusion.

McGRATH: That's exactly right.

CARVER: During the development of sonic holography, one of the things I had to undo was the spatial distortion associated with loudspeaker reproduction. During its development I worked very hard, and I could do complete ventriloquism. I had a person walking around banging castanets together, and I could bring the castanets right up to my right ear, and I could have her whisper in my right ear.

LIPSHITZ: How did you mike this?

CARVER: With two speakers. And...

LIPSHITZ: No, but how did you mike it?

CARVER: I'll tell you in a minute. Another thing I could do was-I made a recording in which I fly combat model airplanes, where we have two model airplanes buzzing around the sky, trying to outmaneuver the other and cut off this ribbon that he tows, and the combined speed is about 240 miles an hour. It's really exciting. You know you've been alive when you've flown a combat match. (General tittering and sotto voce wisecracking.)

LIPSHITZ: What size are these?

CARVER: About this big. (There you go again, Bob.-Ed.) The engines are two horsepower.

LIPSHITZ: Three feet long. (Thank you, Stanley.-Ed.) Okay. (Wisecracks continue in the background.)

"...that's a very powerful demonstration that you've correctly undistorted the spatial distortion, if you can ...walk a sound around the room anywhere at will." EE ear]

CARVER: And when the airplanes do a loop-how round is a loop?-well, it's a large, round loop that goes up, overhead almost, and then back down again. That was the most difficult to achieve. I'd bring the recording inside my house and I would play it back through my processing, and the poor little loop would be squashed- ts00-tsoo, like this, pfoom, pfoom, pfoom

-so0 I could never, and I haven't yet succeeded to this day, to make my round loop in the house; it's always squatted, squished, but [ can, horizontally, have a person come around and whisper in my ear.

LIPSHITZ: How high is your ceiling? That's the problem! You should raise your ceiling. (Laughs.)

CARVER: Believe me, I've done it; I've put my speakers outside; I mean, I've gone through the whole thing...

LIPSHITZ: I know.

CARVER: So... I've sort of forgotten what I was...

EARGLE: Well, you can answer a question about how you recorded this-when a woman came up and whispered in your ear. You were going to tell us later.

CARVER: When she did that, I had her come right up to one microphone, talk into one microphone.

LIPSHITZ: But what were you using? Co incident recording? Spaced mikes? What mike technique were you using?

CARVER: Oh. I had a piece of cardboard, a large piece of cardboard; I glued a pil low, a large pillow, on one side, and I had

-this was, now remember, 10 or 12 years ago-two Nakamichi microphones glued on either side of the pillow, and I had cut some holes in my large piece of cardboard.

LIPSHITZ: These were omnis? Well, it doesn't really matter very much. Okay, so you had a baffle between your two mikes.

CARVER: Uh-huh. I had a baffle between my two mikes.

EDITOR: This is Bob all over. I love it.

CLARK: It's great.

EDITOR: This has your signature.

CARVER: Yeah.

LIPSHITZ: Well, you know, there are people who record with that sort of arrangement. I mean, not the pillow and the holes...

McGRATH: Yes. The Jecklin baffle be tween omnis. What do you think about that? Does that give us some coincidence with omni characteristics?

LIPSHITZ: No, it does not give you coincidence. No way. It is similar to those weird baffles that some of the people with PZM's are using, where they have these PZM's on opposite sides of a piece of Plexiglas. Because they are just a quarter of an inch apart, the Plexiglas being a quarter of an inch thick, they say this is co incident recording. Not at all. The acoustic path length from the one to the other is many feet. It's a baffling; it introduces level differences as a function of angle and frequency due to the diffraction effects, so that you produce level differences and time differences at the mikes. I haven't tried it.

CARVER: I remember what my point was in that story. It was, if we can learn to do that kind of ventriloquism, then perhaps we have a handle on how to undo the spatial distortions in the recordings. Because that's a very powerful demonstration that you've correctly undistorted the spatial distortion, if you can completely do ventriloquism and walk a sound around the room anywhere at will. Except I couldn't in a million years take that and record an orchestra and make the orchestra come out sounding real. It's a puzzle to me just what should be done-how to get the ambience component of the soundfield and the direct sounds and have it all work out. Now what I've heard so far-and I'm learning about making recordings-is two nearly spaced, nearly coincident microphones, two flanking microphones, and there are more microphones in the back sometimes?

EARGLE: If you need them.

CARVER: If you need them. Okay. I understand. I got it. And the two flanking microphones really do contribute mostly to the ambient soundfield, the out-of-phase components of the soundfield.

EARGLE: Probably. Yes.

LIPSHITZ: Well, let me give you an example of one recording which I did of a brass group in a large church. The recording is extremely reverberant, but the imaging is still quite precise-because it's a coincident recording. But the venue is very reverberant. I doubt whether people would say that it's not reverberant enough if they heard it. I think you'd probably feel that it was reasonable-or at least you might agree that it was a reasonable balance.

Most venues are not reverberant enough for a coincident-type recording to sound reverberant.

CARVER: That's right! And that's why they're so disappointing sometimes.

LIPSHITZ: That doesn't mean it's not reverberant; that Nimbus Equale Brass CD I suggested to you does not sound reverberant. But when you ask yourself what causes you to be able to tell that there are enormous differences in distance between you and the instrumentalists, it is the direct/reverb ratio in the recording. That's the only way you can tell that so-and-so is closer than so-and-so. So it is reverberant.

CARVER: So the clue or the cue that allows you to understand that something is soft and close instead of loud and far is the...

LIPSHITZ: Direct/reverb ratio.

CARVER: ...cue or clue associated with the ratio between the direct sound and the reverberant sound.

LIPSHITZ: Yes. Now normally, in a con cert hall, you don't sit down and say, when the orchestra plays, oh what wonderful acoustics. You're not normally aware of the reverberant sound. If you are aware, it usually means something is wrong.

EARGLE: There's too much of it. Yes.

LIPSHITZ: Too much, or nasty things, echoes, or some features that are not nice.

So, the things you're consciously praising about recordings are things whose pres ence you are consciously aware of. I believe that in many ways they're wrong. But I understand the desire for them. What I'm reluctant to give up in order to achieve that sort of sense of involvement in the sound is the precision of the direct sound- because in no natural situation do you have phasiness on direct sound. You can't. And that's what worries me.

CARVER: No, there shouldn't be any phasiness on direct sound. I agree. I have one more question, something I just want to understand. If I had a hard L + R signal, mono signal, between my speakers, and it was a really hard signal, and imagine that I had an echo associated with it but it was completely in an L -R channel, and I had a control so that I can turn off the L -R or turn up the L - R atwill-and let's make it a toy duck quacking, a mechanical toy duck going quack, quack, quack, quack- and there were some echoes associated with it, and if I began to turn the echo up, my subjective perception would be that the duck would begin to move further away?

LIPSHITZ: Oh yes. With a Blumlein recording, coincident figure eights at 90°, and this means really good figure eights- the Schoeps figure eights, by the way, I've always used at 85°, the criterion being the fact that they get slightly sharper at high frequencies, and you've got to trade off a solid central image versus one that's sort of receding a bit and giving you a slight hole in the middle.

McGRATH: How far back do you position the mike? Do you allow the spacing of the musicians to exceed a 90° angle?

LIPSHITZ: Oh, I wouldn't mind because that means the image will go slightly beyond the loudspeakers. That's fine; you can go, you know, 10° more, either side, and within a certain range of antiphase signals in the two channels you get imaging beyond speakers-then it starts becoming oppressive phasiness on the ears.

EDITOR: You haven't finished your thought, Stanley. What about the Blumlein...?

CARVER: What about my duck?

LIPSHITZ: 1 haven't finished my point.

With a good Blumlein recording, with let's say textbook-perfect figure-eight microphones, your volume control is a distance

control to a very large extent.

CARVER: Okay. I believe that.

LIPSHITZ: And you can try that. As you turn the thing up, you'll find that it's almost as if you were moving closer to the sound.

CARVER: Well, in this case, as I turn the L -R up, I move further away.

LIPSHITZ: In fact, Peter Walker is per haps the first person who talked about that, if you remember the instruction books that came with his preamplifiers...

CARVER: Yes, I remember that.

LIPSHITZ: ...or perhaps even come with the current ones; I'm not sure.

"I would like somebody to tell me why the microphoning has to be more location specific, more image-specific, than what we are accustomed to in the concert hall."

McGRATH: The most important control.

LIPSHITZ: He would say that there is generally a correct volume setting.

CARVER: That's right, that's most realistic.

LIPSHITZ: And that is true. If you turn it too high, the lateral perspective is wrong for the apparent distance perspective. If you have it too low, again it's too wide for the apparent distance. There's an "about right"-and if you did the recording your self, you either agree with that or you can fool yourself into believing that you hear what you know is about the correct perspective.

EDITOR: Does that phenomenon still exist with spaced microphones?

LIPSHITZ: No.

CARVER: The effect, though, of volume? The more out-of-phase level, the more distant it becomes...

LIPSHITZ: Not in the sense that I'm talking about it, though. You don't think of the volume control as being that kind of distance perspective control, that things click in just correctly when you set it.

EDITOR: Is there still one correct volume for spaced microphones?

CARVER: Yes.

LIPSHITZ: I don't think so. I don't think SO.

EDITOR: You don't think so.

CARVER: My final question with my duck, Stanley. If I begin turning my LL - R signal up, the duck begins to move further away. Now, obviously I would think it's a louder duck, so in my mind my mechanical duck would move further away and be quacking louder. Obviously, if he is moving away, and I certainly haven't changed the volume of the L + R, then it must be that he is quacking louder. So it's sort of the acoustic analogue of what artists have done for years with railroad tracks. Perspective.

EARGLE: Well, they show an object the same size against a perspective. Yes. And the foreground object is only maybe three units high here. But the room perspective looks like this, you see. So that same object, the height back here...

LIPSHITZ: Looks enormous.

EARGLE: ...looks much larger.

EDITOR: Nobody can see your drawing, John, on the tape.

CARVER: Exhibit B is a classic drawing of an artist's perspective with telephone poles and a street.

EDITOR: Okay. I would still like one of you to tell me-maybe Stanley, since I think you're the most vociferous advocate of this particular approach-I would like somebody to tell me why the microphoning has to be more location-specific, more image-specific, than what we are accustomed to in the concert hall.

LIPSHITZ: Because you have no sense of sight when you're listening.

CARVER: You have to make it come alive.

LIPSHITZ: You're listening in an acoustic which is not a concert hall.

EDITOR: Suppose I'm blind.

CLARK: Stanley said one of the reasons was because the microphones are closer and the microphones can hear those positions better. I don't buy that, but I think that's what he said.

EDITOR: I know he said that, and that...

EARGLE: Well, okay. Can I offer a way around this? Let's say we take anybody's modern recording of an orchestra, where there is good hard left and hard right information, either by virtue of being on the major axis of a figure eight and the null of a figure eight and that sort of thing or whatever-whatever technique you might use. Who in this room would take that recording and take something that we used to call a blend control and pan that in, to make it sound the width which you might hear in Row M?

CARVER: Nobody.

EARGLE: Not a soul!

CARVER: End of argument. That's it.

That's true.

EDITOR: This is the best-seat-in-the house argument. Right?

EARGLE: Yes, yes, yes.

CARVER: Yes. You get a better chance.

EDITOR: But there's more specificity of localization and imaging in some of these recordings than in the best seat in the house.

LIPSHITZ: Well, no, but John's point is this-that if you sit at a distance such that you have a 60° subtended angle for the orchestra, you're not in the best seat of the house; you're much too far forwards. What he is saying is, if you sat that close to the orchestra, the specificity would be pretty good. I think that's what he is saying.

EARGLE: Uh-huh. Yeah.

McGRATH: I'd like to hang from my light bar that I put my mikes on in the concert, sometime; I'd just like to hang from there, just to see what it's like. The sound must be extraordinary!

LIPSHITZ: The sound the conductor gets must really be quite spectacular.

EARGLE: Oh, it is.

EDITOR: And deafening.

EARGLE: I've been out there. I've listened from the podium a number of times, you know.

McGRATH: It's extraordinary.

EARGLE: Anyhow, can I register a very strong complaint against a lot of pop/rock product that's being generated today?

McGRATH: One? Only one?

EARGLE: Well, the main one is this-that all musical values aside, some of the stuff that is mixed down from 24, 32, and 48 tracks is so center-heavy that it just drives me crazy.

CARVER: I know.

EARGLE: I mean, my God, you're paying for stereo-why don't you get stereo for Christ's sake! You know, it's all piled in the middle so much of the time, and maybe you'll hear a lead guitar over here part of the time and something over here. And if you look at it on the scope, it's just really sort of a 45° blob. There's almost no original stereo recording in there; it's all panned mono.

LIPSHITZ: Mono. Yes.

EARGLE: Maybe the output of a Lexicon is in...

CLARK: John, I don't know how you can say that. ['ve recently searched for a bunch of recordings that were extremely phasy, had very little correlation between them, and I had no trouble finding them-and they all had to be pop recordings.

EARGLE: Were they new recordings?

CLARK: Yeah, like Tiffany, for instance

-big, swishy, swirly things, and a little voice in the middle...

EARGLE: Well, good for them.

CARVER: They're getting the hang of it now.

EARGLE: It's about time.

CARVER: The last year there've been a lot them that have come out.

CLARK: They vary all over the place.

EDITOR: Okay. Name a few intelligently recorded rock/pop albums.

EARGLE: I decline because I really don't know. I mean I'm not that familiar with the scene.

McGRATH: Dire Straits?

CLARK: Yes, Dire Straits is another whole thing.

McGRATH: Yes, it's wonderful.

EARGLE: Now, there are a few jazz things that I've done. I wouldn't do what I do if I didn't like it, basically, but that doesn't mean that I praise everything that I've done; Lord knows, we've all made mistakes. But to me the real mark of a good pop or jazz recording-a natural acoustical recording-is that it's a very nice mixture of real stereo recording plus close-panned images when you need them.

You have to mike a soloist at a distance of about 18 inches to two feet.

CARVER: That's where you put the microphone for soloists?

EARGLE: And a bass has to be miked; a guitar has to be miked; otherwise, if you rely upon natural perspectives, they're going to be drowned out by the drums and everything else. But the drums can be miked in stereo, beautifully; a piano can be miked in stereo...

CARVER: Oh, I hate pianos miked in stereo. I hate it; I just hate it.

LIPSHITZ: When you say miked in stereo-what is it that you hate about a stereo piano? You mean you don't like a 60° wide piano?

CARVER: One mike stuck on the treble end and one mike stuck on the bass end- that's what I hate.

EARGLE: Well, you don't quite do that.

CARVER: Okay. Stereo back away is fine.

EARGLE: A piano is normally panned left to center and not from left to right...

CARVER: That's fine.

EARGLE: ...and the drums are normally panned from center to right. And you get a very nice cohesion between them.

CARVER: John, the microphone is 18 inches away from the soloist-do you do that also with opera singers?

"...if you had a digital recording that had hiss while the music was playing and silence between notes, you've got noise modulation...

something not being right."

EARGLE: No!

CARVER: No? Much further out-okay.

Just pop.

EARGLE: Well, I've never done an opera, but I wouldn't be that close.

EDITOR: Let's talk about some of your other equipment in your recording sessions-your digital processors, consoles if anything-what is the current state of the art past the microphone?

EARGLE: For classical-unless you're doing actual panning and need a pan pot- all you need is a summing amplifier on the output of a string of mike preamps, and you need one for the left and one for the right, as simple as you can get it. The only filtering, the only EQ I ever do on a classical date-ever have done-is to roll off the extreme low end if there's a discrete rumble or something like that-and that's a bad room.

McGRATH: Air-conditioning noise, blow Fs

EARGLE: Yes. But otherwise never any putzing around at the high end.

McGRATH: I hold a similar philosophy. I basically don't even have a summing net work, though I've been using it now on my concert recordings, but basically I've just been using-specifically-a Jensen two channel thing...

EARGLE: It's a lovely preamp.

McGRATH: The mikes go into it, and the signal of that goes right on out into the tape recorder.

EDITOR: Those are JE-990's, right?

McGRATH: Yes. And that's it.

EDITOR: Well, what about digital processors? There's been a lot of talk about the Colossus and similar stuff. Any comments on one kind of A-to-D conversion against another?

LIPSHITZ: Well, I think the differences are questions of accuracy. If you want an accurate A-to-D conversion, then make sure that the system is either properly dithered or there's enough noise in the mike feeds to do that.

EDITOR: There's a lot of inaccurate stuff around, isn't there?

LIPSHITZ: Oh, yes, the A-to-D conversion is much more difficult than the D-to A conversion.

CLARK: Before you throw out 16 bits, make sure you've got all 16, right? Before you go to 18, make sure you've got your 16.

LIPSHITZ: Well, 16 good ones is better than 18 bad ones.

EDITOR: That's for sure.

LIPSHITZ: And the same is true of the D to-A's, mind you, where the number of bits is proliferating, and that doesn't mean they're better. But that being the case, I think we're talking engineering details, which in principle I don't believe have significant effects on the sound.

EARGLE: They're all vectoring in on the same solution. I mean, if they're all perfected, whatever they're doing, the answer ought to be the same.

LIPSHITZ: Yes. You've got single-spot CD players and three-spot laser optical systems-that's neither here nor there.

EDITOR: So you don't go for this "hey, man, of course it's better, it was made with the Colossus, man"? What do you think is better?

EARGLE: The only difference between the current processors right now would be the filtering, the monotonicity of the de vices and so forth, and the dithering. Normally, there's enough dithering caused by the self-noise level of microphones.

LIPSHITZ: For classical recordings, I think invariably. The digital system's noise floor is below the mike feeds. I mean, you could try that on PCM-F1 type processors, where you can switch from 14 to 16 bits.

EDITOR: There is hiss on nearly all of my CD's. Where does it come from? It comes from the mikes.

LIPSHITZ: Yes, of course. But there's hiss everywhere. And there ought to be. Be cause, quite honestly, if you had a digital recording that had hiss while the music was playing and silence between notes, you've got noise modulation. I mean, that's an indication of something not being right. There ought to be a steady, thermal, white-noise floor in almost every recording.

EDITOR: But not so high in level that you're aware of it when you're not listening for it.

LIPSHITZ: Well, the mike noise may have been rather high. But anyhow, the point I was making about the PCM-F1 type processors is apropos of this question. You can switch from 14 to 16 bits. The dither is just done by the inherent noise in the circuit, and the noise floor doesn't change when you switch the number of bits. It's properly dithered at 16-slightly over-dithered--and under-dithered at 14. If you ask me the question which should you use in a recording, I'd say to you this: Have your mike faders up at the normal level at which you’ll use them during the recording, crank up your monitoring level, listen to the sound, and switch between 14 and 16 bits. And if you cannot hear any change in the noise floor, the digital noise floor is dominated by the noise coming in on your mikes-use the 14-bit mode because that gives you greater error protection and it's being well dithered by the incoming signal on the mike lead. If you hear any change in the noise floor when you switch from 14 to 16, the digital system's floor is dominating the mike feeds-use the 16-bit mode because now you need the dither in the recorder to dither the system properly; the mikes' hiss is not going to do it. But here we're talking about noise that's close to the digital floor; what Peter just mentioned there was a noise floor in the recording that's way above the 16-bit floor. Well, that's being limited by something that's not the digital recorder, unless there's something vastly wrong with the digital recorder.

McGRATH: Generally, that's mike preamps-and microphone self-noise, too, to some extent.

EDITOR: What about consoles? There have been a lot of bad consoles.

EARGLE: Well, of course, the minute you speak of the necessity of a console, you're talking about pop recording because for most classical you're really looking at, at the very most, two sets of summing amplifiers-a whole bunch of preamps feeding into one sum and a whole bunch of preamps feeding into another in the other channel. But the trouble with consoles is that console builders play with architecture the way I noodle an eclectic factory in Los Angeles. Anytime they want to change direction-I go around the corner-they put in a summing junction to go from here to there. And they overdo it-there are so many active devices in a console that it's almost frightening. And I guess the reason why that happens is that once you have enough of these, you can double the num ber and get a certain decrement in performance; then you double that entire number for the same decrement-so the more you have, the more you can use with impunity.

At least it seems that way. The real test of a console is to have a hard-wire bypass around it-around as much of the architecture as you intend to use-and switch it.

The way I've done this in the past is to take a little L pad with a 200-ohm light across the middle to simulate the source impedance of a microphone and to feed into this a stereo program and to bring the level down so it really corresponds to what a good capacitor microphone might be giving me. The figure I take is 94 dB SPL, in other words one Pascal, to give you 7 mV out. And I simply run that through the board, bring it up to zero level, and then have a path around the board. Then take a good preamp and have both of these coming in, set levels, and then switch between the two-put headphones on or speakers, whatever you want to listen to at that point. If you hear a difference when you do that, that's not attributable to level, then you've got a problem.

CARVER: Does that happen often, the problem part of it?

EARGLE: Well, the answer is that I've found very few consoles, at least using the part of the architecture that I normally use, which is very, very simple.

McGRATH: Okay, that's a big difference.

EARGLE: I try to bypass all the monitor section, the mark and mix, and all that bullshit; I try to avoid that.

McGRATH: So you're really barely using it.

EARGLE: That's right. I'm using mainly the mike preamps, up to a pan pot.

LIPSHITZ: I don't use consoles...

EARGLE: You don't have to use them...

LIPSHITZ: ...because I have very simple recording things; I built some of my own mike preamps and so on...

EARGLE: But most of the problems are the ones I've run across.

"The real test of a console is to have a hard-wire bypass around it-around as much of the architecture as you intend to use-and switch it."

LIPSHITZ: ...but I would do exactly that.

Take it out as soon as you can.

EARGLE: As soon as you can. And by pass as much as you can internally. But when I don't hear a difference in the sonic quality, what I do is turn off the music and crank the gain way up, and then listen to the noise of the system. There are a lot of consoles that are prone to local RF problems, TV sync coming in, and things like that, that will drive you crazy. There are some consoles that will take long micro phone cables coming in, and you'll have a problem. The same cable on another con sole won't have a problem. You disconnect the cable-the problem goes away.

You're looking at electrically unbalanced inputs or transformer inputs. There's nothing wrong with a good transformer input, I'll tell you that much.

McGRATH: It's just good engineering versus bad engineering.

CARVER: Aren't the transformer inputs the best?

LIPSHITZ: There are clever transformer inputs, too, you know. There's an idea that I think Peter Baxandall came up with, for a transformer input with feedback around the transformer to keep it operating in the zero current mode.

CLARK: Barry Blesser did that years ago.

LIPSHITZ: So that it can have zero distortion, no magnetic saturation things, and frequency response to as low a frequency as you like-you know, hundredths of a hertz.

CLARK: A question. You mentioned some allegedly bad consoles. How did you verify, or how does anybody verify, that they're bad? Now, I just heard what John said about RF problems and noise floors and noise quality and so forth-is that what we're talking about? In that case I can agree with it; there are differences. But I have designed quite a number of consoles--I used to work for a console manufacturer and designed them, and they had many, many parts, not so many in those days--but what is a bad console and how do you know? What is this bad sound? Is it bad imaging or something big like that? In other words, I think a lot of them are given a bad rap without really testing them.

EARGLE: Like so many things in this business, you know, like anything large and expensive-the first thing you do is go up to it and kick it, or pick it up, or touch it; you feel how cold it is to make sure it's really metal and not plastic, and things like that. In other words, a lot of it has to do with the quality of the switches, the feel of the switches-the tactile qualities, as opposed to the audible qualities.

CLARK: Well, that certainly makes my point.

EARGLE: The point is that the word bad is defined by all of these things. I mean a switch that will put a click in the signal when you engage it or disengage it-some will, some won't, it depends on how they're made-some consoles you switch in the middle of the recording and you'll have a level shift going out or a click in your output; others you won't...

CLARK: But these audiophile recording companies-I like Telarc-are making a big deal of what kind of console they use and what kind of op amps are in it and all of these exotic things...

EDITOR: And what kind of wire in it...

CLARK: ...and what kind of wire in it- not seemingly the practical kinds of things you're talking about. And also people say, as you have, there's so much in the signal path. Now, there is a topology or call it a configuration of mixing buses, when you have a large number of inputs, where you can do separate summing junctions summing into other ones and come out with less noise than feeding them all into one summing junction-in which case I would say it's clear-cut that you can use more parts and achieve better performance in that area. You go through a hundred of these console-grade op amps and you come up with less distortion, far less distortion, than some of the most highly re viewed tube-type audiophile amplifiers. So where's this badness? Have we actually found it, or are we just sure they are bad?

EDITOR: Well, I'm not an active practitioner, as you know. When I said "bad con soles," this was hearsay, and the hearsay was based on early op amps that were apparently of less good performance than the more recent ones.

McGRATH: Would you agree with that?

CLARK: Twenty years ago, they used op amps with a slew rate of 0.5V/us, and in a system that has a nominal operating level of 1.23 V [1 to 3 V?-Ed.], that can under extreme conditions cause a problem.

McGRATH: But if those extreme conditions were not met, would said console using those op amps be less good than a current one, designed today?

CLARK: Yes.

McGRATH: It would be less good-or as good?

CLARK: Less good.

McGRATH: Why?

CLARK: Because you could come up with a worst-case scenario instrument, a real instrument with fast rise times or an electronic instrument plugged into it that would exceed the 0.5V/us slew rate of these 20-year old op amps.

McGRATH: But assuming that you're simply implementing it with just micro phones, recording music. Would it sound less good than a current, contemporary design? And if so, why?

CLARK: My belief is that with distant microphones, classical miking techniques such as you talk about, you probably would not be able to hear the difference using these 0.5V/us op amps.

LIPSHITZ: The current op amps are better, I think, in all measurable respects. Like the output stages of the best ones are quite low in crossover distortion, without negative feedback, so that the nonlinearities inherently, I think, are lower. But your point, Peter, is that these would not be stressed, probably, under normal conditions-but the modern ones would also certainly be quieter, probably less prone to RFI, especially the FET input types. That must be a major headache in consoles.

CLARK: The better old ones, the kinds that I designed, were just as good as modern ones in noise and RF and distortion.

They used discrete-component op amps with very wide bandwidth products.

LIPSHITZ: Ah. Okay. I was automatically thinking integrated op amps, I guess.

CLARK: Well, you know-what's bad? I take the worst, which is 20-year old op amp consoles-I don't want to make a case for those. I don't like those at all.

Those weren't the kind that I designed.

EDITOR: Aren't they still in use, though, quite widely?

CLARK: Are they? Audio Designs-was that who made some of those?

EARGLE: I don't remember.

LIPSHITZ: You mean the consoles using 741's?

CLARK: Yeah, and 1558's and that sort of thing.

LIPSHITZ: I don't know whether they're still in use. They probably are in some places, but they're not being used by any of us, and I doubt whether any of the recordings by any of the companies that...

CLARK: I'd rather keep them out of the conversation; I don't wish to defend them.

EDITOR: Let me ask you this. Let's take a console that you totally approve of-that there's nothing wrong with according to your lights.

CLARK: Okay.

EDITOR: And you put every control in neutral, so that the console isn't doing any thing except passing the signal through its amplification stages-but not altering the signal, at least not deliberately. And then you bypass it in the manner that John just mentioned. Would you say that it would pass that test?

CLARK: If you turn the gain up, or had a condition where you could hear the noise, it undoubtedly would have more noise than a bypass, almost by definition. If the noise issue was sufficiently low in level, I don't think that in a scientific test you would be able to hear a difference between the by pass and dozens of amplifiers in a console.

EDITOR: So the so-called minimal con sole recording or no-console recording, according to you, is a tweako fantasy?

CLARK: In my opinion, yes. I think all of this discussion about where to put the microphones is decidedly not a tweako fantasy-that's the real thing.

EDITOR: Oh, nobody even suggested that.

LIPSHITZ: Peter, I think, look-John was saying a little while ago that he would take a signal out as early as possible, use as few stages as he needed to use...

EARGLE: Well, I think that's good engineering practice in any discipline.

LIPSHITZ: ...and I think I would do the same. Not because I believe that the other

"So the so-called minimal-console recording or no-console recording, according to you, is a tweako fantasy?"

"In my opinion, yes

stages make audible degradations; they will add noise, even if it's not audible noise-there's just no point in going through them unnecessarily.

McGRATH: It's a lot cheaper to do with out them.

LIPSHITZ: And if you're buying, you wouldn't buy the console with those features if you didn't need them. I think there's a difference between your question, whether one could demonstrate that these were degrading the sound, or whether one therefore, being unable to demonstrate that, must necessarily use them because they do not degrade the sound.

EDITOR: Nobody is saying that.

LIPSHITZ: I agree with Dave. I suspect that many of the cases where people have criticized consoles are probably totally unwarranted in the way they were being used, but they were not consoles that are as good as ones that we can make nowadays.

McGRATH: But there is a pattern. Even if one accepts the premise that they're not audible, the evidence does speak to this- and perhaps it's just a function of a consistency of philosophy-that people who are inclined to position microphones carefully and use the proper microphones and per haps minimize their number are also people who are generally inclined not to have a lot of miscellaneous electronic stuff. So that it naturally evolves that people put the credit where it really isn't necessarily due; where the real credit is that they're just doing the basic things correctly and leaving all of the crap out of the picture.

EARGLE: There's a lot be said for operating as simply as you can, if it's consistent with your miking philosophy. Because it's less stuff to schlepp, believe me, and it's less stuff to go wrong.

LIPSHITZ: Yes.

EDITOR: But there's this broader issue here; consoles are just one example. I mean the prejudice on the part of purists against a multiplicity of active stages in the signal path. The philosophy of--let's remove from the signal path as many active stages as possible.

LIPSHITZ: But I suspect that if you put out an audiophile recording saying, you know, made with custom-designed micro phone preamplifiers feeding directly into custom-modified blah-blah-blah...

EARGLE: That's a record-jacket hype that's been around for 30 years.

LIPSHITZ: ...and you put out the same re cording, identical disc, with a jacket which says, recorded on an ABC console fed into a factory-standard XYZ digital recorder...

CARVER: Forty-seven op amps.

LIPSHITZ: ...most people would find the latter decidedly inferior in sound to the former, even though they're identical, in other words the same disc.

EDITOR: All right. I can tell in what direction we're headed. (Laughter.) And we'll dig into that.

EARGLE: I remember, once at RCA back in the old days, back in the vacuum-tube days, I once counted the simplest path from a microphone to a stereo master lacquer-the path of that single input from that microphone, how many transformers it had to go through. And it was well up in the 20's, I mean something like 23 or 24.

By the time it went through the console, and the console combining networks, and then into the next thing-the tape machine didn't have one, but the remix board had several, and the lacquer cutting chain had its share of transformers-and yet the end product wasn't that bad. I mean they were all fairly good transformers, let's face it- and you can make them.

LIPSHITZ: You can make very fine trans formers, absolutely.

EARGLE: Yes, superb.

LIPSHITZ: They're better in many ways than active circuitry...

EARGLE: Yes, they are.

LIPSHITZ: ...because they don't have the ground problems and common-mode things. But there's one point that might be worth making, and that is that the addition al circuitry in the console, which may not be audibly degrading sound in the sort of aspects we've been referring to now, will probably be passing it through a large number of coupling capacitors. Now, I'm not talking about that because I'm saying or implying that there's something about a capacitor that is nasty; I'm not implying that at all. I'm saying that there is probably a large number of high-pass filters. They may all be set at 5 Hz, but there is a very large accumulated phase shift that is introduced in the lower part of the audio band, getting up to a few hundred Hz, by those high-pass filters, those coupling circuits.

And there is evidence, published in the Journal of the Audio Engineering Society, some experiments that Laurie Fincham of KEF did, for example, that those things potentially are audible; those phase shifts are audible. Maybe we need more people to do some more experiments, down at the bottom end. Phase effects normally are very subtle, but the implications of this paper are that, yes, we want high-pass filters-you don't want to be feeding 0.5 Hz and noise and junk and room pressure changes to your subwoofers-but perhaps we need linear-phase high-pass filters that will roll off below 5 Hz or 10 or 20 or whatever but maintain the phase response as linear phase down to very low frequencies, so that the phase distortion isn't a factor. You see, a high-pass is much worse than a low-pass. Almost any low-pass filter is essentially linear-phase in its passband.

A linear-phase filter is not time-dispersing; it's waveform-preserving, almost totally.

And it doesn't matter whether the low-pass filter accumulates quite a large phase shift; it's linear phase shift, pure time delay, a few microseconds or milliseconds-totally non-distorting. But a high-pass filter with the same amount of phase shift, because it's happening in the inverse direction as a function of frequency, is very phase distorting, very waveform-distorting, and potentially audibly so.

CLARK: I've read that KEF thing; in fact I was there when he gave it, and I've studied that. I think that the audibility of that phase shift at low frequencies possibly is better looked at as an audibility of the time dispersal rather than the degrees of phase shift...

LIPSHITZ: The group delay.

CLARK: Yes, the group delay. Because 45° or 90° or some such small amount of phase shift at 20 Hz amounts to many milliseconds. And it's not too surprising when you start talking about the milliseconds involved that it might be audible. Again, I think group delay wins out as a concept over phase shift in terms of what's audible and what isn't.

LIPSHITZ: Well, as long as you know how to interpret the numbers, both of them are ways of quantifying it. An interesting point, of course, is that making linear phase analog circuits, or phase-corrected analog circuits, is relatively expensive.

Digitally it's very easy to do these things.

EDITOR: Dean Jensen put out a pretty good paper on what you just said, Dave- what kind of phase shift is relevant and what kind is not. Basically that flat group delay is not audible.

LIPSHITZ: With one proviso, one qualifying comment. Yes, you're talking about meaningful and relevant parameters, or something, in high-frequency phase distortion...

EDITOR: I have it around somewhere.

LIPSHITZ: Yes, I know the paper you mean. And, yes, that is all perfectly correct. The implication, however, that these high-frequency phase shifts are audible, which is in that paper, is quite unsubstantiated.

CLARK: That's what I thought. He always implies that.

LIPSHITZ: Yes, he implies that, and I've asked him, and he can't substantiate it and won't even try to. He'll say, for example, I know you disagree with me on that and don't want to discuss that issue. That's all very well and good, but it doesn't prove anything one way or the other. (The recent tragic death of Dean Jensen leaves this controversy without a conclusion.-Ed.) All the evidence is that phase shifts above a few kHz-phase shifts of modest amounts, which is what we're talking about-are inaudible. But phase shifts of modest amounts below 500 Hz can be audible-and there's indication that per haps significantly so at low frequencies- and need to be thought about. Something of interest-I haven't had a chance to listen to the disc yet-but there's a very interesting CD that has just been put out by the Acoustical Society of America. It's available for 20 dollars, 17 each if you buy five or more-that's for members at least-and it's audio demonstration tests, 39 tracks on this disc. It's been produced in conjunction with Philips/Polygram, the Institute for Perception Research in the

----

"All the evidence is that phase shifts [of modest amounts] above a few kHz are inaudible. But phase shifts of modest amounts below 500 Hz can be audible..."

----

Netherlands, and Russing at--is it Northern Illinois or whatever?-and they have digitally regenerated all the Harvard psychoacoustic tests, on tapes released by Harvard in "78 and rapidly distributed- none of them are available. They digitally generated all these things, and it contains demonstrations of many psychoacoustic phenomena. And that could be extremely interesting to many people; you will find demonstrations of interchannel level differences, interchannel time differences, the audibility of reverberation; for example they have an anechoic and a reverberant voice-I think it's voice-track, where you're aware that this is a reverberant voice-it's in a room, it's not an anechoic voice-but you're not normally very aware of the reverb. And then they play the voice backwards.

EARGLE: And you hear the tail coming in.

LIPSHITZ: Yes, the tail comes in first.

You're absolutely aware of it then-this is reverb there, absolutely.

CLARK: I have that disc. I've listened to about the first half of it; it's fascinating. A lot of it is pretty dry; it's not really entertaining; it's instructional material.

LIPSHITZ: But if people are interested in some of these psychoacoustic things, there is a 93-page booklet that comes with it.

EDITOR: Acoustical Society of America? How long has it been available?

LIPSHITZ: A few months; it's dated 1988.

EDITOR: Let's talk about storage media- LP records, analog cassettes, CD's, digital audio tape, and so forth. What do you think the future holds? What do you think is going to happen to these various media? Are they all going to survive?

EARGLE: Can I tell you? The data as I see it tells me that the LP-outside of a few specialty houses like Reference Recordings and Dave Wilson and people like that-is really doomed. I think major companies will put it out for certain market segments, like jazz, or rhythm and blues, or country and western, where the population base hasn't had the infusion of CD players that the rest of the world has. But I think the classical business in the hands of the majors is going to go basically com pact disc and cassette. The Philips cassette remains the strongest medium that we have.

EDITOR: That's what you're being immortalized on right on.

EARGLE: Right. And I think that you're going to find more and more people duplicating cassettes from a digital source and hard memory, instead of having a running master tape, a duplicating master. I think that the vote is still out on the DAT as a possible newcomer, or certainly as a re placement for the Philips cassette. First off, the Philips cassette is much better than the LP was relative to what wants to replace it. In other words, the LP was re placed by something that was much better, much more easily manufactured, quieter, and so forth.

EDITOR: Many tweaks would disagree with you, but...

EARGLE: That's right, that's right. But by and large the Philips cassette, well duplicated, especially in Dolby C, really rivals, for quick A/B checking, a CD, except on certain kinds of music-you know, a piano recording, something like that. So the need for the DAT, as a carrier for prerecorded music, is pretty slim. As a device at our level, for recording, I think the DAT is really going to just take off; it already has taken off. I think it's going to replace a lot of machinery out in the field. But I don't really know that it's ever going to make it as a carrier of prerecorded program material.

McGRATH: It might be doomed to the same oblivion that the F1 was.

EARGLE: Right.

McGRATH: Who knows? I mean, it may fall into the hands of those of us who adore its functionality and its qualities, but it may...

EARGLE: Absolutely.

LIPSHITZ: Of course, the F1 was intended, I think, for the semiprofessional market. It was not intended as a consumer product.

McGRATH: Well, interestingly, I was a dealer for Sony, and that wasn't even marketed through their ES division, which was their high-end home division. It was strictly a consumer product.

LIPSHITZ: Well, in Canada now, for example, the PCM-601 is available from Sony, but only Sony professional. It's not avail able as a consumer product, as far as I know. But if you want prognostications, Peter, on these things, I think the LP is clearly going and will soon be gone.

EDITOR: Do any of you regret that?

CLARK: No.

EARGLE: I don't.

LIPSHITZ: Well, not greatly...

McGRATH: I do.

EARGLE: There's a little nostalgia, but...

LIPSHITZ: ...because I don't believe that the new media are inferior, you see. I sup pose you regret it in the way you regret something you're long familiar with...

EDITOR: I don't think that's what Peter means.

LIPSHITZ: I know.

McGRATH: I'm still one of those backward flat-earth or black-vinyl types.

EDITOR: You're a tree-worshiping analog druid.

McGRATH: That's it, exactly.

LIPSHITZ: I do disagree with John, though, about cassettes. But I guess I'm a bit of an unusual individual here-I do not own a cassette deck. I do not own a cassette deck because I've had a lot of experience with cassette decks. Our chamber music society every year, in order to make up for deficits, gets permission from some of the artists-the best concerts we've had during the year-to make copies on cassette and sell them to our members for donations.

EARGLE: In what way do you not agree with me? You don't think it's going to hang on as a strong medium?

McGRATH: They sound terrible-please! They're awful.

LIPSHITZ: They're not flat; they change from day to day; after a few months they start wow-and-fluttering...

EARGLE: Oh, of course.

LIPSHITZ: I mean, just a month ago, we had a violin-piano recital; the violinist lives in San Francisco, and he got his cassette copy-you know, we give a copy to each of these people free-and I get a call from him: it's a semitone sharp! A copy a semitone sharp that's duplicated directly from the digital master tape? On all our cassette decks we match carefully with a test tape, so that when we do our dubbing one is not going to run out ten seconds before the music ends for side one and that sort of thing. And I measured all our cassette decks, and none of them was more than 1% off in speed, which is 1/6 of a semitone. But this person has absolute pitch-maybe he didn't mean a semitone; maybe he meant a significant fraction of a semitone; I'm not sure. A 1% error for a cassette deck is par for the course. And after a while these things, wow and flutter, become intolerable. The president of our chamber music society has reached the stage where he can hear wow and flutter on cassette decks; it's an acquired ability.

Oboe is perhaps the best instrument to demonstrate that. He can even hear wow and flutter on live oboes. (Laughter.)

EARGLE: Aha. He's really good.

LIPSHITZ: It takes great skill. But for a perfectionist it's a flawed medium. For most people it is a very convenient, good quality medium.

EARGLE: Yes. True. I agree. Okay, let me say this about it, since I'm the one who implied that it had no problems. The thing is that there are more cassettes sold-more prerecorded cassettes sold-than any other single recorded carrier of sound. So that the consciousness level out there must be geared into what the cassette is offering.

The question we have to ask ourselves is, if there were dissatisfaction with the cassette as a medium out there, then it would bid fair to be replaced by DAT or some thing. I don't think that level of dissatisfaction exists with the prerecorded cassette because if it did, it wouldn't be selling the way it is. Now that may be a circular argument; [ don't know.]

LIPSHITZ: But I agree with John that the future of DAT for the consumer is highly questionable because I don't see its clear function. It is recordable, and that is continually made great weather of, yet the LP survived for 40 years and it was not recordable. I believe the cassette is where it is now because there was no other feasible medium for use in cars, and certainly for portable use when the Walkman was invented. The CD may never be as impervious to vibration and so on as a DAT or cassette can be; I don't know. But it seems

"They [cassettes] sound terrible-please! They're awful." "They're not flat;

they change from day to day; after a few months they start wow-and-fluttering..."

silly to have double inventory, one for your home and one for your car.

EARGLE: Record companies hate it. They hate the promotion of extra inventory.

LIPSHITZ: So, really, it makes much more sense to me to have a CD player or changer in the trunk of your car and use the same discs at home, and to have some kind of portable CD player, than to have two sets of software. And that would probably not mean buying two sets of software but buy ing the one and transcribing to the other.

Now is it worth really paying $20 or $15 for a two-hour blank DAT cassette to tran scribe a $15 CD? And that requires a

$2000 recorder as well, in order to enable you to do it, plus presumably a player, an other $1000 DAT player in your car, so you can play it. It seems to me it's a technology which the professionals and semi pros will be using. Many artists already, 1 understand, would request copies on DAT to take home to listen to, to approve or otherwise...

EARGLE: Most of the ones I know would take a cassette of that.

EDITOR: The hardware is still a gray market product. Nobody has really authoritative information on this-whether that can be expected to change or...

(Remember, this seminar took place before the great mid-1989 compromise in Athens, Greece, between the recording industry and the DAT hardware makers. Some of what follows here is therefore out-of-date, but the opinions expressed are still of in terest.-Fd.)

CLARK: I have some authoritative infor mation regarding professional units. There are Technics units on the shelf-they're by Panasonic-that have XLR connectors;

they are imported; they're for professional use. So it's here for the pro use, but any body can buy them.

LIPSHITZ: But you have to buy through a pro dealer, don't you?

McGRATH: No. There are, in fact, two or three gray-market companies now that are selling them to anybody over the counter.

Even home or pro units.

LIPSHITZ: In Canada they're available through Sony.

CARVER: The Carver dealers have pretty much killed it. They say no, it doesn't work right to the customer's interest. Of course they don't have them to sell.

LIPSHITZ: What doesn't work properly?

CARVER: DAT's.

LIPSHITZ: You have a DAT machine?

CARVER: Yes, I have a DAT machine.

LIPSHITZ: I didn't know you had a DAT machine.

CARVER: Yeah.

LIPSHITZ: Oh. Is it advertised?

CARVER: No. I don't make one. Carver Corporation doesn't. I have one.

LIPSHITZ: You want to be sued then by this conglomerate? Wait a minute, is he saying that he personally owns one?

EARGLE: Yes. He personally owns one machine.

CARVER: I personally only have one ma chine.

LIPSHITZ: Oh! I thought he meant Carver Corporation was offering one for sale.

CARVER: No, no, no.

LIPSHITZ: Ah! I thought, goodness, he's going to be sued. Then we'll see what hap pens.

McGRATH: But Nakamichi will have one out. They showed it at the show and they are intending to bring it in.

EDITOR: And they're willing to be sued.

McGRATH: We're going to take it on.

CARVER: The success of the DAT ma chine, at least now when it's expensive, depends a lot on how the dealers get be hind it.

EDITOR: Well, more power to them. I think somebody should take on the RIAA and beat them. And I think they will if they take them on.

LIPSHITZ: Well, Sony in Canada has been selling R-DAT's for a year and a half now-the consumer and the professional R-DAT's. But only through Sony Canada Professional, and that means you've got to buy it through a professional audio dealer.

CLARK: But that means you have to walk in off the street, plunk down your money, and buy it-the same as retail.

LIPSHITZ: Probably. I don't know. I haven't tried it.

CLARK: My office is in with a pro audio retailer, and anybody with, well, $2600,

$2700 can walk in and buy this profession al Technics DAT. Actually a couple of different models-and the Sony. There is no problem.

LIPSHITZ: The thing that riles me about this whole DAT controversy-and [ am sympathetic to the whole copyright ques tion, and there's no doubt that a lot of these machines will be used for copying copyrighted material, although the U.S.

Supreme Court has already ruled on the Betamax situation, and it's not clear that for personal use that is not acceptable-but what annoys me is the way this is treated as a new, unprecedented threat because it's a high-quality copy. They're implying that low-quality copies are all right.

EDITOR: Exactly. Not only that; they're implying that a very good copy is all right, but a very, very, very good copy-that's bad! That's the implication and that's nonsensical.

McGRATH: It's absurd.

LIPSHITZ: And on top of that, having the double lockout-that means the following two things: One, you can't do a digital-to digital copy because the consumer R-DAT machines will not digitally copy at 44.1 kHz sampling rate. That's locked out two ways. There's a copy-prohibit flag in the R-DAT format, which would automatically prevent copying at any sampling rate if that flag is set. Virtually every CD made has in its digital subcode a copy-prohibit flag, so you couldn't copy a CD onto your R-DAT even if your R-DAT were to allow digital dubbing at 44.1 because the CD digital output says "copy protected," and your R-DAT will say, "Sorry, I won't dub it for you." But on top of that, they have prevented recording at 44.1 and decided it's got to be 48. Now that means, we have-the chamber music society that I'm part of has-250 or 300 Beta-format master digital tapes. We could not transfer them to R-DAT format.

EARGLE: I'm sure you can find a way to do it.

LIPSHITZ: Oh, we could find a way to do it. But that's not my point. It's a double lockout; it is pointless. That lockout was already present in the digital interface for mat. Why put an additional one in to appease an industry that weren't appeased? They weren't appeased- -let's go over...

EDITOR: To the best of your knowledge, Stanley, how would the president of RIAA answer that one question?

LIPSHITZ: Oh, I've no idea. He probably couldn't really answer it without making some other sort of similar things...

McGRATH: Right there. (He shows his Sony professional DAT recorder deck.)

LIPSHITZ: But this is the professional...? This is yours?

McGRATH: Yes.

LIPSHITZ: We have in front of us the...

Well, you better say what...

CLARK: Exhibit C.

McGRATH: No, it's legal. It's legal. It's a pro unit.

LIPSHITZ: It's perfectly legal. It's the Sony portable pro R-DAT machine, which will do these things.

CARVER: They're all legal.

LIPSHITZ: But, for example, not all the pro units will do what you think they do.

Professional people have a need to transcribe material for legitimate purposes.

McGRATH: You are right.

LIPSHITZ: I mean some recording studios are being sent CD's by the copyright hold ers and told, "Here's the digital master. I'd like you please to rearrange the titles, the songs, in the following order." If they have a CD player with a digital output, they can't dump this down onto digital tape and rearrange the order because that CD has got a copy-protect flag on it, and they can't dub without doing various things...

McGRATH: By the way, this will not accept a 44.1 signal in the digital domain.

It will record in 44.1; it has a digital input;

it will take 48 and 44.056, but it will not accept 44.1.

LIPSHITZ: If it's like the Sony R-DAT, the 2500 I had experience with, it will take 44.1 if it does not have a copy-protect flag.

McGRATH: But I tried it from another R-DAT to this, where I've never recorded the copycode.

LIPSHITZ: Oh. And?

McGRATH: In fact, what I did was, I recorded a master on this, from my analog, and then before I mailed the master off 1 wanted to make a 44.1...

LIPSHITZ: Which digital input format are you using? Which interface format?

McGRATH: I think it's...

"...it isn't likely that a company is going to keep a 48 kHz master for making R-DAT's, and then transcode that down to 44.1 and edit it, and then make CD's."

LIPSHITZ: Sony-Philips? Mc GRATH: ...the EBU.

LIPSHITZ: The EBU? The XLR thing?

McGRATH: Yeah. Exactly.

LIPSHITZ: Ah. Well, that's not the case with the other Sony professional R-DAT.

EDITOR: I've heard through various sources that there are some consumer-type DAT machines that will record at 44.1 with a very minor unauthorized modification.

McGRATH: There are dealers, in fact, that are advertising that service.

LIPSHITZ: I think you must be quite careful in what you decide to publish in your magazine about this whole issue. I think we wouldn't, perhaps you wouldn't, want to be seen to be advocating such a modification of equipment.

EDITOR: I'm not advocating; I'm just reporting what I've heard.

LIPSHITZ: That is true. One can indeed make modifications, depending on the internal architecture of various equipment, to get around some of these things.

EDITOR: That still doesn't get rid of the copy-protect flag on the CD itself.

LIPSHITZ: No, but in principle somebody could make a black box that takes the digital feed in, finds the copy-protect flag but modifies that bit, and puts the same digital feed out. I mean, one can make those things; it's not something that the average person can easily do. There are ways of defeating almost anything you want to defeat. My point is, the people for whom I see a clear use for the R-DAT are the pros and the semi-pros who do their own record ing. If you're doing your own recording, it's a magnificent recording service. It's more convenient to carry around than a video recorder and a digital adapter, or an open-reel format digital recorder.

EDITOR: What about editing?

LIPSHITZ: It doesn't have editing capability at the moment. But for our use, the chamber music society's use, these are a lot of functions we don't need. We don't have multiple takes; we can't edit.

McGRATH: It's the same for all the broadcast tapes I do. It's not a require ment.

EDITOR: You have editing facilities on that machine?

McGRATH: Not at all.

LIPSHITZ: You can edit between tracks by just, you know, record pause, just over writing. It's not glitch-free, but if it's silent when you do that, that's fine.

EDITOR: Or you could copy onto another machine, digitally.

LIPSHITZ: For the average consumer, I don't see the need to do the recording. You see, my point was, the cassette was successful because they needed a compact for mat for car or portable use. It needed to be recordable because there weren't prerecorded cassettes available initially. So peo ple become used to a recordable format. If you give them a recording-type R-DAT, I can see it could conceivably be used for dubbing material, possibly for use in car or portable environments where people don't want to use CD's. If it did not record and was a playback-only medium, I really don't see it. It doesn't really have advantages over the CD; it has disadvantages in most respects with respect to the CD.

McGRATH: Really.

CLARK: It's better for the car.

LIPSHITZ: It might be cheaper to duplicate in the long run than the CD is, but I doubt it. The CD has had quite a few years lead time to get its process down.

EARGLE: Yes, the tape cost alone is not likely to go down very, very much...

EDITOR: There have been suggestions to the effect that the higher sampling rate produces a minor improvement.

LIPSHITZ: But you won't get the higher sampling rate because the material that would be released on prerecorded CD's will have been...

McGRATH: It's all 44.1.

EDITOR: But the prerecorded DAT, conceivably...

LIPSHITZ: But the 44.1 dubs that people get will come from the CD masters.

EDITOR: ...where, say, a Mitsubishi master is made at the 48 kHz sampling rate and then duplicated digitally on a DAT.

EARGLE: Peter, it isn't likely that a company is going to keep a 48 kHz master for making R-DAT's, and then transcode that down to 44.1 and edit it, and then make CD's.

EDITOR: dmp does that.

EARGLE: Well, dmp is a very special little company, and one man runs it; he can do anything he wants to do. But this will not become an industry standard because it would require two different edits.

LIPSHITZ: And not only that. You've got to watch it, though, because there are many professional digital recorders that are switchable 48/44.1. But it does not follow that if you switch it to record at 48, the analog filters are switched, so that your bandwidth increases. It may still use the 20 kHz filter it uses at 44.1.

EARGLE: In which case you gain nothing by doing it at 48.

EDITOR: All you're saying is that it's not designed as a total two-format machine.

LIPSHITZ: No-and quite honestly, if I'm given the following choice, even assuming the analog filters are switched, to record at 44.1 or record at 48, for CD release I would record at 44.1. There's no point in doing digital arithmetic unnecessarily when you're digitally filtered to get it down to the lower bandwidth. It's point less.

EDITOR: That makes a lot of sense.

McGRATH: Back to dmp, though, what he does-he edits in the 48 mode and then number-crunches to the 44.1, so he doesn't have to edit in both domains.

EARGLE: Is he doing razor-blade editing?

MCcGRATH: Yeah, on his Mitsubishi. Yes.

LIPSHITZ: But there are the Mitsubishi 96-kHz sampling machines that are now available. What on earth are people going to do with those?

EDITOR: I think that happens to be a very good product, if you like the music-dmp I'mean.

McGRATH: I think it's wonderful. I think he makes the best-sounding product I know of.

EDITOR: I suggested to Tom Jung that he dig up some of these rapidly fading jazz greats that hang out in various places in New York, half forgotten. They're not be ing replaced; there is no new generation of jazzmen who play like that, and they would deserve his kind of recording, but so far he is only doing this New Age stuff- or whatever they call it.

LIPSHITZ: Well, quite honestly, you know, with a portable R-DAT machine and a good mike/preamp thing-you could al ways carry it all in under your arm-you could record in clubs and produce pretty high-quality masters, I think, if you can get permission to do that.

EDITOR: Sure.

CLARK: Can I ask a question? Why do we even consider a flawed medium like the cassette-amongst ourselves, why do you

-or LP for that matter, that has demonstrable flaws in it, when at the same time you're really questioning a console, with these series of amplifiers, that's demonstrably okay?

McGRATH: I'm not sure I follow.

EARGLE: Would you ask that again? I missed the line of the question.

CLARK: Well, we have something that specification-wise--and to me sound-wise--is great, like digital audio tape...

EARGLE: Okay.

CLARK: ...and you were saying that the cassette is sort of okay, but this is...

EARGLE: Well, I would say, look-I ought to be ashamed, but I have never bought a cassette tape. I use it for making copies for those people who need them.

Everybody that I know who wants a copy of something wants a goddam cassette.

McGRATH: That's right. And I don't own one either, and I live with them for exactly the same evil purpose.

EARGLE: Okay. And the only reason I put forth the cassette as a viable medium-not likely to be replaced by the expensive R DAT, which is the only thing that would replace it in kind-is that there are more prerecorded cassettes sold than any other thing in this country. They must be doing something right-either the expectation of the masses out there is so low that they will accept a cassette with all of its gar bage, or we're being a little bit too critical of it.

LIPSHITZ: Isn't the answer, John-are you not actually implying-that if these people go off their cassettes at some point, they'll probably go on to CD, not to R-DAT?

EARGLE: Well, I know they're not going to go to R-DAT because of the expense, the high cost of buying prerecorded materi al. You can pick up a prerecorded cassette

"...we have an unfortunate situation... The R-DAT was never intended as a professional digital format and may become a professional, not a consumer, format." for three or four dollars.

CLARK: Yes, but that's like saying com puters are expensive.

LIPSHITZ: But the R-DAT material would come down if it were mass-produced.

EARGLE: It's not going to come down.

The tape cost will not come down. The duplicating cost will.

LIPSHITZ: But it's not clear to me why you couldn't produce a 75-minute R-DAT and sell it for the same or less than the price they sell CD's for. I'm sure you could.

CLARK: I didn't get that point I was trying to raise satisfied.

EARGLE: Dave, what is your question again?

CLARK: The question, which turns out not to be really much of a question after it's been brought up, is why do we accept the cassette? Because, well, we really don't accept the cassette. But I'd like to make a point because we're talking high end, state of the art in audio here, right? I would just like to make the point that the cassette does not represent the state of the art in audio, and it never did.

EARGLE: I think we all agree to that. And I think we all say that it's an inferior re corded medium that just so happens to have elevated itself to the highest figures in the country.

McGRATH: But the other half of your question was, how could someone criticize consoles...

EARGLE: ...rail against the console and accept cassettes.

McGRATH: And LP's are so bad, too-I mean, was that the way it went?

CLARK: Yeah, I was throwing that in. But can we just say, when cassettes are so bad?

EARGLE: Well, it's a matter of accepting.

The thing is that the public doesn't accept consoles; only professionals accept them.

The cassette exists because of the vast public acceptance, and we just have to ac knowledge that. The question that original ly got this going was Peter's question about mediums, about recorded media.

And any question about recorded media must indicate the largest-selling one, even though it's terrible.

LIPSHITZ: I think we have an unfortunate situation with the DAT, in the same way that the cassette was unfortunate. It was never intended for music and became the largest music character that we know, in terms of numbers-of sales. The R-DAT was never intended as a professional digital format and may become a professional, not a consumer, format.

EARGLE: Well, that's the only area where it's been allowed to grow. It does it so well!

LIPSHITZ: It does not have the time code capability that one would want; at least it's not formalized in some standardized way et...

EARGLE: You don't need it at that stage;

you dump it to an editor.

LIPSHITZ: One could. But if you were designing a professional R-DAT, you probably wouldn't design it the way the R-DAT format is...

EARGLE: Well, it's an another funny screw-up. I mean it's brilliant engineering, improperly brought to the market.

LIPSHITZ: ...so it's another case of this situation. It's a beautiful thing-I mean, I initially was very skeptical about its robust ness, but apparently it's surprisingly robust.

It's very nice. And it probably is going to be around for some time. I don't know whether it will become the portable professional format; I suspect it will because so much money has gone into the development of the technology, the standardizing of it and so on. I think the Japanese will tend to stick to it.

McGRATH: How old are your earliest F1 masters? When did you start using it?

LIPSHITZ: When did it come out? Was it

'81?

McGRATH: Yes. I think I got one of the first in the United States...

LIPSHITZ: Probably '82.

EARGLE: "82 was when I got mine.

LIPSHITZ: I think '82.

EARGLE: Seven-year period.

McGRATH: Mine are imported from Japan. And the masters I made from them...

LIPSHITZ: No problem.

McGRATH: No problem. They still play.

Absolutely no problem.

LIPSHITZ: I always used high-grade types.

I had bad experiences with other types because of head clogging.

McGRATH: But everyone said that they're wonderful-but... And wait a few years or wait a couple of hundred plays... I made maybe a thousand masters.

LIPSHITZ: Oh yes, but Doug Sax also said that you couldn't make a single digital-to digital copy without errors. I mean, people have said all sorts of things. I've reached the stage now where I don't care what people say-I care what people can prove.

EDITOR: Peter, I would like to ask you, before we quit this subject, to state the "tree worshiping analog druid" party line.

EARGLE: The what, now...?

McGRATH: The tree-worshiping analog druid party line. Whatever that is.

LIPSHITZ: The initials spell TWAD.

McGRATH: On what issues?

EDITOR: You still consider the very best analog achievable to be better than the very best digital recording achievable? I believe that's it.

McGRATH: That's essentially my position.

EARGLE: Are you talking about finished product for the consumer or a master tape which you made?

McGRATH: I'm talking about having made recently, using the analog master tape as a starting point-which may or may not be a valid premise for some people- but taking that tape and taking the original machine that recorded it, using "brown rice" type wires, and dumping it straight into a Sony 1630 with no intervention, just finding where the peak level is and then making a transfer...

LIPSHITZ: Nothing wrong with that.

McGRATH: That's exactly what I've done.

LIPSHITZ: 1 mean, I've been buying a whole lot of these old historic releases...

McGRATH: No, no. Let me finish the whole thing.

EARGLE: He's not through yet.

McGRATH: I'm building a case. And then we've taken that tape and dumped it into a 1630. The 1630, then-all that's done to it at that point is just time code is put on it, and the CD...

EARGLE: It's making it from analog to digital at the same time. It's a very important step.

LIPSHITZ: A-to-D conversion, certainly.

McGRATH: No, no. What I'm saying is, after that transfer we just put the time code on it. Then CD's are made. Then that same original master is taken to Doug Sax and it is cut. And then, comparing...

CARVER: Cut? An LP?

LIPSHITZ: From the CD? From the digital master?

McGRATH: No, no. Cutting from the original analog master-we are then making an LP. And then comparing the finished LP product to the CD product, back to the original analog master tape.

LIPSHITZ: Played on the same machine every time?

McGRATH: Played on the same machine throughout.

LIPSHITZ: Even at Doug Sax's?

McGRATH: No, it was cut on a different tape recorder there. I sent something like 12 different reference tones and so forth, so I'm pretty sure that the EQ was right- and I'm not exactly sure what he does there.

LIPSHITZ: Was he able, by the way, to cut the disc with a nonpolarity?

CARVER: I have a lot to say about this stuff. (Muttered under his breath.)

McGRATH: We tried both, in terms of a comparison...

EARGLE: Well, what are your objections to this? (Addressed possibly to Lipshitz or Carver, not clear.-Ed.)

McGRATH: ...and he inverts.

LIPSHITZ: He inverts?

McGRATH: The bottom line is, when we played the LP and compared it to the com pact disc using a--you name it-an Accu phase with a Wadia D/A converter, or an Accuphase straight, or any one of a variety of different CD players, and using the LP on one of my better turntables, such as a Goldmund, the LP more closely mirrors what's coming off the analog tape than does the CD. And based on that, I would say I still would have to consider myself an analog druid tree worshiper or whatever.

LIPSHITZ: What do you conclude from that experiment, then?

McGRATH: Well, what I conclude is that if the digital technology cannot indisputably mirror even the deficit-laden analog original, then far be it from doing so with

"What I have done is that I've compared this [professional DAT] making duplicates of analog masters to the original analog master, and you can still hear a difference." live original music.

EARGLE: Okay, can I ask you another question? Have you taken a 1630 and per formed the following experiment? Because the 1630 is the only element in the chain that could really be in question-or it's certainly one element that could be in question-the other element would be the CD player, the Wadia and the Accuphase.

Have you ever jumpered from the composite digital in and out on the back of the...

LIPSHITZ: Is the Wadia a player?

McGARTH: I'm sorry. This is a Wadia D/A converter.

LIPSHITZ: Yes, I know of it.

EARGLE: Okay. But the point is that if you take the digital encoding part of this whole thing and run its output into its input and use it as a link in a comparator and A/B through it or around it, you think you're going to hear much of a difference?

McGRATH: I don't know.

EARGLE: Okay. Have you ever taken your F1 unit and A/B-ed it like that?

LIPSHITZ: Yes. And I've challenged Ivor Tiefenbrun to that experiment, and he could not tell the difference between A and B.

EDITOR: Yes. That was written up in the Boston Audio Society newsletter.

EARGLE: I can't either. And I'm sure you couldn't tell the difference there.

CARVER: I sympathize, Peter, I sympathize a lot with your experience...

LIPSHITZ: I have an explanation for what you're saying, a potential explanation for your conundrum, and that is-as John has mentioned, there are two possible places where the change could have occurred.

Because when you transcribed your analog tape to digital, you used your analog tape recorder...

McGRATH: Sure, there is a possibility...

LIPSHITZ: ...when you listened to it for the comparison subsequently, you used your analog tape recorder.

McGRATH: Correct.

LIPSHITZ: Forget the LP; we'll try to compare the digital copy with the analog original-and you're saying they sound different.

McGRATH: They do.

LIPSHITZ: So the problem is either the 1630 or the CD player, as far as I'm concerned. Now I can say this about the Wadia CD player [he means DJA processor- Ed. ]-it is not flat in frequency response.

McGRATH: Well, the differences are also audible when just coming straight out of the DP-70, which is presumably a pretty good CD player. That's the Accuphase.

And, if anything, the sound gets closer to the subjective impression of what's on the analog tape when going through the Wadia.

LIPSHITZ: But forget that. Why not play it on your 1630 and see if it sounds like the analog tape? Because that eliminates the one extra step.

McGRATH: What I have done-for the sake of argument, what I have done is that I've compared this (pointing to his Sony professional DAT deck) making duplicates of analog masters to the original analog master, and you can still hear a difference.

And this at either 44.1...

LIPSHITZ: And how did you set your levels and so on, Peter?

McGRATH: What you do is, you simply take a 1 kHz tone, which I have on the analog tape, and run it into here; then, when running this back, we match the two levels precisely-because I have a trim on the output of the Stellavox...

LIPSHITZ: That's measured at the loud speaker terminals?

McGRATH: Yes, exactly. And you hear a difference. Now, I'm not sure if this phase inverts-there's no documentation...

LIPSHITZ: I'm sure it doesn't.

McGRATH: I don't know that it does or doesn't.

LIPSHITZ: I doubt it.

McGRATH: I hear a difference. I mean, it's there. I mean, in terms of what I've...

Maybe, again, I've not done an ABX...

EARGLE: Okay, if you walked into the room, gone out of the room, had somebody toss a coin, and...

McGRATH: Yes, yes. Because what hap pens, I come in and I sit down and I say,

"Scramble!" 1 don't want to know what I'm listening to. And I can pick it every time. It's very obvious to me.

CARVER: I sympathize with that. I've had similar experiences-but from the up side of the coin. My experience has been like yours when I listen to the LP and listen to the CD-the LP is much nicer to listen to. But I've not been able to have your experience, in which the analog copy compared to the digital copy is superior. I've always found that they're indistinguishable, when ever I've done experiments such as you've described. So my results are different than yours, and I can't explain them. All I know is-both experiences are anecdotal. But what I did do is a series of converging experiments, and it actually was written up.

When my CD didn't sound as good as my LP, I set about to make them sound the same. And it wasn't very difficult; it wasn't very hard at all. I took a very pedestrian approach. I didn't have the kind of equipment that you had; I just went out and bought a bunch records-LP's and a bunch of CD's-and tried to make them sound the same. And I found that I could, if I simply modeled into my signal chain the specific frequency response of my cartridge/LP interaction-and also the fact that the vertical information coming off the LP was significantly larger than the equivalent vertical information coming off the CD. And so I had to make a little circuit that...

McGRATH: Introduced some phase shift?

CARVER: No. It didn't introduce phase shift.

LIPSHITZ: Increased the L - R.

CARVER: Increased the L - R. And also, I had to do some other funny things. I had to equalize the L + R channel a little differently than the L - R channel. That introduced a little time delay. Then I had to make a time-delay compensator-sort of the same thing we talked about earlier- before it went back into the matrix. By the time I was finished, I had a circuit board that had a lot of stuff on it. I was looking at it-shit, it takes all that...

LIPSHITZ: To degrade the CD player to what your cartridges do when you play an LP-to what cartridges and cutters do to the original signal.

CARVER: Yeah. The thing of it is, I genuinely enjoyed the LP better. The things that I like about my hi-fi set-which is sort of a warm bass, and a nice curtain of sound, lots of bloom-my cartridge did all of that.

And it didn't take long to figure out that it was my cartridge/LP interaction that pro duced the soundstage that I liked.

EARGLE: I hope you're not putting forth the idea that that's more accurate.

CARVER: Heavens, no!

EARGLE: Okay.

EDITOR: He's saying he likes inaccurate sound.

CARVER: I'm saying that I liked it better, but it's not difficult. I mean, in the next breath-you know, unless I sit on my hands and bite my tongue-it would be very easy to say, you know, it must be more accurate, it sounds more lifelike. Be cause it does sound more lifelike. I mean it has depth-and more depth. But it can't be more accurate; obviously it isn't.

McGRATH: Well, generally, when I get that kind of customer coming into my store-saying that CD's sound like gar bage, and this and that and the other-I generally take a very hard-line posture and just simply ask, on what basis are you say ing that? Here I'm just simply offering what I have experienced. That's why I said, about making these comparisons...

CARVER: The point of what I was saying is that the LP is not superior to the CD. I can make the CD sound like the LP; I could never make the LP sound like the CD. So, by definition, that means that as a recording medium...

McGRATH: Well, I'm still not convinced that that isn't true [viz that the LP is su perior to the CD-Ed.]. 1 still feel that an LP, meticulously manufactured, using what hopefully is a state-of-the-art lacquer, state of-the-art cutting techniques, and then played back on a state-of-the-art turntable...

CARVER: Oh, yeah...

(Everybody talking at the same time.)

EARGLE: It's hard to get all of the state-of the-arts to happen at the same time.

LIPSHITZ: You see, it's rather like Mitch Cotter telling me...

CARVER: It is possible. I agree; it's possible. It's amazing how good an LP can be. I mean, it can be virtually indistinguishable from a CD.

LIPSHITZ: So can an analog tape.

CARVER: Sure.

McGRATH: Well, again, all I'm saying is that the LP produced from that resulted in sound more closely mirroring what was on the analog tape. That to me makes it better.

"The point of what I was saying is that the LP is not superior to the CD. I can make the CD sound like the LP; I could never make the LP sound like the CD."

CLARK: I agree than an LP is better than a cassette.

LIPSHITZ: Well, I don't know the answer to the difference you heard. We're second guessing; we can't answer that question.

Sounds like a perfectly reasonable thing you did, and if you indeed heard the difference, I can't explain it now. However, I would...

EDITOR: There is a possible explanation.

LIPSHITZ: ...say-my belief-there is an explanation; I don't believe it's esoteric, and I'm sure that if...

CARVER: If we were there, we would sort it out.

LIPSHITZ: ...we could repeat that experiment, we'd be able to find the reason. And whether it relates to frequency response or a polarity thing, a distortion thing, an over load, I don't know what it conceivably could be, but it's one of them.

CARVER: But we would find it, and it wouldn't be mysterious.

McGRATH: Well, I'm not suggesting that it is.

CARVER: And it wouldn't mean that the LP storage medium is somehow more magnificent or superior. That's what it is...

EDITOR: Here's a possible explanation.

Your digital recorder is more bandwidth limited than your pure analog system.

CLARK: That's true. Maybe it's...

LIPSHITZ: But so are his ears, I think.

How high do you hear?

McGRATH: What?

LIPSHITZ: How high do you hear?

McGRATH: What? (Everybody laughs at the joke except Lipshitz.)

CLARK: Almost up to voice range.

EARGLE: How old are you?

LIPSHITZ: Very good. I didn't guess that you were being silly.

McGRATH: I have no idea. I would guess 10 kHz, 14 kHz.

LIPSHITZ: Yeah, but I mean you're not claiming more than 20 kHz, like Laurie Fincham-21 kHz he used to claim-I don't know if he still does.

McGRATH: I'm not claiming anything at all. No, no. I've no idea what my ears do.

I'm 40 years of age, and I suspect that it's normal...

LIPSHITZ: I'm below 15,000. I wouldn't be able to listen to the TV whistles...

EDITOR: Okay. One could still suggest that there's a lot of out-of-band energy in the analog program material, which, when properly reproduced, may not be directly audible but may somehow interact with the audible spectrum.

EARGLE: No! You'll have to explain that one. That's a long article.

McGRATH: Come on, Mitchell!

LIPSHITZ: A nonlinearity. Okay, I will give you an example that happened three, four months ago of exactly that. A friend of mine at the university comes to me, and he says, "I borrowed a CD from a friend. I tried to transcribe it onto cassette tape- because I liked it. Can't do it," he says.

The CD is perfectly fine, he says. The cassette tape has got this horrible distortion on it. So I started asking various questions.

Yes, even recording at lower level, it's still like that. Problem. So I asked him to bring me the CD and the cassette copy. I listen to the CD; it's a pop recording with a female vocalist, and so on...

McGRATH: There's a 12 Hz signal in there or something...

LIPSHITZ: No. No. Perfectly fine-sound ing but clearly gimmicked recording-the cassette tape has got this ghastly, gritty, horrible, twittery sound on the voice.

Things going on. Ha! Sounds like an inter modulation thing-something beating with the bias? Because analog recording is not very linear, certainly not with cassettes. So we do a spectrum analysis of the CD-oh, goodness gracious, look what's happening above 5, 10 kHz-I forget exactly where the comer was going up. So a student of mine says, "Oh," he says, "looks like they might have used an Aphex Aural Exciter on that." Now you know. You listen to the voice-it's got this excessive breathiness to it. I guess this is Aural Excitation. There is a huge high-frequency thing there; cassette decks cannot record that without overload and beating.

EARGLE: Well, if you bring the level down, it ought to disappear.

LIPSHITZ: It did not.

CARVER: By 20 dB. You have to bring it down 20 dB.

LIPSHITZ: 1 think his cassette deck was bad. I took it home. I've got my modified Revox B77. I've been doing digital recordings for the last seven years or whatever, but before that, you know, I've got dozens of master tapes, and I would challenge many people to tell the difference between the analog and the digital, except on the very loud passages. When you set the volumes within tenths of a dB-and you have to be very accurate on that-they can be very close. But no problem recording it on my machine, so I think it's just his cassette deck, and/or the cassette medium couldn't handle that. But there's an example where there is no obvious explanation for the horrible distortion. And it wasn't out-of-band stuff because the CD produces no out-of band signals, nothing above 20-odd kHz.

EDITOR: You said something, Stanley, that I would like you to explain. It doesn't really belong in the discussion of storage media, but it came up and we might as well handle it. You said that our ears are even more bandwidth-limited than the dig ital system and therefore the bandwidth of the digital system is perfectly adequate. I have heard suggestions to the effect that the ear can detect rise times corresponding to higher frequencies than we're able to hear as pure tones. Do you...?

LIPSHITZ: I know many hypotheses. I know many people who say, may be able, or could. These are all possible. I just hap pen to believe they're not true. [ mean...

EARGLE: Well, nobody has ever proven them at least.

CLARK: Some of those experiments- I've heard some of that stuff. What it turned out to be when it got into the fine print-somebody says this in an offhand manner, and I followed it up-they're talk ing about presenting two things almost simultaneously, and when could you detect a shift, and the amount of change required in the two presentations.

LIPSHITZ: At your two ears, you say?

CLARK: And the time, when translated to a frequency, was above the audio band.

EARGLE: Oh, yes. Well, the ear, you know-in terms of microsecond-resolving of a very sloppy onset of the signal-you know, it will shift...

LIPSHITZ: Interaural is one thing. But monaural or intra-aural...

EARGLE: If you do what you're saying and translate that into an equivalent band width...

CLARK: It gets very high.

EARGLE: ...it gets very, very high.

CLARK: Like 40-50 kHz or something.

EARGLE: It's a remarkable property of the ears acting in concert with each other and improving on the bandwidth.

McGRATH: Couldn't there be some relation to that?

EDITOR: Couldn't there be that type of energy in actual program material?

CLARK: No, we're not talking energy;

we're talking about a difference with which low-frequency energy is presented to the ears, and then the very mathematical, artificial thing of saying, "What's the time difference?"-and then looking at that time difference, the reciprocal of it, as a frequency. That's the only explanation; the one thing that came up that I understood with these things. That's the best I could do as a translation.

CARVER: You know-and 50 microseconds is a long interaural time space...

LIPSHITZ: Oh, no, no, no. You can hear a couple of microseconds. What you can hear is down in the less than 10 microsecond range.

CARVER: That's what I said-50 microseconds is a long time. It would be very easy...

"If you present a signal to both ears with a small time offset between the two, that time...

can be less than the rise time of the ear, and you can still detect the timing difference."

LIPSHITZ: 50 is very big. But now 20 kHz bandwidth would correspond to- what's your .35 over... You know, what is the standard first-order rise time of a 20 kHz bandwidth signal? I forget the number offhand, but it's...it's...

(A somewhat tentative calculation discussion ensues between Clark, Lipshitz, and Carver. They end up getting it close, but no cigar; the correct answer is 18 us.)

CARVER: ...It's 20 microseconds, and 20 microseconds is a big time for your brain to...

EARGLE: A long time, yes.

CARVER: I mean, your brain will be able to pick up a small fraction of 20 microseconds, so...

LIPSHITZ: So the point, I think, that's being made is that if you present a signal at one of our ears, the mechanical system that lets it into the inner ear has a frequency response cutoff-an incredible plum meting low-pass filter that chops off, if your hearing is still good, at maybe 20 kHz-and the inner-ear transduction mechanism converts all this into neuro-signals that go off to the brain, and the rise time of the system based on that frequency response cutoff is not very fast. However, if you present the signal to both ears with a small time offset between the two, that time off set can be less than the rise time of the ear, and you can still detect the timing difference.

EARGLE: The firing takes place at different times.

LIPSHITZ: Because the signals are cut or delayed going into both ears the same amount, but the one was additionally offset by a small amount towards the other. No contradiction there.

EDITOR: But are you suggesting that this kind of phenomenon cannot be reflected in actual program material?

CARVER: No, no, no. It can.

CLARK: It could be.

CARVER: It can, easily.

EARGLE: If you could derive a binaural experiment to do this...

CARVER: It's a binaural... It's not a test of how we hear rise times or high frequencies; it's a test of how we hear binaurally.

That's what it is.

CLARK: And you could do it with a 20 kHz band-limited system. You could run this experiment.

LIPSHITZ: Yes, exactly. That's a good int.

EDITOR: I see. Okay.

(Unfortunately we must stop in medias res, having run out of space. To be continued, and concluded, in Issue No. 15.)

-------

[adapted from TAC 14]

---------

Also see:

Seminar 1989: Exploring the Current Best Thinking on Audio, Part III of the Continuing Transcript

Correct Speaker Placement for Optimum Bass Response: A Simple Mathematical Method

Various audio and high-fidelity magazines