Stravinsky's Firebird "Revision"
Here’s a rant on a niche topic: In 1919 and 1945, Stravinsky arranged his Firebird ballet into suites for concert performance. Both times, for copyright reasons, much of the music was reorchestrated as well. Possibly reflecting the legal rather than creative origin of the changes, in my opinion, the “updated” orchestrations are much worse than in the original ballet.
The idea for this post was from a visit to the Hollywood Bowl, where an okay LA Phil performance of the 1919 suite, along with Britten’s Sea Interludes, and Tchaikovsky’s Violin Concerto took place.
The original ballet is scored for a massive, Mahler-scale orchestra. 16 woodwinds, 7 onstage brass, 3 harps, a piano that’s barely used1, and a large percussion section, surely stretching the limits of the orchestra pits whenever it’s performed. The suites’ scorings are reduced for a more standard orchestra, but my issues with them don’t really have to do with that.
Here’s a recording of the 1919 suite for reference (both of the suites’ orchestrations are mostly similar to each other):
and the original 1910 ballet, for comparison:
Here’s a list of complaints (my opinions, of course):
- 2:58 (1919 suite) / 2:47 (original ballet): Piano sounds less “enchanting” than the original’s celeste. I do like the use of the piano in the following L’Oiseau de feu et sa danse, where it and a harp replace the celeste and three harps of the original. It’s a different sound, still good, though the white-key glissandos on the piano are very dissonant with the rest of the orchestra, which is playing in 6 sharps.
- 10:00/33:01: More of a comment on the selection of excerpts and how they’re (not) stitched together. I can see the intention behind going from the quiet ending of the Khorovode to jumpscaring the audience with the Infernal Dance, but it just doesn’t quite work for me. The buildup to the Infernal Dance in the original ballet is just too good, from the violins + English horn line being built up to an incredibly ornate section, before everyone but the timpani and basses drop out. To cut that part out entirely when it’s not even that long seems wrong. Plus, it’s nice for one of the percussionists to be able to show off that xylophone excerpt they’ve practiced so much for auditions.
- 10:38/33:30: Piano+flute is weaker than the original piccolo clarinet. And what’s with the pauses that stall the momentum?
- 11:00/33:45: As a general rule of thumb, the orchestral piano usually shouldn’t double just a single voice, or at the very least should double with octaves and not just one note at a time. (An exception might be arpeggiated chords). It (usually) sounds awkward if it’s done that way, as I think it does here; maybe that’s a consequence of the piano not being as naturally melodic of an instrument as orchestral ones2. I also have no clue why the piano stops playing every other measure.
- 11:52/34:45: Comparing the scores, you can see the more sparse orchestration of the suite compared to the original. No harps! I feel that this passage with its gorgeous string writing benefits with a more dense accompanying orhestration.
- 12:21/35:11 Something about the added timpani rolls sounds off to me. It could just be this recording, but it feels like it’s overwhelming the main texture of the “machine gun” brass and strings.
- 13:00/35:49: Again, like 11:38/33:30, what’s up with the missing first beat? I can maybe see the intention, but it’s just distracting.
- 13:07/35:57: The orchestration of the brass is reversed here; in the original, the trumpets and trombones loudly proclaim the three-note figure three times with the orchestra, then the horns are completely alone on the last statement. In the suite, the horns play with the rest of the orchestra while the trumpets+trombones are alone. I could go back on forth on which is preferable, but right now I appreciate how the quieter (but still sharp) sound of the horns leads nicer to the following section.
Fortunately, the berceuse and finale are mostly left alone; some of Stravinsky’s best and certainly most romantic string writing remain. Still, there are a couple changes that are worth pointing out:
- The section of the original ballet preceding the tremolo chords right before the finale (41:53) is left out, unfortunately. A shame, because the build-up of the first inversion D-sharp minor chord is implemented so well. Instead of just being one note after another going up and up as a less clever composer might, Stravinsky goes back down the chord once a measure, repeating notes before going back down again but to a note higher than the last time. The tremolo rumble provided by the timpani and basses provide a feeling of transformation as the chord expands from low viola to high violins.
- The original ballet’s finale has some string writing that would make any music school orchestration professor absolutely lose it if a student wrote it (see 45:03 for example). The double stops require so much note-to-note movement and stretching that I imagine actual musicians playing this would play them divisi despite the explicit instruction not to, because of how awkward they are. The suites thankfully change this into something much more playable (20:45).
- The 1945 version does something completely different with the final maestoso statement of the finale theme. It’s certainly a choice.
Now that the original ballet is public domain, I’d be interested to see a version that’s maybe scaled down to improve feasibility of performance (2 or 1 harps, more standard woodwind section, etc.), but otherwise adheres to the original orchestrations as much as possible. This way, a group performing the work can get closer to Stravinsky’s original orchestration, but without the extra length associated with performing the full-length ballet, including the more transitory or recitative-like moments that would make more sense with theatrical scene changes and dancers acting out a story. That said, I think even those moments are musically sound on their own, and contain some of the best moments such as the aforementioned transition to the Infernal Dance.
I do think these opinions would hold even if you changed the recordings (i.e I don’t think these are not issues with the specific recordings). I do wonder to what extent these feelings would change if live performances were compared instead, if the effect of the changed orchestrations make more sense in an in-person concert hall setting. Hopefully this post isn’t that wrong…
-
The piano and celeste play at the same time at least once, so you need separate players for both to adhere to Stravinsky’s original orchestration. That said, it wouldn’t be the end of the world to eliminate the piano part entirely since it always doubles another instrument, unlike the instrument’s huge rule in Petrouchka. ↩
-
Due to the piano’s inherent attack-and-decay sound. Now that I think of it, that’s exactly why piano fits in so naturally with the percussion instruments. I’m sure many people far more knowledgeable than I am have written volumes about the orchestral piano. ↩