MCS Review On-Line Reprints

More About Recording Perspective

by William Sommerwerck

Reprinted from MCS Review, Vol.6, No.3, Winter 1985.

Paul Alter's letter in the Autumn '84 issue defending the point of view that a recording should reflect the composer's intentions about the music, raises some points that deserve longer discussion.  As a music lover and amateur recordist, I'd 1ike to throw in my own two cents' worth.

The ultimate esthetic question of music recording (particularly for classical and jazz, which are concerned with external 'reality', rather than an arbitrary studio situation) is one of how the music is presented to the listener.  That is, what is the point of view (both esthetic and acoustic) of the record producer, and how will the listener perceive it?

Paradoxically, the question directly concerns engineering, as well as esthetic matters.  The sound of a performance cannot be magically translated to the listening room; an instrumentality is needed, and the nature of the mechanisn has a profound effect on the esthetics of recording.  Here are two examples, both obvious but nonetheless valid for it.

Until 1925 all recording was acoustical, with extremely restricted frequency and dynamic range.  Acoustical recording could not resolve fine sonic detail, so reverberation only tended to muddy instead of enliven the sound; hence, orchestral recordings were made in acoustically dead studios.

The integrated circuit has made practical recording consoles with 48 (or more) inputs, as well as sophisticated (and very natural-sounding) digital reverb.  This technology has encouraged many producers to jam a mike 'down the throat' of each orchestral section (or even each instrument), place each on a separate track, then position the sounds with pan pots (rather than by their locations in front of a pair of mikes), and cover the results with a glaze of synthetic ambience.  The results, of course, are not exactly what one would obtain with two mikes behind the conductor.

In both cases, the music (or more precisely, the presentation of the music) becomes subservient to the technology.  In the former case, it was unavoidable; in the latter, it is equivalent to willfully gorging oneself at an unattended candy box.

For some record producers, whatever technology makes possible, it also makes imperative.  It is not enough that a particular effect be available when needed, but it be used as often as possible.  For the producer who espouses aggressive multi-miking, this leads to a new approach to the recording process (as well as justification for the producer's job).  The composer didn't really know what he wanted, so the score becomes a kind of sonic 'tabula raza' (blank slate) on which 32 mikes and a talented mixing engineer can write their own view of the work.

Such producers make no pretense of claiming that the finished product — in the music biz, an album is called 'product' as if it were a can of paint or a box of bobby pins — is anything like what one would hear at a live performance.  It is, instead, 'a new musical experience', without any reference to live sound, and with every detail of the score mercilessly exposed, dotted and underlined, regardless of musical sense (or lack of same).

This approach can create embarassing situations.  A British composer was supervising the first recording of one of his works.  The producer pointed out that a certain oboe passage was inaudible and suggested 'spotlighting' the oboe at the appropriate time.  The composer politely informed him that said passage was not supposed to be audible, since it was a warm-up for a more difficult passage that followed.

This kind of 'sonic Lego-land', where the score is torn apart and reassembled, can give musically-rewarding results if it is applied to the right scores and with a modicum of taste and sensitivity to the composer's wishes.  The 'right scores' are usually operas, and the classic examples of doing it 'the right way' are heard in recordings produced by the late John Culshaw.

When stereo tape recorders became available in the middle '5Os, Culshaw realized it was possible to record opera with natural stage motion and depth, as well as an optimized balance among orchestra, chorus and singers.  In short, the producer could present the listener with a experience that was much closer to a live performance.  Incredible as it may seem, some other producers thought Culshaw was looney and recorded operas in stereo with the performers standing stage center, all the time!

In the late '5Os, Decca decided to record Wagner's complete 'Ring' (four operas which run over 12 hours and cover 19 LPs!) for the first time.  Culshaw knew that many of the sonic and visual effects Wagner desired could be achieved only imperfectly, or not at all, on the stage.  (This massive work, based on Norse and Germanic myth, abounds in magical effects and atmospheric scene-setting.)  So he went through the score, bar by bar, determining exactly what Wagner wanted, then deciding how it could be realized on record.

In most cases the decisions were trivial.  If Wagner wanted the Nibelungen to beat on a certain number of anvils with their hammers, then that number of anvils was obtained.  Where a particular type of peasant horn was required, the correct horn and a person skilled on playing it were called in.  And so on.

The controversial decisions revolved around Culshaw's representation of visual effects in sound.  For example, in Act II of 'Das Rheingold', Alberich becomes invisible and chases Mime around the stage, beating him with a whip.  Although there is no practical way of realizing this in the theater, it nevertheless was what Wagner wanted, so Culshaw simulated invisibility sonically by manipulating Alberich's voice to produce a hazy, other-worldly effect.

One effect in particular raised a real stink with some critics.  In his notes which appeared on the album jacket, Culshaw described the matter:

"The most challenging, difficult and controversial effect in 'Götterdämerung' occurs at the end of Act One, when Siegfried returns to Brunnhilde, disguised by the magic of Tarnhelm, as Gunther.  Wagner's stage direction here reads that Siegfried must sing 'with a feigned voice deeper than his own.'  (Some scores quote the word 'rauherer', which means harsher, rather than 'tieferer', which means deeper, and which is in the original libretto; but the intention in clear in both cases — that Siegfried should sound frightening to Brunnhilde, and should assume a baritone quality.)
"Now here, perhaps, Wagner was asking too much, or dreaming of some theatrical possibility beyond the scope of his time or ours.  He helped the tenor by writing the part as low as he dared, but on the stage one is always forced to the conclusion that the Tarnhelm's magic, though fine for turning people into toads, serpents and dragons, falls far short of converting a high tenor into a baritone.
"In this recording we have been able to do something in that direction — without, of course, resorting to transposition.  The voice is Windgassen's [the tenor playing Siegfried] — just recognisably so, but more by inflexion than quality.  The timbre has been changed, the whole colour of the voice altered toward a baritone structure.  (It has nothing to do with echo, by the way, nor is it a question of microphone characteristic.)
"The change is at its most startling when, following Wagner's directions, Siegfried reverts to his natural voice for the last four lines of Act One.  There will doubtless be disapproval from some quarters on the grounds that this sort of thing could not possibly be done in the theatre; but since Wagner was never deterred by the impossible if it was what he wanted imaginatively, and since it can be done on a recording, why should it not be done?"

Why not, indeed? An opera is conceived in principally dramatic terms, and I can't argue with the idea that anything which properly conveys the dramatic sense of the opera to the phonograph listener is justified.

By the time 'Salome' was recorded in 1965, Decca had developed an elaborate mixing console, giving even finer control over the recording process.  Culshaw wanted the listener to hear every detail of Strauss's complex score, especially those things that cannot be easily heard in a theater.  But there was no desire to sonically 'dissect' the music.  Again, in Culshaw's own words:

"Strauss enclosed Wilde's drama in music that is, at one and the same time, extraordinarily concise yet lavish in detail.  Mood is everything:  from the first clarinet notes one is plunged into the Byzantine night and the tension is never relaxed.  In approaching this score for a modern recording, we had one major idea in mind:  that, although its detail and subtle colouring are almost without precedent, the fact remains that much of it, and hence much of the record and intensity of the drama is lost in the average theatre.  'Salome' is a symphonic poem with vocal parts added,' wrote Gabriel Fauré in 1907.  'There is not a character whose physical individuality, whose morality, or immortality, whose thoughts and acts are not minutely translated.'
"To this end an altogether new technical approach was devised, through which every desirable detail in the score might become audible in its correct proportion to the rest of the vast orchestra (it averaged 115 musicians) and to the voices.
"In scoring 'Salome', Strauss asked for the most extreme blends of tone colour, not all of what are practicable in a theatre during the course of one performance, since their success in aural terms often depends on variable orchestral development.  The technique employed in this recording was devoted to capturing these nuances.  They are not exaggerated; they are not 'monitored':  they are exactly as Strauss wrote them, and it was gratifying to hear an eminent professor of music, a Strauss expert, remark [at] the first public play-through of this recording in Vienna:  'I had simply no idea until to-night of the fantastic detail that one should, but does not, normally hear.'"

The important thing to note is that it was Cuishaw's intent to clarify the composers scoring, and do so in a way that did not call attention to itself.  Moreover, opera can tolerate this treatment, not so much because operatic scores are more complex than purely orchestral ones (they may or may not be), but rather because operas are conceived as dramatic entities, rather than as a piece of music to be listened to purely as music.  In other words, the principal end of an opera is its dramatic, rather than solely musical, effect.

If this seems like a rather subtle distinction, consider that in opera, music is only one of many things (singers, chorus, sets, lighting, sound effects, dancers) which combine in the total theatrical effect.  But for jazz, chamber and orchestral music, the music itself is the total message.  And so it needs to be treated with even more respect.  Lack of respect for the music is not limited to the recording engineer or producer, but can extend to the conductor, sometimes in subtle ways.

Even in the concert hall, we do not always hear a piece the way a composer intended it.  Until the second or third decade of the 19th century, most works were scored for small orchestras (by modern standards) in small halls.  Modern orchestras are large enough for works scored for 100 or more players, and performed in correspondingly large halls.  So when performing 'older' works, the conductor may simply 'double up', in order to use all the instruments available (the players of which are being paid, regardless).  Eighty instruments in a large hall are obviously not the same as 40 in a small one.  So we're starting to see things such as Beethoven symphonies performed on original instruments with the number of performers LVB specified.

Of course, the opposite may happen; there can be too few performers, altering the sonority of large-scale Romantic works.  (If you've never heard a 30-player chamber orchestra of amateur musicians wading through Tchaikovsky's 4th, you've missed something.  You might search out the Hoffnung Festival records on Angel, which include such things as the '1812 Overture' played on kazoo.)  And to complete our consideration of every possible mismatch of score and orchestra, there is the classic example of Schumann's symphonies, so thickly scored that conductors often choose to cut the number of players per part.

Now, if the conductor and the orchestra (let alone the composer himself) can't always get the composer's intentions right, where does that leave the recording engineer?  Is he the final arbiter of matters musical?  Consider the some of the liner notes for CBS's recording of Bartók's 'Concerto for Orchestra' (MQ-32132), the first really controversial quad recording:

"This recording puts the listener in the best seat in the house...  It's the center of the symphony orchestra, inside the music — inside, perhaps, the creative spirit of Bela Bartók.
"In various sections of the third and fifth movements, mikes 22 and 23 were patched from track 7 into track 6.  In these cases, it was to gain needed antiphonal deployment between clarinets and bassoons as they related to flutes and oboes.  (We await the argument that clarinets and bassoons cannot move around during the course of a performance.  Yes they can, on a recording whose sole purpose is to clarify and intensify the musical performance.)"

These remarks are especially interesting when one considers the musical nature of Bartók's work.  The 'Concerto' alternates between treating instrumental groups in a concertistic fashion (contrasting them with the rest of the orchesta) and antiphonally (playing them against other groups).  If Bartók thought there was insufficient antiphony among the woodwinds and reeds, why didn't he do something about it?  He was free to request any instrumental deployment he liked.  So is the producer ever justified in second-guessing the composer?

To some extent, yes.  It's not uncommon for composers to admit that conductors find things in their works they didn't know were there, so one cannot exclude the possibility of similar revelations from 'creative miking'.  But there's a paradox involved with the 'Concerto for Orchestra'.  Although calling for 97 instruments, it's a light-textured piece which lends itself to the kind of sonic treatment CBS gave it.  (In richly-textured works, such as Mahler symphonies, I don't think we'd enjoy hearing the instruments dissected into groups and spread around us.)  But it is precisely the clarity of Bartók's writing which offsets any real need to surround the listener.  In other words, those works which can best tolerate a sonic dismantling are those which need it the least.

I guess my real objection to total surround sound for classical music is not so much that it shouldn't be done (after all, it is quite effective here, even if not really necessary) but rather that producers tend to think that it ought to be done.  I simply object to the technological cart driving the musical horse.

I don't completely agree with Boris Goldovsky that the sound at the podium is an undifferentiated assault — it can be quite exciting — but then again, I'm not a conductor, trying to make musical sense of the performance I'm directing.  Fortunately, this is not what the listener hears if the mikes are placed there.

As we all (should) know, conventional stereo is fundamentally flawed because it mixed the ambience with the direct sound in such a way that it cannot be separated or differentiated.  The effect of this misdirected reverb is to blur, soften and blend the sound, sometimes to the extent that it seems one is seated row M, rather than behind the conductor's head!  It is quite common for 'purist' engineers and amateurs (like myself) to place the mikes at the podium, and it is the sonic defects of stereo which save the listener from the aural assault described by Goldovsky.

Reprinted from MCS Review, Vol.6, No.3, Winter 1985.
Copyright © 1985, 2000 by Laurence A. Clifton

In the Autumn 1984 issue, Paul A. Alter wrote ...

Scott Henderson's letter (Summer 1984) takes me back to the 1950s, when the question of 'recording perspective' received a great deal of attention.  In those days of mono, it seemed that each of the major companies had its own favored perspective, ranging from 'first row of the first balcony' to 'aisle seat, row H', or even — in the case of the British — 'as if you were listening through an open window to a live concert in the next room.'  None of the major companies, however, recorded from the perspective of the podium ... and probably for good cause.

Boris Goldovsky — venerable and venerated opera producer, director, conductor, sage, etc. — has written about the terrible, undifferentiated, opaque wash of sound that assaults anyone on the conductor's podium.  It takes a special ear, according to Goldovsky, to make anything at all out of that noise.  Even Goldovsky himself, although an experienced opera conductor, does not feel inclined to venture out of the pit, onto the podium, to brave it.

I sympathize with Henderson and wish that he, and all the others I know who want to feel themselves in or in front of the orchestra, could have their wish.  (Perhaps the dominance control that Peter Fellgett discusses on page 12 of the Summer [1984] issue is the answer.)

But I want no part of it.  Until composers start writing compositions to be heard from the conductor's desk, I'll stick to listening to them all from where they were written to be heard — somewhere in the auditorium.

Record makers take note.  To me 'high fidelity' means fidelity to the composer's intention.  I want to hear it the way he/she meant me to (of course, this business of a barge out on the Thames ...).

In the Summer 1984 issue, Scott Henderson wrote ...

I would like to suggest that MCS Review run a poll to find out what type of surround sound the readers prefer.  I'm not talking individual systems, but recording techniques.  As an example, I personally prefer full surround, with instruments and ambience all around me (instead of limiting the back channels to ambience alone).  I would also like to hear some dead-center signals, such as bass, vocals and snare drums, mixed between all four speaker pairs. This would free the listener to face any direction he chooses.  It's hard for me to believe that so many new orchestral recordings done in multi-channel sound use the medium just to display ambience in the rear channels.  If rear-channel ambience is your thing, then why not just abstract it from stereo recordings using the QS Vario-Matrix Hall position, the Tate II, or stereo enhance on the Ambisonic decoders?  In my opinion, the best place to hear music is not the front row facing the audience; it's right up there with the conductor where the instruments and ambience surround you.

So how about it!  I think it's time we told the record producers what we want to hear from our surround-sound recordings, and this publication is certainly the right place for such an undertaking.  Perhaps a simple leaflet could be mailed with one of your issues which could be filled out and mailed back in.  The results of such a poll would be very interesting to all your readers.

Last updated: August 6, 2000

MCS Review On-Line Reprints