Bach Aria, and Some Recordings of the Well-Tempered Clavier

Let me preface this post with what I take to be a wonderful coincidence. As some of you know, a new Bach aria was recently discovered. If you click on this link, you can go to the NPR story on it, and even listen to it. The aria is the first discovery of a vocal composition of Bach in over 70 years, and it was performed by National Public Radio 292 years after the year of its first performance. The story of its discovery is interesting with some amusing twists. But what I think is the best thing about this discovery is the fact that the announcement was basically made on my birthday. It's a rather nice birthday present, and it gave me an added incentive to write these brief notes on recordings of the Well-Tempered Clavier

* * * *

Recently, I've been comparing recordings of Bach's Well-Tempered Clavier in an effort to figure out who has the best recording of the work. Of course, in order to determine what's the best recording one has to develop some sort of criteria. This is actually rather difficult to do. A number of different factors influences one's reception of a recording. There is obviously the skill of the pianist, but there is also the quality of the recording, the quality of the mixing, whether the microphones have been placed properly, the hall, how well the recording was remastered, the question of whether or not a historically accurate recording is better than a modern variant. Also there are considerations about the stereo that you choose to play the recording. As I discovered listening to a number of different recordings, some sound far better on particular sorts of stereos. There are further complications with Bach. Unlike more contemporary composers, Bach gave very few instructions as to how his pieces should be played. Tempos are a matter of conjecture, and the same pieces can be played at vastly different speeds. Further, how should one play a piece that might have been written for a clavichord on a modern piano?

Before I say what recording I like best, let me take you through the recordings I listened to.

First, there was no doubt that I had to include what many people take to be the standard recording of the WTC: Glenn Gould's recording from the 1960s. I love these recordings, as do many other listeners. They are very well-engineered, and they sound very good on most stereos. Further, there is no doubt that Gould was one of the greatest performers of Bach the past century has given us. But, I think that Gould's recordings have become so popular, that many people now fail to recognize what's distinctive about Gould's recordings. Some people have just come to think of Gould's sound as Bach's sound; they assume that when they listen to Bach they are having transparent access to the way Bach wanted his pieces to sound, or if they are more agnostic when it comes to historical accuracy, they think that Gould somehow reveals to us the heart and core of baroque music in a way no one else can. But I think that it is important to realize that what made Gould's recording of the Goldberg Variations in 1955 such a revolution is the fact that Gould played Bach in a new way that is often quite idiosyncratic. I like to think of Gould's performances as a modernist reaction to his predecessors. I think that the best contrast is seen against a lesser-known player like Walter Rummel. Walter Rummel plays Bach much in the way that one would expect a Romantic pianist to play Bach: he uses lots of peddle to create dissonance or to smooth out lines, and incorporates dramatic changes in emotion and intensity. I don't know much about Gould's early education, but I could imagine that his style is a reaction to this sort of Romantic excess. By usually adopting a peculiar almost staccato style of playing - which apparently was produced by the unusual way Gould held his wrists below his fingers - and by eliminating peddles, and what's more, by eliminating much of the emotional contours to individual melodic lines, he presented a very pure style of playing that showed a tremendous amount of respect for counterpoint. His style of playing created a revolution.

Like most revolutionaries, I happen to think that Gould ended up producing a product that went a bit too far in the opposite extreme from his contemporaries and predecessors. For one thing, I don't think that Gould shows sufficient respect for the phrasing of the melodic lines in the pieces that he plays. Bach was obviously concerned with this, since Bach's own introduction to the Inventions says that one should most of all try to develop a cantabile style of playing ("am allermeisten aber eine cantable Art in Spielen zu erlangern"). Gould's particular style of playing is perhaps one of the reasons why Gould famously thought that the preludes in the WTC were not particularly interesting, and instead concentrated on the fugues. A fugue seems a very natural composition for Gould to play. His style of playing means that none of the lines in the fugue dominate the others, and that it is possible to clearly hear every note as it is played. But the preludes require a different sort of playing. Here we don't have the complex counterpoint of the fugues, and instead we often find a more unified phrasing structure, such as that required by a particular dance form. And I think Gould is at his weakest with this sort of phrasing structure. Sometimes he just races through a prelude in order to make them a little more challenging, at other times he omits repeats to get to the fugue a little faster. (Of course, I generalize a bit too broadly. For instance, Gould takes the prelude to the B-flat minor prelude of WTC slower than just about anyone else I've listened to, and he occasionally can bring out a melody in a wonderful way, I am obviously speaking here of general impressions.) And what's more, so far as Bach's compositions are an exploration of the particular qualities of each of the keys in the Well-Tempered scale of the 18th Century, I don't think Gould does a great job of bringing these qualities out. This probably has to do with the way his peculiar style of playing de-emphasizes the phrasing of the melodic line, and the harmonic connections between notes along the "horizontal" lines of the piece. But there is no question that Gould's recordings should be on the shelves of anyone who truly appreciates Bach's WTC. And sometimes his modernist interpretation of Bach can sound just as refreshing as it did with his monumental 1955 recording. After listening to twenty minutes of B minor fugue recordings, when Gould's begins it seems a gust of fresh air has entered the room, and one just is called to attention. Perhaps it was this quality that led NASA to include a recording of Gould playing the Well-Tempered Clavier on each of the Voyager Space missions. Since in general variety is something to be praised, one has to praise Gould inventiveness. Nonetheless, Gould is not my favorite recording of these pieces.

Another set of recordings that I've been listening to are those of Ralph Kirkpatrick in his recording of the first book of the WTC. The unique thing about Kirkpatrick is that he plays the piece on a clavichord. Since the clavichord was THE instrument for secular keyboard playing in 18th Century Germany (In contrast to the harpsichord, which was more popular in France and Italy), these recordings have special historical interest. Bach himself may have played these pieces on a clavichord.

It is worth making a few observations about clavichords. I have only had the opportunity to play one once (I hope to play one again at some point), so most of my knowledge of the instrument comes from hearing it on these recordings. Unlike the harpsichord, the mechanism by which a clavichord makes the sound allows for an added degree of expressiveness in terms of the sound quality of the note and the dynamic range. For instance, it is possible to produce a slight tremolo effect by wiggling the key a little when one hits a note, and it is also possible to produce louder tones by striking the keys harder. The sound of a clavichord is also closer to that of a guitar. But clavichords have a number of weaknesses. For one thing they cannot project sound very well, and thus need to be played in smaller spaces than the piano. (Harpsichords also are better adapted to smaller places: much of the beauty of a harpsichord's sound is lost in even a moderately-sized concert hall. In a larger space a harpsichord often sounds very percussive, and the beauty of the rich resonance of the instrument is lost. If you want to hear a recording of what a truly beautiful harpsichord sounds like, I would suggest Pierre Hantai's Opus111 recording (The cover to my copy is different, but I think the preceding link goes to the right album.) of Bach's Goldberg variations: a recording that I think may well be my favorite recording of the piece - I should note that unfortunately Hantai's Re-recording of the same piece is not nearly as good, get the Opus111 if you can.) To get back to the clavichord, it also has the misfortune that the mechanism itself can be distracting. This is something that one notes with Kirkpatrick's performance. The recording is full of clicking sounds and squeaks produced by the instrument itself.

But Kirkpatrick's recording is quite good. He plays difficult pieces with hardly a hitch - see the further notes on this below - and plus one gets to hear all the extra resonances that a clavichord can provide. Minor keys sound darker and richer, and major keys can have an unusual brightness at times. In listening to this recording, one really gets a sense of why Bach chose to explore the different keys in the way that he did. Although Kirkpatrick is not my favorite recording, I think that it could also arguably have a place on most listeners’ shelves, if only for the shear joy of hearing Bach performed on the clavichord.

I also had the opportunity to listen to Sviatoslav Richter's recording of the entirety of the WTC. Richter is without a doubt one of the greatest pianists of the 20th Century, and therefore any recording that he makes of a piece of music deserves a hearing. Richter's recording of the WTC is also very good in itself. Although Richter does play Bach in a more Romantic style than any of the other recordings I listened to, it is not an excessive Romanticism - more like a Prokofiev than a Rachmaninoff - and represents a nicely balanced midpoint between the recordings of Gould and those of Rummel. Unlike Gould, he can bring out melodic lines with remarkable effectiveness, and even in a complicated fugue he can manage to make each of the voices sing clearly and independently. And although he allows himself some freedom when it comes to tempos, and dynamics, his rhythms are often nearly perfect. So, although his version of the C# minor fugue from Book I shows some signs of his Romantic piano training, it still manages to be a masterpiece of clarity and precision. What's more, the engineering on his recording may well be the best of the recordings I've listened to - certainly on a par with the engineering on Gould's recording - on a good stereo the sound is truly divine. And Richter is also the only one of the recordings I've listened to that is performed on a Bösendorfer. I don't know how many people have had the opportunity to play this brand of piano, but it does have a very distinctive sound. It's hard to describe. The piano body is often more square than that of a Steinway - a little closer to the shape of a harpsichord - and because of the shape I've often described a Bösendorfer as having a sort of "square sound," although I've never been able to quite explain what I mean. I suppose it is a fantastic way of saying that the notes are quite sharply delimited and clear. This sound brings an interesting perspective to a performance of the WTC.

When it comes to tempo, Richter is closer to Tureck than Gould, and he seems to relish playing the repeats, and in what I'd like to call the "construction" of a sound canvas: the different voices in a fugue are like different colors in an abstract painting, and Richter seems to enjoy arranging them in the way an abstract painter might arrange color on a canvas. We hear one line being played smoothly with the notes connected, and another line of the fugue is played with a sharp and loud staccato, and when a third lines enter, the preceding lines change to accommodate their new arrival. I think this does obscure the voices in fugues a bit, but the effect can be quite appealing - particularly, late at night.

So Richter's is also a recording worth adding to a collection. In particular, people with very good stereos might want to own this recording if only so they can relish in the beautifully pure sound of the Bösendorfer and the level of professionalism that the engineers bring to bear on the recording. But I don't think Richter is my favorite recording of these pieces. For one thing, the rich canvases of sound he paints often dilute the strength of the rhythmic lines, and therefore detracts from the forward momentum of the pieces he plays. This is not always the case, but more often than not I think it is. Nevertheless, his recordings are masterpieces, and worth listening to.

Yes, I've been saving the recording I think is best for last. As far as I am concerned no one can bring out the beauty of the WTC better than Rosalyn Tureck. It is worth making a few comments about Tureck (See the Tureck Bach Institute website for additional info about her). On the one hand she is a skilled textual scholar who often consults original manuscripts of the pieces that she plays, and she has almost perfect control of the more technical aspects of Baroque keyboard works: she has a remarkably contolled pace and momentum, and can play perfect cantabile phrasing. Her trills are also very precise, and two simultaneous trills often interlock perfectly. Although Tureck is a scholar and remarkable technician when it comes to baroque keyboard music, she doesn't sound as if she is confined by these traditions, and one really feels that the music itself is what dictates her style of playing. Of all the recordings I discuss here discuss here, I find that Tureck's recording is the one that really reveals the quality of the key and the individuality of each of the pieces. Many times another performer will emphasize the same things that Tureck does, but they are just brought out more distinctly and clearly in Tureck's recording.

Perhaps her greatest masterpiece is her performance of the F minor prelude and fugue from the second book of the WTC. She takes these pieces incredibly slow. For a comparison, Glenn Gould plays the prelude in just over a minute and a half: 1 minute and 45 seconds. Tureck however takes over 7 minutes to play the same piece: 7 minutes and 18 seconds! And one doesn't feel like it is too slow - something I occasionally feel when Gould or Richter takes a piece very slowly. It is true there are places where I don't think her slow tempos sound appealing to a modern listener - her recording of the Goldberg Variations, for instance. But for the WTC and for the Partitas, no recording seems better.

If you want to hear more about Tureck's philosophy of playing, the best place to look is her educational piece: An Introduction to the Performance of Bach. Among the unusual things that she explains there is her use of the pedal. Tureck uses the pedal to bring out some aspects of the piece, but she uses it in a way that conforms with the musical structure of baroque music: not to blend notes together, or to emphasize dissonance, but to emphasize a cantabile style of playing, and to bring out a particular quality of a note. Unlike a Romantic pianist who might hit the pedal before or during the execution of a note, Tureck steps on the pedal AFTER she plays the note, in order to catch the trailing resonance of the note. In this way, the initial articulation of a note is sharp and precise, but one still appreciates the way it's sound can blend with other notes that are played concurrently. I really don't know how often Tureck does this, but sometimes you can hear it. I think far more usually her cantabile phrasing is sufficient on its own to achieve this effect. Anyway, this is an interesting technique that is faithful to baroque musical structure, but nonetheless makes use of the full sonority of the piano.

Almost every piece Rosalyn plays seems nearly perfect. Her performance of the C# minor fugue from book I was initially an exception to this for me, and perhaps brings out a weakness of her recording: the engineering on the album isn't very good. I first heard it on a very good stereo, and was surprised how jarring the sound quality was. If you have a $5,000 dollar stereo and you are just looking for a recording to show off what your stereo is capable of, you may want to skip the Tureck. Her recording really needs to be tweaked with an equalizer on really good components. And it occasionally can sound disastrous on some cheaper stereos. You should really try her album on a few different stereos before making a judgment. These facts are really surprising when one considers Tureck was recorded in the 70s - you think they could have done much much better. Anyway, the best sound I've got so far from this recording has been from my old Nakamichi RE-10 receiver with an MB-8 compact disc player and Acoustic Energy One speakers - a good sort of beginner's level high-fidelity system (although my CD player has serious problems - far too many skips for my taste). But, another thing that made me initially not like Tureck's recording of the C# minor fugue, is a note at the climax of the piece that can sould like a clear mistake. This occurs in the very difficult five-part fugue at measures 100-101 of the fugue. At this point all five voices of the fugue are in play, and there is a complicated combination of half notes, quarter notes, and eighth notes that seems to demand some accent since it is the most intricate part of the fugue. In a five-part fugue one wants to bring out each of the voices, but one only has so many fingers! Here Tureck's playing seems to stutter slightly. I've been trying to figure out what is going on there, and I'm guessing she plays the first chord in measure 101 with one note ahead of the others. Certainly, none of the other recordings I talk about do this. Gould plays the piece fast enough and staccato enough that he doesn't get caught in this thicket. Richter gets through it by playing a simple and solid chord, and thereby de-emphasizes the difficulty of the individual melodic lines. Surprisingly enough, the best recording of these bars seems to be Kirkpatrick. He shows no hesitation, and doesn't overly emphasize anything, but runs through the measures without a hitch. Although, there are elements of the recording to obscure this: unfortunately with five voices going at once there is a lot of squeaking and clicking coming from the clavichord keyboard - one can't play Kirkpatrick very loudly on a stereo here! Anyway, I am uncertain at the moment whether to say Tureck makes a clear mistake in these measures - perhaps someone who is more familiar with this piece can tell me whether she was merely trying to bring out a rhythmic feature that I am unaware of? Anyway, I do wish that she managed to come closer to the smoothness of Kirkpatrick, or Richter here. Her faithfulness to the melodic lines gets her into a bit of a thicket. But the rest of the piece is so good this detail really doesn't matter. It is perhaps a testimony to the quality of these recordings that I am forced to limit my complaints to single notes!

So, I think that no one performs these pieces as well as Tureck, and that she has been legitimately hailed as the "high priestess of Bach" (New York Times). But there is something to be said for not limiting oneself to her recordings. As Tureck herself says, she thinks of Bach's music as fundamentally abstract. What Bach does is give us a template of well-composed pieces, and leaves performers a great deal of freedom in how to bring these pieces to life. I think that there is a parallel with Shakespeare. I have been in a number of different productions of Shakespeare, but I always marvel at the freedom that is left to a director and to the actors of a production. This perhaps is an accidental feature of the way the texts have been transmitted to us: history has withheld many details about performance, thereby leaving many important decisions completely in the hands of the director and the actors. Bach's music gives us a similar freedom: there are no designated tempos, there are very few phrasing marks, and no phrasing remarks. Performers are called upon to make these pieces their own. And each of the recordings that I've listened to manage to convey a lot about the performers. It is very easy to distinguish each of these keyboard player's recordings - even disregarding things like Gould's occasional vocal accompaniments, or the fact that Kirkpatrick is on a clavichord. So, there is an advantage to listening to more than one recording of the WTC, because the pieces are as much defined by their differences as their similarities.

As an aside, I remember being in a piano master class in high school at Carnegie Mellon. I wasn't a particularly good pianist at the time, and was probably the worst pianist in the room, now that I think about it. I certainly wasn't playing anything difficult at the time, being mainly a composition student. Anyway, in one of our discussions in the master class the professor asked us what we thought about Bach's music. Much to my surprise, everyone thought that Bach was no fun to play. His music was described by a number of students as "mere technical exercises" or as not particularly interesting, unexpressive. Everyone was much more in favor of Rachmaninoff and Chopin. I think I well may have been the only exception to this other than the professor himself, who reprimanded the class for undervaluing some of the greatest keyboard music ever written. I don't know why students thought of Bach as mere technical exercises. I would guess that it was a combination of how they were introduced to Bach through piano lessons (almost everyone starts with Bach), and some prevailing prejudices about what baroque music should sound like. For me, however, no music is as expressive or emotionally powerful as Bach, mainly because it is so abstract: the abstractness leaves space for emotional expression of the widest variety. Often - perhaps under the influence of Gould and some comments he has made - people think of baroque keyboard music as unemotional and technical. It is true that a fugue doesn't leave the composer as much room for variation as other less demanding forms provide. But, it is important to recognize that Bach could attain great emotional depth in his compositions: his cantatas and choral works in particular emphasize this. There are the glorious dark tones of "Aus der Tiefe," which Brahms made so much of, or there are the bright happy tones in the setting of "Mache dich, mein Herze, rein" in St. Matthew's passion. Furthermore, Bach's instrument par excellence was an instrument capable of enormous depth and power: the organ. With an organ one can produce sounds so powerful that they can shake a cathedral. It is difficult to believe that Bach could do all this, but would revert to a confined, technical, and unemotional playing style when it came to the clavichord, harpsichord, and fortepiano. Hence, I personally would encourage performers to appreciate the expressive ability of Bach's music, although it is true that some restraint is actually necessary: one doesn't want a pianist to mangle the rhythmic or harmonic structure of a work. But then it is often true that tieing something down to a structure can be a vehicle for emotional expression. And, there is a wide range of possibilities in the performance of Bach, possibilities that enable performers to make their recordings deeply personal. And perhaps ultimately it is who we are ourselves, as subjective listeners, that determines what the "best" recording is, for just as performers bring themselves to the performances, so we bring ourselves as listeners.

Let me end this posting by mentioning a few games that can be played with the WTC. If you load your albums onto itunes, it is possible to make a playlist that enables you to listen to each of the prelude and fugue combinations by each performer in a series. So, for instance, you can play Kirkpatrick, Richter, Gould and Tureck all playing the C major prelude and fugue. You can then listen to the corresponding A minor prelude and fugue, then perhaps listen to the A major fugue. This seems like an excellent way to come to know the sound of each of the different keys. One has to remember that in addition to being a great pianist and composer, Bach was also a great teacher. How many young pianists start out playing a piece from the Anna Magdalena Büchlein or the Applicatio in C major from the Wilhelm Friedemann Clavierbüchlein? I think that Bach's foremost role may well have been that of an educator. This is perhaps the reason why Bach's music often seems so transparent: in addition to performing melodies Bach always seems to be trying to teach his audience something. For instance, when you hear something in F minor, it feels like it could never be written in any other key, and that Bach is trying to teach us something about the different keys as he composes for them. I suppose this has been one of my two main criteria in evaluating these performances. The other one is just the shear musicality of the performances. In both these categories I don't think anyone is as good as Rosalyn Tureck. But there are many more recordings out there for me to listen to. I suppose Landowska and Schiff are next, but I also would like to listen to Keith Jarrett and Hantai's versions more closely.