Thursday, October 07, 2004

Orchestral Arcana

Recently, in listening to records and attending concerts, I've become more aware of differences in orchestral layout, and in particular the advantages of dividing the violin sections: first violins on the conductor's left, second violins on his right. This is not the standard layout, but it's becoming more and more popular, and I'll sometimes find that a conductor and/or orchestra has "converted" to the seating.

Up until the early part of the 20th century, most composers would have expected the violin sections to be divided left and right, and they accordingly wrote in a lot of antiphonal effects, with themes thrown back and forth between the violin sections. Elgar's "Enigma Variations" has many such effects; the second variation, "H.D.S-P.," begins with the first violins playing a phrase that is then echoed by the seconds, and the variation mostly consists of the firsts and seconds bouncing phrases back and forth. When the violins are divided, as they would have been in Elgar's time, this is a great antiphonal effect, as you can hear in recordings by Adrian Boult and Pierre Monteux. When the first and second violins are seated together on the conductor's left, this effect doesn't really come off, and on a recording, it's completely lost (because it's hard to hear the first and second violins as distinct sections). There are many other examples, e.g. Mahler's ninth symphony, where the opening theme is begun by the second violins, not the first violins, which -- with divided violins -- has an element of surprise built into it, because the tune isn't coming from the side of the hall that we subconsciously expect a violin melody to come from.

As illustrated here, Leopold Stokowski (in the '30s with the Philadelphia Orchestra) was the one who really popularized the practice of seating all the violins together on the left; he did it, in part, to create a big, rich, string-heavy sound. Record producers liked the seating because it made an orchestra easier to record (I'm not sure of the technical reasons for this). More importantly, violinists liked it because it made it easier for one section to hear the other, and because it didn't require the second violinists to sit, in effect, with their backs to the audience. I think the split-violin seating became more and more inconvenient as orchestras got bigger and bigger.

So with a combination of pressure from the musicians and the record companies, and probably other factors that I'm not remembering right now, more and more conductors started switching to the massed-violin setup. But it was kind of gradual. For example, in Fritz Reiner's earliest stereo recordings with the Chicago symphony, you can hear that the violins are divided left and right; a few years later he stopped doing it. The ironic thing is that by the time stereo recording came in, there were few split-violin setups around to take advantage of stereo separation.

There were a few holdouts who just refused to give up the divided-violin seating. Toscanini did it to the end of his life (though he didn't get to record in stereo, so you can't tell this on his recordings); so did Otto Klemperer; so did Pierre Monteux; and so did Adrian Boult. Boult actually wrote a letter to the GRAMOPHONE magazine in the '60s, soliciting readers' opinions on whether he should continue with the seating; he'd gotten so many complaints from orchestral violinists who didn't like the seating that he was unsure whether it was worth going on with it. Among younger conductors, Rafael Kubelik was one who liked the divided-violin seating enough that he tried to use it with any orchestra that would let him (all his recordings with the Bavarian Radio Symphony Orchestra, including his Mahler cycle, use the seating).

The seating started to creep back in in the late '70s/early '80s, but it was a haphazard thing. Now it seems to be gradually on its way to becoming the norm again. Most of the period-instrument groups use the setup most of the time. Among the modern-instrument conductors who have adopted the setup are James Levine (with the Met orchestra), Michael Tilson Thomas (in San Francisco), Riccardo Muti (at La Scala) and Daniel Barenboim (in Chicago).


Clifford Loo said...

Kubelik at first didn't do it this way (c.f. his BRSO recordings of the Weber overtures, Mendelssohn's A Midsummer Night's Dream, and the Dvořák 8th). Other early proponents of this layout include the Boult disciple Vernon Handley, Norman del Mar, Giuseppe Sinopoli, and the trumpeter-turned conductor Gerard Schwarz, who managed to make a US orchestra, the Seattle Symphony, adopt it where Kubelik and Sinopoli had failed (with the Boston Symphony in Ma Vlast and the NYPO in R. Strauss, respectively).

nick said...

Just stumbled upon this very informative post. The flaw with your analysis is that most modern concert halls are too large for the antiphonal violin seating to make an acoustic impact. Unless we're sitting close to the stage, we hear mostly reflected sound, and not directly radiated sound. In the smaller, dryer halls of Beethoven's time, it's a different story. So it seems that any perceived acoustic advantage of antiphonal violins is psychological. In reality, the antiphonal effects usually work worse, because the seconds have their f-holes facing the wrong way. This DOES reduce the amount of reflected reverberation, and it means that the 2nd violin "answer" usually sounds weaker. (In order for the voices to sound equal, the seconds need to play about 125% louder than the firsts, which entails a strong second violin section and also a conductor with a good ear for balance). This is rarely a problem when the firsts and seconds are adjacent, as the f-holes are all pointed outward.

But antiphonal "gags" (like the ending of Beethoven 7) are the exception rather than the norm; more often, the first and violin sections play together, and antiphonal seating makes this much more difficult. Playing difficult, rapid passagework together is an impractical challenge, especially when an ensemble is on tour and has to quickly adjust to an unfamiliar or problematic hall. Other problems of ensemble crop up as well: the cello and bass are moved to the left side, which means they are now on opposite sides from the lower brass.
Finally, in Romantic music, the second violins often play an octave lower than the firsts, supporting their sound and stabilizing the intonation. To this end, the conductor will often ask the second violins to play their lower octave louder. With antiphonal seating, it's hard for the firsts to hear the seconds, and as a result, the massed violin sound is usually thinner and less in tune. For juicy Romantic music, it's rarely an attractive option.
That being said, I'm not sure why anyone actually endorses antiphonal seating. I think part of it is based on an unreflective appeal to tradition, saying "this is how it was always done" without taking into account how concert halls have changed, and so on. Everyone likes to give the "antiphonal effects" reason, but I've hopefully demonstrated how this reasoning is based on acoustically unsound principles. If anything, antiphonal effects work better with the modern seating....

Nick said...

P.s. I agree completely about the antiphonal effect being great on stereo recordings, but this is precisely why I think mono recordings are, in principle, superior. Stereo recordings rely on an acoustical fallacy that we experience directly radiated sound, and this is not how concert halls work. In a room or small hall, sure, but not when you're sitting in the balcony at Carnegie Hall.