by John Culshaw
LONDON--It is a pleasant experience to find one's own thoughts emanating from another mind, and a double bounty when the mind is that of an old adversary. In other words, welcome home, Conrad L. Osborne, for all is forgiven. Or, rather, almost all, for I have not yet closed the door be hind him and may well find reason in the future to boot him out again.
Of course it is open to readers to conclude that Conrad and I have grown long of tooth and hard of hearing, for it is many years (it was 1968) since we joined in that glorious battle over my production techniques in London's recording of Richard Strauss's Elektra-a battle from which I emerged claiming total victory. (So, incidentally, and with typical impertinence, did he.) But some of his words concerning the CBS Louise (February 1977) force me to withdraw my fangs, bury my hatchet, and what ever else one does in the face of plain common sense. I quote from his review:
The recording ... is only fair by current standards, missing any real breadth or weight of sound. I do not at all care for the runny textures and strange balances of the street scene-the vendors' voices do not sound off, but only separate (the same is true of the supposedly distant chorus in Acts III and IV), the worst instance being that of the Carrot Vendor: The whole point is that he is crying out at the top of his lungs, but at a considerable distance, whereas he is clearly not more than two feet away, crooning. All the staging here sounds like mixing, and it's pretty unconvincing [my italics].
There are only two things slightly amiss with that statement. The first is that Louise is, on the whole, representative of current standards in operatic recording, and the second is that when Osborne writes about mixing he does not mean what is done during recording, but rather reduction from multitrack to two-track afterward which process is rapidly turning from the helpful adjunct it was meant to be into the blight of modern recordings, operatic or not. It makes for flat records. Not flat in pitch (which is merci fully rare) nor physically flat in relation to the turntable (which with advantage could have a higher incidence), but flat--utterly, boringly flat--in perspective. Louise is not the only victim. Both the recent versions of Meistersinger are flat in my sense of the word, and CBS and London have each achieved the impossible by flattening Daphnis almost out of recognition, although London confuses the is sue by having shortly before produced a recording of Prokofiev's Romeo and Juliet with the same conductor and orchestra (Maazel/Cleveland) that is a model of what orchestral perspectives should be. Most RCA operas over the past few years have been as flat as the proverbial pancake, and the reason for all this is that producers are not using space any more.
They record in one perspective and then attempt to "reduce" their multi track to create a spatial effect; and the truth is, brethren, that it doesn't work.
I can think of a parallel from some twenty years ago when what is known as "panpotting" was introduced. No longer, I was told, need Tosca move in order to exchange intimacies with Cavaradossi; on the contrary, she could stay right where she was by her own microphone and be "panned" over to him. It worked, in the sense that her voice certainly traversed the space between them. The only trouble was that, instead of sounding like a woman heading toward her lover, it sounded like a man turning an electronic control; in other words, her disembodied voice floated uncannily across the sonic spectrum. I likened it to underwater opera, if you can imagine such a thing, with Tristan paddling down a forest stream to greet Isolde, or Amonasro backstroking up the Nile to intercept Aida and Ra dames. It was an easy but self-defeating way of avoiding a problem, for in stereo there is no point in moving characters at all unless a dramatic purpose is served.
What Osborne noticed in Louise is the same canker in a much more advanced state. You need more from the trombones? Then don't be old-fashioned by doing anything so primitive as asking them to play a little louder, or slightly increasing the level on an appropriate microphone (which in terms of good sound may not be the one nearest to them), or placing them in a more advantageous position. No; all you do is to "pull them out" on track six (or whichever it is) at the reduction stage and then pop them back when they've finished. The fact that the result sounds dry, artificial, and flat seems to strike few of today's producers. The trombones are there, aren't they? You can hear them, can't you? Okay, then what's wrong? What's wrong is the texture, and the place within that texture of all the strands it comprises. I have said it be fore, and now Osborne has said it again (reference the Carrot Vendor), but it is still worth saying for a third time: There is all the difference in the world between a soft sound recorded close and a loud sound recorded from a distance. It is so utterly obvious that it hardly seems worth writing down, and indeed it would not be worth writing down were not so many producers ignoring it, month by month.
The word most frequently used by reviewers to describe this and associated phenomena is "congested," and a congested sound is exactly what in sensitive reduction produces. It creates an apparent rather than a real space between orchestral sections, and choruses tend not to integrate with the rest of the sound. The fact is that nobody has yet invented a satisfactory way of conveying spaciousness by artificial means. A new version of the War Requiem, be it ever so multitracked and echo-plated, would not, I wager, compare with the Kingsway Hall two-track original.
But please do not misunderstand me or Osborne, for although we are united on this issue we are not reactionaries; we are not against multi track per se any more than we are against microphones. We are simply against the misuse of multitrack and all the laziness and self-indulgence that such misuse implies.
(High Fidelity, Oct. 1977)
Also see: