Liszt’s Sonata in B Minor is such a magnificent and famous piece that huge number of recordings have been made over decades, let alone countless recitals by so many pianists. It is an interesting exercise to compare performances of the Sonata by many well-known pianists, and the differences of interpretation amongst them often can be so significant and astonishing.
Here I’m sampling a few performances found on Youtube; and to make the comparison somewhat meaningful, I will pick a few passages from the piece, as shown below:
Performance by following artists are used for the comparison:
Claudio Arrau (https://www.youtube.com/watch?v=FMjZqEBBseo)
Andre Laplante (https://www.youtube.com/watch?v=VCHE-UPwBJA)
Yuja Wang (https://www.youtube.com/watch?v=qnPMXk_GaN8)
Krystian Zimerman (https://www.youtube.com/watch?v=TKoAN426YAo)
Valentina Lisitsa (https://www.youtube.com/watch?v=w65QgjWHNDA)
Khatia Buniatishvili (https://www.youtube.com/watch?v=S4D3GfF2jOs)
Vladimir Horowitz (https://www.youtube.com/watch?v=H_6Rkm-ZvsI)
Evgeny Kissin (https://www.youtube.com/watch?v=VfHgLyuO0hY)
Just as any performance is infused with much of the performer’s own interpretation and feelings, such exercise of comparing those performances are highly subjective. It is by no means a judgment call on how “good” a performer is, rather only an expression of personal taste on the piece and how (I think) it should be performed.
Subject A (m.32)
This is the beginning of exposition section, presenting the main subject A on the tonic key. As we know this subject combines the forceful theme B (m.8) and the bold hammer-blow of theme C (m.13), giving a sense of tension and defiance. Such mood is probably better expressed by highlighting the four hammering notes (marked as “marcato”), and indeed that is how most good players handle it. Overly fast tempo will sound sloppy and reckless, too much a slower tempo will not bring out the energy required to portrait the intended character.
To me the better ones are by Wang, Laplante and Zimerman, with proper tempo and accents on theme C; the slight ritardando leading to bar 32 opening is just right to underscore the arrival of the theme. Here’s Laplante:
I wish he placed a just bit more emphasis on the hammer-blow, which is what Wang does here and somewhat better:
I’m surprised to hear what Horowitz rendered in this 1977 recording, in which the passage came out with an awkward, almost stuttering quality:
What’s completely “off the base” is that of Buniatishvili (and Lisitsa as well), with extremely fast tempo rushing through the notes, letting the fine details of the theme’s character completely squashed:
Allegro Energico (m.205)
This passage is one of a series of transformations on theme B (m.8), which is prominently featured throughout the entire piece; the theme’s variation intermingles with those of subject C, the lyrical cantando espressivo (m.153). With a recitative-style and fast ascending scale in bar 204, this passage rides on that crescendo wave and enters with a powerful cord (similar to its introduction at bar 8). The ensuing measures are filled with passion leading to the climax in bar 232.
Zimerman does a great job here, the crescendo of ascending cords (e.g. m.209~212) is appropriately prominent, and every stoke of the fast triplets on the left hand sounds clear yet all smoothly flow through:
Both Wang and Arrau do an almost equally fine job, too.
Contrasting with the above are below two unpleasant rendering, each goes sort of the opposite of the other. In Horowitz’s case, the progression is bit tardy, but more problematic is that it does not sound continuous, sometimes you hear sudden slowdown in between bars. Some people may like it, but not me:
Now here’s the other extreme, the excessive fast tempo with a boisterous sound lacking not only proper dynamic variation but clarity of phrasing (for example the left hand descending cords starting at bar 221):
This is another ingenious transformation of subject C (cantando espressivo, m.153) after the previous development on theme B above. The character of this passage is such a drastic departure from the original material; that tranquil singing melody is now turned into stiff and noisy shout-outs, leading back to the strong claim of theme B at bar 286.
Laplante does a superb job:
But I prefer Wang’s interpretation a bit better, her slightly slower yet steady tempo with strong staccatos is perfect:
Kissin overall played it well but I think he started acceleration a bit too early from bar 263, whereas the stringendo marking doesn’t start until bar 272:
The one by Lisitsa simply just fly through the passage too fast, we hear the loud thumping quickly passing by, without getting (feeling) much out of it. :
Middle Climax (m.397)
This climax occurs around the middle of the development section, which begins with the Andante sostenuto at bar 331. It is essentially a new subject (D) based on theme B. After recalling the lyrical subject C, this passage, starting at bar 363, gradually builds momentum on the Grandioso (subject B), repeating twice, then invokes a passionate set of octaves that leads to a glorious climax at bar 397, in uplifting major mode, which also starts repeat of the Andante subject, with a pleasant sense of relief. A good interpretation I think should highlight the progressive buildup of tension and the gratifying release at the climax.
That’s why I think Arrau’s rendering here is excessively slow, as if to make announcement every step of the way, and it actually undermine the emotional power it would have otherwise:
I’d consider a faster tempo more appropriate; the key is, it should present a structure, an emotional buildup and release, regardless the exact playing time being a bit longer or shorter. Laplante plays the fastest here compared to others, but overall structure is there (though I wish bar 398 not as rushing):
Zimerman and Wang both do well, with Wang sounds more elaborative on the octave buildup. Here’s Wang:
and here’s Zimerman:
In case of Lisitsa, even though her playing time is longer than Laplante and Zimerman, the passage comes out rather simplistic and flat; it lacks the gradual dynamic ups and downs to depict the emotional tension and release:
Coda Climax (Presto – Prestissimo, m.673)
The general consensus is that the coda section starts from bar 682 and soon leads to the final climax of the work, bar 710. The energy starts building up actually from bar 673, where the opening theme A is called out on powerful octaves with double forte force, followed by the Presitissimo, which is a transformation based on theme B, with a condensed rhythmic pattern on fast octave progression; after the octaves are repeated on left hand, they push through a frantic line of chromatic scales that leads to a final proclamation of the Grandiose, with the power of full triple forte, and highlighted twice by the sforzando on the last cord (m.701 & m.703), the followed by the ascending cords all the way to the climactic final note at bar 710. With a long pause, the Andante sostenuto repeats with a much lighter tone, shifting the mood down to quiet and solemn reflection of the opening measures of the Sonata and fading to the end.
It definitely sounds better to start the first measure of the Presitissimo (m.682) a bit slower then pick up the pace; it gives a sense of the “beginning of the end”, so to speak. That’s exactly how Arrau did it, execution was superb:
For Wang, I wish the Grandiose repeat had been a bit clearer with the first sforzando (m.701) standing out a bit more, otherwise it’s equally excellent:
Zimerman’s is very good as well, started clear and balanced, until bar 700 where the pace suddenly becomes lagging, the power of that final climactic wave surge turns out not being as strong and convincing:
As for Horowitz’s interpretation of this passage, I don’t like it at all. The descending scales of theme A came in with a sudden soft crescendo, and also sounded fragmented, which is completely out of character; the final two lines of ascending scales just rushed through without any of the necessary slowdown in tempo, as a result the ending lost that triumphant sense of climax:
In Buniatishvili’s playing, I only hear a passage of fast, loud yet emotionally flat line of notes shuffling through my ears; for example, the Andante was by no means any softer than previous bars:
In summary, it is almost impossible to single out one performance that can be deemed the “best”, simply because not all movements, not all sections, and not all passages for that matter are played equally well by a performer. Overall on this list though, I’d put Laplante, Wang and Zimerman on the top tier.