Beethoven, Diabelli Variations, S. Kovacevich (piano)
(released on February 10, 2009)
Onyx 4035 | 53'31"
Beethoven, Diabelli Variations, P. Lewis (piano)
(released on June 14, 2011)
HMC 902071 | 52'46"
Beethoven, Diabelli Variations, B. Levin (piano)
(released on June 28, 2011)
Centaur CRC 3046| 60'21"
Beethoven, Diabelli Variations, L. Cabasso (piano)
(released on November 15, 2011)
Naïve V5282 | 50'40"
Beethoven, Diabelli Variations, A. Staier (fortepiano)
(released on May 8, 2012)
HMC 902091 | 67'30"
Most of the things we admired about Paul Lewis's Beethoven sonata cycle are true also of his recording of the Diabelli Variations. Lewis has a vigorous way with all of the variations, pushing the extremes of tempo and articulation, not to blow them out of proportion but to bring their contrasts into focus. His approach is in line with the booklet essay of scholar Jean-Paul Montagnier, which describes the variations -- almost all of them lasting only a minute or slightly more -- as something akin to bagatelles. The contradiction is shocking, between a vast architecture and the individual building blocks, which are often drolly terse. In the same way that he did with the sonatas, Lewis does not attempt to overwhelm the piano with vicious bite either.
Stephen Kovacevich, in his second recording of the Diabelli Variations (the first was made in the 1960s, for Philips), comes close to Lewis's very short timing. By contrast to Lewis, many of Kovacevich's tempo choices seem a little too fast or too slow, as he looks for more daring or unconventional or enigmatic interpretations, not necessarily the most polished technically and often overemphasizing both the jolting and the unctuous. Even shorter than either Lewis or Kovacevich is French pianist Laurent Cabasso, whom we have read about but never reviewed until now -- and the experience has me wanting to hear much more. The approach is a more impish take on the variation-as-bagatelle, with some real romps in the fast movements -- just as virtuosic as Lewis and maybe a little odder, captured in that warm Naïve sound with all of the crunch in the fortes and the plushness in the softer moments. He pairs the Beethoven with Schubert's own single variation on Diabelli's waltz, plus another technical showdown with the Wandererfantasie.
Beth Levin comes in with the longest timing in this batch of recordings from the last few years, approaching the time of another fine and enigmatic recording by Grigory Sokolov (live, Opus111). The caution in Levin's approach is evident in her lilting, waltz-like, but ultimately not very exciting rendition of the original waltz and the ponderous rendition of the first variation, marked Maestoso, but also Alla Marcia, a shocking shift of meter and character for the initial variation. Levin adds many nuances to each movement, drawing out some interesting ideas, but among the competition this rendition does not hold my ear for very long. Too much flair, as on the trills of Variation 6, is missing, the maestosos and pesantes are all exaggerated, and inferior sound quality does not help.
Sokolov takes slightly longer than Levin (62'07"), but rather than sounding cautious, his choices are just odd and fun and often marked by almost impetuous virtuosity: Variation 1, just as slow as Levin, is pompous, with some distant echos from the cannons; no. 2 walks on eggshells; and it goes on from there, as all sorts of unexpected voicings and textures pop out at you. While Levin's version is slower in almost all the movements, the idiosyncratic reading by Piotr Anderszewski (Virgin, 2001) is mostly as fast or faster than other recordings. To add up to the longest timing of 62'51", Anderszewski takes just a few movements at outrageously slow paces, like Variation 14, which is the most Grave e maestoso I have ever heard. All of the slow movements are slower and longer than most other recordings, but Variation 31 is taken just as glacially as no. 14, also ending up over six minutes in length. All in all, it is one of the stranger Diabellii Variations available on disc.
The newest recording at this point is by Andreas Staier, made on a reconstruction of a Graf fortepiano, an instrument that still has a thunderous forte side, albeit not as strong as a modern piano, but also a beautifully nuanced soft side. As demonstrated at recitals by Richard Egarr and Davitt Moroney in the last couple months, playing on historical instruments, especially when they are actually instruments the composer may have known, can help illumine our understanding of the sounds and effects the composer was after in a piece. The modern piano can just do some of the demanding things better and more easily -- the trills all sound a little clunky and wooden -- but anyone who has an interest in this piece, either player or listener, should listen to this recording. The use of the moderator (forerunner of the una corda pedal) and shift pedal (Verschiebung) in Variation 20 creates an almost otherworldly soft sound, and the bassoon stop (touches of reedy buzz adding a sung quality) in the comic Variation 22 and the janissary stop (a crash of percussion on big chords) in Variation 23 are not to be missed.
The only other fortepiano version of the piece to have reached my ears was by Edmund Battersby (Naxos, 2005), also on a modern reconstruction of a Graf instrument, although not as exotic in tone as the one Staier plays and not captured in the same quality of sound. Battersby does not spice up the score with exotic stops, and the fortepiano version is paired somewhat oddly with the same artist's performance on a modern Steinway. Furthermore, Staier rounds out his disc with eleven variations by some of the 50 composers Diabelli solicited variations from, plus Staier's own Introduzione to the theme, based on ideas Beethoven sketched out in his manuscript but did not ultimately use, leading immediately to his performance of the whole Beethoven set.