Bach’s Goldberg Variations, through Rheinberger

Semper-Aula, ETH Zurich, 2015-11-24

4-star rating

2015-11-26 — Original posting
2016-08-15 — Brushed up for better readability

TwoPianists (source:
TwoPianists — Nina Schumann & Luis Magalhães (source:

Table of Contents

A Personal Preamble

Let me say this at the onset: I have major reservations towards performances of Bach’s harpsichord works on a modern concert grand. Nothing against playing Bach on whatever instrument one has available, of course — but if I go to a concert to hear Bach, I have the expectation to experience Music that relates to, represents the composer’s intent.

What’s Wrong with Playing Bach on the Modern Piano?

Compared to what Bach could possibly imagine, or to what Bach must have pictured when writing his works, i.e., what we now consider “Bach’s original sound”, one loses a lot of “substance”. In the case of the harpsichord, the modern piano lacks the specifics of the harpsichord articulation. One ignores the specifics of harpsichord “language”: the transparency and clarity, the internal sound balance, the richness and brilliance of its sound.

Just as an aside: modern pianos are explicitly built with the aim to make the sound across the tonal range as seamless and equal as possible. This is quite different with early keyboard instruments (even up to fortepianos in the early 19th century). These feature more distinctly different “registers” across the range of the keys.

In summary: to me, playing works such as Bach’s Goldberg Variations on a modern piano can at best present an artists personal view on specific, limited aspects of Bach’s music. But ultimately, to me, it primarily remains music by Gould, Schiff, Sokolov, etc., rather than by Bach. In a way, it’s more of a momentary adaptation of, a “reinvention” according to a given work. Hence, it is not strictly comparable to performances on original instruments, such as a harpsichord.

There is Always Some Degree of Abstraction.

Of course, even when today’s great harpsichordists play the Goldberg Variations, we don’t strictly know whether this is how it sounded at Bach’s time. There is always some degree of abstraction from Bach’s original. Hence, where one draws the limit between “authentic” and “non-authentic” is ultimately a personal decision. While my personal limit is should be clear from the above thoughts: I can enjoy Gould, Sokolov, etc. playing Bach’s works. But just as their very personal (often ingenious) view on the baroque original.

Why this Concert (Still) Caught my Interest

Despite the above reservations, I was interested in attending this concert: it did not “just” feature “the” Goldberg Variations, BWV 988, by Johann Sebastian Bach (1685 – 1750), but a transcription for two pianos by Josef Rheinberger (1839 – 1901). That transcription was further edited by Max Reger (1873 – 1916).

Therefore, (to me) it is not “just” Bach’s music, but rather a late-romantic transcription. A “translation” that retains Bach’s original text as much as possible, but adds extra harmonization, additional voices, and so forth. As such, it does not claim to represent Bach’s original. It rather becomes an independent re-invention that deserves the term “authentic composition” by itself. That’s not to be confused with authenticity in interpretation, of course. In that sense, even for a “HIP purist” like myself, that concert featured a legitimate transfer of Bach’s work onto the (two) modern concert grand(s).

About the Artists

The artists in this concert were the pianists Nina Schumann (South Africa) and Luis Magalhães (Portugal). The two met while studying with Vladimir Viardo at the University of North Texas; they married and now live in Stellenbosch.

In 1999, they formed the duo “TwoPianists“, which since gas gained high reputation in the concert (and recording) scene. This evening, they played the above “translation” of Bach’s Goldberg Variations onto two pianos.

The venue was the Semper-Aula in the main building of the ETH in Zurich (see my earlier reviews for additional information on the hall). This is a small hall that ideally supports the sound of the two Steinway concert grands, especially with Rheinberger’s dense, romantic musical texture. Acoustic clarity and transparency were not the prime requirement here, but rather a rounded, full, harmonious soundscape. That’s what we could indeed enjoy in that concert!

About the Sole Work in this Concert

The theme in Bach’s Goldberg Variations, BWV 988 (“Aria mit 30 Veränderungen“, i.e., an Air with 30 alterations) is a simple, two-part Aria in G major (16 + 16 bars, both parts with repeat signs). The 30 Variations are organized in groups of three. The last variation in each group is a two-part canon, first in unison, the second one with the two melodies at different pitch, a second apart, then a third, etc., up to a ninth in variation 27. For variation 30 see below. Most canons have an extra accompanying voice).

The variations in-between the canons are a rich collection of popular musical forms, such as a french Overture, dance movements, fughettas (little fugues), slow forms, as well as virtuosic variations, covering the entire spectrum of keyboard music and techniques. Bach clearly arranged the variations such that they form build-ups, big arches, from the simpler, playful first variations, or from slow, more meditative variations in G minor, up to dense and virtuosic, resounding (particularly on the harpsichord), even brilliant moments towards the end. The cycle ends in calm serenity, not triumphantly, though — see below.

About the Transcription

As much as I could see from looking at Bach’s original score, Josef Rheinberger has taken over / integrated Bach’s score in its entirety, essentially unaltered (with few exceptions, such as occasional octave shifts). One component of the adaptation consists of the distribution of the score onto two instruments. This offers the benefit that in variations such as No.5 the artists can avoid the crossing of hands. Also, segments that Bach wrote for a harpsichord with two keyboards can be played effortlessly on two instruments, without entanglement of fingers.

Performance & Interpretation

TwoPianists played the Aria and its 30 variations (plus the Aria da Capo) essentially without pause (i.e., attacca). They played all repeats, without exception: over 75 minutes of music in a single performance.


The first bars were still purely the original Bach score, played on one piano only. But then, after 8 bars, the music is seamlessly switching over to the second piano. In the second part of the Aria, very inconspicuously, some discreet additional notes further fill the harmonies, making the harmony richer than in the original. The basic harmonic structure remains untouched, though.


In terms or ornamentation, the transcription and its interpretation stay relatively close to the original. Rarely, additional ornaments are added, occasionally (primarily in repeat sections), appoggiaturas (and written-out transitional notes) are left out. This may feel somewhat counter-intuitive: for baroque music. most commonly, one expects the artist(s) to play the first pass is as is. In the repeat, ornaments may be added at the player’s discretion. I liked the fact that the pianists avoided an excess of romantic rubato. The moderate agogics used by the two pianists remained in line with baroque (and Bach’s) music as perceived and practiced today. On a harpsichord, even stronger agogics may be in order.

Harmonic “Filling”, Additional Voices

The distribution of Bach’s 2 – 4 voices onto four hands and two keyboards doesn’t just allow for additional “harmonic filling”. It also opens the opportunity for melodic / polyphonic intensification. Frequently, Rheinberger adds additional, perfectly fitting, and often virtuosic voices. Such parts would typically be impossible to play and manage on a single keyboard (let alone a harpsichord).

Dynamics, “Choreography”

Compared to the original, the sound overall was dramatically intensified, altered through the concert grands, towards a romantic soundscape; the instruments may open up the temptation for dynamic excesses, which the artists luckily avoided. Within a given variation, the dynamics more or less followed the density of the texture. On a larger scale, I felt that Rheinberger enhanced the build-ups that already Bach conceived into big, harmonious, large-scale arches across many variations. But he retained Bach’s resting points in the form of slow, more introverted variations in-between (see below for more notes on those).

Intensification, Quodlibet and Final Da Capo

Over the whole, I got the impression of a polyphonic and expressive intensification towards the last, virtuosic variations. That’s in agreement with the original score. Bach ultimately (in variation 30) transfigured this intensification into an ironic quodlibet. That’s a kind of hodgepodge of quoting several popular (almost vulgar) folk songs in the Aria‘s harmonic framework. Within this quodlibet, the music calms down towards the Aria da Capo. The return of the Aria (simplified in the ornaments) then offers a serene, peaceful closure to the entire cycle.

Max Reger’s Contribution?

In the aftermath, I asked myself about Max Reger’s contribution / editing in this (I did not check the sources). My impression was that the music is largely Rheinberger’s transcription. But I felt that occasionally (particularly in the second, denser part), Reger could not resist “intensifying” (“loading”) one of the other chord in his style. Even if that guess is correct, Reger added these chords in a very discreet manner. There’s nothing comparable to the dramatic, heavily dissonant transitions into a final harmonic ending that we see in some of his organ works.

Pitfalls & Shortcomings of the Transcription

Returning to my initial remarks, I cannot resist adding some critical notes, largely / mostly on Rheinberger’s account / contribution. Compared to the original (and with the Aria, after all), the tempo sometimes appeared slightly exaggerated towards romantic virtuosity. But OK, this is the romantic view of Bach’s variations.

On the other hand, for example in variations 13 and 25, the “translation” falls into the same trap as many pianists playing Bach’s original version on the modern piano. The slow 3/4 meter appeared altered into a fast 12/16 meter, in which Bach’s ornamental passagework Becomes the melody voice. In variation 13 the slowly striding duet in the left hand (in Bach’s harpsichord score) is stretched out beyond recognition. As stated, many, if not most pianists make this mistake. So, Rheinberger and TwoPianists are in good company here.

Finally, in the quodlibet (Variation 30), Rheinberger’s intensification, i.e., the denser texture, is obscuring the ironic character of the underlying, popular songs.

A Final Comment on the Performance

The two pianos (Steinway D) were facing each other, such that the two pianists were sitting at either side of the stage area. They could not see each other’s hands. However, the visual contacts over the rim of the score proved to be entirely sufficient. It was clear that the two artists were intimately familiar with the other’s internal rhythmic feel. In retrospect I seem to remember maybe one, two short moments where the coordination may not have been entirely perfect.

On top of that, over the years, the two players obviously have mutually adapted their playing. They have done this to a degree that made it difficult to figure out who was playing. At least from the rear of the hall I sometimes could only tell from watching their hands. Sure, the acoustics have helped rounding the sound and mixing the contributions of the two instruments — but that doesn’t diminish the quality of these artist’s performance: congrats, and thanks for such a nice & interesting concert experience!


The artists played two encores:

I very much appreciate the artist’s desire to return something in reward for the lively applause. But as in other concerts, I also somehow regret that this partly covers up the fresh memory from the last piece in the program. Especially here, where this had such a calm, serene ending! But OK, the applause itself does that, too. Hence, the “purist’s only solution” might be to ask people not to applaud…

Addendum 1

For the same concert, I have also written a (much shorter) review in German for See also the note at the top. This posting is not a translation of the Bachtrack review.

Addendum 2

TwoPianists have recorded the above transcription of the Bach Goldberg Variations for their own record label; I have not reviewed or listened to this CD yet, the information is provided merely for reference:

Bach / Rheinberger / Reger: Goldberg Variations — TwoPianists; CD cover

Bach: Goldberg Variations BWV 988,
arranged for Two Pianos by Josef Rheinberger & Max Reger

TwoPianists (Nina Schumann, Luis Magalhães)

TwoPianists records TP1039213 (CD, stereo); ℗ / © 2013
Booklet: 12 pp., English

Bach / Rheinberger / Reger: Goldberg Variations — TwoPianists; CD, EAN-13 barcode
amazon media link

AboutImpressum, LegalSite Policy | TestimonialsAcknowledgementsBlog Timeline
Typography, ConventionsWordPress Setup | Resources, ToolsTech/Methods/Pics/Photography

Feel free to comment — feedback is welcome!