This post is part two of a group that began here.
I’m currently in rehearsals for a performance of the ever-popular Ravel orchestration of Mussorgsky’s Pictures at an Exhibition.
I doubt there’s any classical fan reading this who hasn’t heard this beloved warhorse many times- in fact, there’s no doubt that Ravel’s version has long since eclipsed Mussorgsky’s original piano piece in popularity. Out of the entire piece, of course, the last movement is the best known- “The Great Gate of Kiev.”
However, this very well known piece offers us a interesting example of the simple challenges of reading music. Perhaps after thinking about this example, you might conclude that you’ve rarely heard the notation for this music read accurately.
If one looks at a page of the score, you’ll quickly see that there are lots and lots of grace notes. Grace-notes are one of the most troublesome bits of standard notation, not because their meaning is imprecise, but because their meaning is flexible.
Other kinds of notes, crotchets, eighth-notes, whatever you like to call them, all express mathematical relationships to the unit of pulse. In the Great Gate, the rhythmic language is quite simple and foursquare, so all of the rhythmic relationships are easily figured out by the players and conductor.
Grace-notes, on the other hand, express a duration of time that can only be read given the context that they are in. In classical music (Mozart and Haydn), we have elaborate rules for knowing when a grace-note is on or before the beat, and what it’s duration is. In this music you quickly learn that a gracenote does not simply mean to play the note as quickly as possible.
On the other hand, very often grace-notes should be played very close to the beat and very fast, so often that many musicians forget that is not always the case.
The problem in orchestra is that often we play them so fast that they are no longer heard at all.
So how does one know how fast to play a grace-note? Is it always as fast as possible? If that’s wrong, how are we supposed to know that it’s wrong? Surely this is an example of the limitation of notation?
Nope, sorry. The problem in “Great Gate is that the orchestra version is a transcription, so the performers are reading the notation out of context. Notation creates context, so notation out of context loses some of its clarity. If one goes back to the piano version, you can see that the pianist has to jump and reset the hands in a new block chord after the grace note, so the grace note has to be played before the beat and not very fast. Most orchestras play these notes either so fast they’re not heard at all, or even worse, on the beat (this is a mis-reading of Ravel’s indication to play the grace-notes down bow in the strings near the end. Musicians look at those down bows and think “aha! he wanted those on the beat,” but what he wanted was for them to be really, really loud so they would have a similar prominence to what they have in the piano version).
Ravel could easily have omitted the grace-notes altogether as his not limited to having on one person to play all the notes in the chords, and, when you don’t hear them in an orchestral performance, you wouldn’t know you were missing out on them. However, if you look at the score, for instance the last two pages, you can see that he took a great deal of care to transcribe the grace-notes in the piano version as honestly and imaginatively as possible. He knew they were an important part of the original, so HE MADE THEM AN IMPORTANT PART of the transcription. (In fact, there’s a lot of important harmonic information in the gracenotes. At first Mussorgsky just uses them to lay down an e-flat pedal, but later, the harmonies move in the grace-notes).
If one knows or has even played though the piano version, you won’t be tempted to play the grace-notes any faster than a fairly broad eight-note, not the thirty-secondish note you usually hear. You’ll also know that, as the writing gets more massive and the leaps in the piano part get bigger, the grace notes must get slower and heavier.
Have a listen first to a decent performance of the orchestral version, then visit Evgeny Kissin’s piano performance. Do the fast, largely in-audible grace-notes in the orchestral performance still seem like an accurate reflection of the notation?
Graace notes speed should adjust to the acoustics of the hall.
You might take a view that Its a coloring or blending to imbue ambiguity or suspension. Same with trills…
Similar considerations apply to the other “inexact” notations… Take the trill, for example: Pianists know a trill that lies low in the left hand should generally be slower than a trill in the upper register. This might be a case where the acoustics of the instrument suggest a realization, but there are also cases where the musical context should guide. A sensitive tympanist might, for example, play a roll more slowly in one of those Mahler “trauermusik” sections than he might in a more martial setting.
Hi Bill
You’re absolutely right, and the parellel might then be looking at what happens with a trill when you orchestrate a piano work.
Brahms often used piano writing to dictate rubato- if he writes a leap to chord that is impossible to play in time, he clearly did not intend the performer to play it in time. In the solo piano works, that’s not a problem, but in the chamber music you’ll sometimes here string players react snootily when the pianist says “I need a bit of time for that leap.” Brahms knew the pianist would need that time, and probably put that leap there to make the time happen, because there were probably other ways he could have written it to make the leap easier.
Then imagine what happens when you transcribe a Brahms piano work for orchestra, like Schoenberg’s transcription of the G minor quartet. How many conductors who do this piece really know the piano writing of the original well enough to judge the rubatos that Brahms wanted?
Cheers Bill, and thanks for reading.
Ken