Estampes:

La soirée dans Grenade

Claude Debussy

Original recording by Ricardo Viñes:

My recreated performance:

Cross-cut video:

Annotated score:

The recreating experience

Much like with the preceding two fragmented Debussy recordings, I have marked the general tempo modifications of Viñes by thick red arrows in the annotated score. I have also marked the most relevant examples of “dislocated attacks” between the melody and accompaniment, but not all – as this is once again a general practice in Viñes’s playing. Some “areas of interest” are marked by red circles and explained in detail in this text. Omitted notes in Viñes’s performance are marked with red crossings. Once more, I chose to divide Viñes’s use of “arpeggiation” and “dislocation” into three categories: 

1. “Dislocated” chords - (“un-rolled”), marked as red circles with arrows showing the “dislocation.”
2. Expressive arpeggios - used to separate the melody from the chord notes, or to create expressiveness in the melodic line.
3. Directional arpeggios with "dislocated" melody - marked as arpeggio lines with arrows to demonstrate if they are played upwards or downwards, and with blue circles and arrows showing the “dislocated” melody notes.

At the beginning of the piece, Viñes establishes the habanera rhythm. Although not quite as brisk as in the Albéniz Tango-piece, it has a slightly rushed pace to it. Despite this, there is some flexibility in the timing of the rhythmic character, where the slower bars are slightly restrained, whereas the faster ones are closer to the “double-dotted” tendency I found in the Albéniz Tango-piece. Viñes quickly releases the C sharp octaves in the beginning from the keys and creates the legato sound through a rich use of the sustain pedal. It is very interesting to observe that these octaves are not “arpeggiated” in bars 2-4, but constantly “arpeggiated” in bars 5-14. This once more supports my claim that Viñes is using “arpeggiation” to emphasize the polyphony in chords/intervals which consist of independent voices. Thus, he clearly separates the habanera rhythm in the upper middle voice from the high C sharp notes. The melody in the left hand from bar 7-14 has a very dominant character which includes a subtle use of rubato. Through his polyphonic playing, Viñes separates the three individual voices in this section not only by dynamics and articulation, but also by giving them an individual timing (“polyphonic rubato”). I have only marked the general rushing/slowing tendencies which applies to all voices in the annotated score. In the Tempo giusto section, Viñes greatly rushes from bar 17-18 and 19-20. There is also in my experience a very subtle use of “dislocation” on the first beat of these bars. In bar 21, Viñes once again uses “arpeggiation” to separate the top C sharp note from the habanera rhythm. It is interesting to observe that Debussy marked the two octaves in bar 22 with arpeggios. Could it also be Debussy’s intention to distinguish between the individual voices of the octaves? Bars 23-28 are seemingly quite straight forward, but the sonority was quite tricky to achieve. It is Viñes’s vast use of the sustain pedal that makes this section difficult to imitate. As in earlier recreations, I found that I had to stop playing the chords tenuto with the fingers and rather make brisk and accentuated attacks, so that the legato is made through the sustain pedal. In addition, the chords must be “dislocated” from the habanera rhythm to achieve the clear polyphony we hear in Viñes’s performance. From bar 38, Viñes plays very briskly and accentuated. He “arpeggiates” the chords on the first beat of bars 38-43 and then continues with the “unrolled,” “dislocated” approach.” In bars 39, 41 and 43, he does not play the second semiquaver A natural note in both hands (although in 41, he slightly "dislocates" the A’s in both hands on the second beat so that it almost sounds as if written, hence the parentheses in the annotated score). We have seen similar kind of such “re-arrangements” with Viñes before. It might be that Viñes felt that the second semiquaver (A natural) was weakening the energy achieved by the close-to-double-dotting approach on the preceding two E natural’s. It could also be a result of the hierarchy in Viñes’s playing – meaning that these notes are simply rushed in favor of the main melody. From the last chord of bar 41 and onwards, the chords in the right hand becomes polyphonic. Consequently, Viñes does not “arpeggiate” the chords, but rather “dislocates” the octave melody from the harmonic middle notes (marked in blue in the annotated score). This practice is carried on throughout bar 49. In this same section, he also omits the middle voice chord notes on the second beat of bars 45-47, although the identical notes from the first beat are more or less sustained throughout the bar by the pedal. He does however add middle-voice chord notes on the second last semiquaver octaves in bars 48 and 49. Thus, the entire section of 45-49 is executed in the same manner as bar 44. What is less so in line with this apparent consistency though, is bar 44, where Viñes plays a D natural instead of the written E natural in the habanera rhythm of the left hand. This is one of the rare cases where I suspect that Viñes does play a wrong note, because I can find no other logical explanation. In bar 50, Viñes plays in the right hand both the E natural that is the third note of the triplet and the semiquaver E natural which is doubling the habanera rhythm of the left hand. In other words, he plays two E naturals to separate the voices from each other. This is very much in line with what Ignacy Friedman is doing frequently in the Chopin Nocturne. The entire section from bar 51-58 is played with a very subtle “dislocation” in the right hand, so that the melody note at the bottom of the chord is very slightly dealyed. Thus, it is clearly separated from the accompaniment in the sonority. In 59, we see an example of a downwards arpeggio, probably to make the habanera rhythm delayed from the high C sharp. In bar 60, Viñes shifts into “arpeggiating” in a typical upwards manner. Bars 61-66 are played very similar as 23-28, but now with the high C sharp on top of the habanera rhythm of the left hand in bars 62, 64 and 66. It is interesting to note that bars 64 and 66 are marked with arpeggios in the original score. In my 2006 Durand edition, bar 62 also has a marked arpeggio in parenthesis[1]. I would argue that Viñes’s practice of “arpeggiating” these octaves can be seen in relation to his “arpeggiation” in bars 5-14. Debussy’s arpeggio indication in 64 and 66 is in relation to this very much in line with Viñes’s coherent practice of emphasizing the polyphony. Bars 67-75 is another good example of a section that is far more complex than it seems at first. I believe that Viñes is “dislocating” the chords from the bass in 67-68, but then, when the chords become polyphonic in bar 69, he starts to “dislocate” the octave melody. However, the tempo is fast, so the two practices occasionally sort of melt together. I practiced this section very slowly with an exaggerated use of “dislocation”. In tempo, the “dislocation” felt more “gestural”, in a similar manner as with the high tempos in the Borodin and Scarlatti pieces. By the end of bar 73 and through 77, Viñes goes back to “dislocating” the entire chord. The section from bar 67 to 77 is also a place where he applies rubato, both by slowing and rushing. The proceeding section of bars 78-91 is very complex, and ended up being the most time consuming part in the recreating process. In 78-81, the habanera rhythm is now located in the middle of the left-hand thirds. As we have seen before, Viñes is often using “dislocation” and “arpeggiation” to bring out such middle-voices. The standard practice is that he will delay the most important voice in the polyphony so that it is played lastly in the sonority. In this section, he manages to clearly bring out both the main melody and the habanera rhythm. Furthermore, he manages to maintain the coherent characteristic rhythm in the polyphony without letting it be affected by the chords (which are played by using both hands). Once more, this is made possible by the interplay between “dislocation,” accentuated as well as brisk finger attacks, and a rich use of the sustain pedal. In a slow-motion perspective, he plays the voices in the following order: 1) the notes that make up the habanera rhythm, 2) the thirds in the left hand, 3) the right hand chords. All these notes are played briskly and accentuated. They are not held down by the fingers because this will “muddy” the sonority in the pedal. From bar 82-86, we arrive at what is in my view the most complex section with regards to polyphony and sonority in the entire piece. First, Viñes briskly “arpeggiates” the half note chord notes (in bar 82 these are the A natural and B sharp in the left and right hand) so that the C sharp of the habanera rhythm in the left hand is delayed. He then plays the octave melody of two D sharps (which is also “dislocated” and thus delayed both from the chord notes and the habanera rhythm). Like the preceding bars, all these notes are released quickly from the key and “captured” in the sustain pedal. And after all these stages, we are still only on the first note of the triplet! Viñes continues to “dislocate” the octave melody so that they are delayed after the C sharps of the left hand. In bar 84, matters get even more complicated: in addition to the practice seen on the first beat of bar 82, Viñes now plays the second beat of the bar by letting the C sharp from the habanera rhythm come first, then the middle chord notes of A sharp and B in the left hand and the A sharp in the right hand, and finally the octave melody of two D sharps in the right hand. The proceeding bars are played in a similar manner. My markings in the annotated score will hopefully make some sense of this. All of these very subtle “dislocations” are played very fast, and I would not have been able to hear the order in which they are played, had I not experimented with several different approaches in different tempos. My solution is the one I believe came closest in imitating Viñes, and this is a good example which shows that it is indeed possible to extract more information from the recordings by imitating and actively engaging in them, rather than just listening. In bars 96 and 97, Viñes plays an arpeggio in both hands on the first beat. The second low C sharps in the bass are omitted. In bars 98-106, the sonority becomes complex again. In line with what I have already described, Viñes “dislocates” the octave melody from the middle voice chords, as well as the habanera rhythm chords in the upper register. The low E natural notes in the bass are also “dislocated” from the right-hand chords. All of this is played with a rich use of the sustain pedal. In bar 109, Viñes plays the C natural on the first beat before the bass line. He then greatly rushes bar 110-111. In 115-118, Viñes “arpeggiates” the left-hand chords and rushes so greatly that he seemingly omits the top note of the four-note chords. In 122-127, Viñes embellishes the guitar-like arpeggios by adding and changing the notes of the chords, not unlike the ending of the Debussy etude. In bar 122, instead of the written C sharp, A, C sharp, E, A, C sharp, E - he adds an E between the lowest C sharp and A, so that it becomes C sharp, E, A, C sharp in the left hand (in other words he adds the fifth of the A major chord). In 124 and onwards, the written C sharp, A, D, F sharp, A, D, F sharp - is changed to C sharp, F sharp, A, C sharp, D, F sharp, A, D (in other words, two complete four-note chords of F sharp minor + D major). These arpeggios are released quickly from the keys and sustained in the pedal. In 128, Viñes adds an arpeggio on the first beat and in 129, he “dislocates” the melody in the left hand by playing it after the first beat and before the second.In the final bars of the piece, the octave melody notes are very subtly “dislocated” from the chords. In 134, the chords are not tied from the preceding bar.

Comparison to other recordings

To compare the Ricardo Viñes playing style in this piece to other, I have chosen three recordings. In addition to this, I also included the Welte reproducing piano roll recording of Debussy himself.

In Daniel Barenboim’s recording, with a duration close to two minutes longer than that of Viñes, it becomes very clear that Barenboim is playing in a much slower tempo.[2] In addition, Barenboim also has a tendency of generally slowing down in his rubatos. Also, Viñes plays the different sections of the piece with a sense of “tempo relationship” between them, Barenboim interpret these sections in greater contrast to each other tempo wise. There seems to be no intentional use of either unmarked “arpeggiation” nor “dislocation” to emphasize the polyphony.

In this recording of Sviatoslav Richter, we hear a faster tempo, but still quite far from Viñes’s.[3] Also in this recording, there is no practice of intentionally adding either unmarked arpeggios nor “dislocating” the voices to emphasize the polyphony. Also, Richter plays rather strict in time and does not add significant rubato in his performance.  

In the historical recording of Marius-François Gaillard from 1928, we hear a general tempo that is very much in line with Viñes’s.[4] With regards to Viñes’s use of “arpeggiation” and “dislocation”, we do not hear this to be a coherent practice in Gaillard’s playing. There is an exemption to this though: in bars 82-89 he has a practice that is very similar to that of Viñes, clearly separating the polyphonic material by “dislocated attacks.” In bars 96 and 97, he does add an arpeggio to the first beat of the bars, just like Viñes, but otherwise he is more restrained in his use adding such things. Although Gaillard’s playing is clearly polyphonic, it lacks some of the “polyphonic rubato” we hear in Viñes. However, his more “balanced” use of rushing versus slowing down is very much in line with Viñes. Indeed, they seem to share some ideas on the general form of the piece’s timing, and from bar 109 and throughout the ending, there is a remarkable likeness in the extreme use of rushing and slowing down – even at the expense of notes and beats. The arpeggios in bar 122 and onwards are also surprisingly similar both in the context of the sonority and the timing. Like Viñes, Gaillard does not tie the chord from 133-134, whereas both Barenboim and Richter plays as written in the score.

In the words of Robert Philip from his book Performing Music in the Age of Recording:

Subtlety is a quality which comes across from Vines's recordings, subtlety not only of pedalling, but also of the layering of dynamics, and the placing of rhythms, in order to clarify textures. In Debussy's 'Soiree dans Grenade' (1930), he sets up the habanera rhythm at the opening with a slightly overdotted rhythm. The melody below it is played with delicate rhythmic freedom, so that it seems quite independent of the habanera rhythm. Comparing this with Debussy's own piano roll of the same piece, it is tempting to think that Debussy's rhythmic freedom would have sounded similarly subtle if accurately reproduced.[5]

Because the Welte reproducing piano roll recording of Claude Debussy is often referred to in discussions of performance “authenticity” of La Soirée dans Grenade, such as in the 2006 Durand edition of the score, I decided to include it in this comparison.[6] I must however stress the carefulness that should be taken into consideration when analyzing piano roll recordings. There have been several discussions about the value of piano roll recordings in the relation of performance practice sources. Neal Peres Da Costa provides extensive information about the various reproducing systems and their strength/weaknesses in his book Off the record: Performing Practices in Romantic Piano Playing.[7] Without going too far into this discussion, there are examples where performers from the early 1900s made both piano roll recordings and acoustic recordings of the same piece, such as Edvard Grieg and Ignacy Paderewski. Without analyzing these in detail in this project, what we can say for sure is that the various reproducing piano roll systems were not able to capture the more subtle qualities in historical performances. What we can dissect from the Debussy piano roll recording in this perspective however, is the use of “dislocation” and “arpeggiation”. Although we may not be able to hear the subtle differences in the practices, we can nonetheless establish if the practices are present or not. Also, although the tempo of the piano roll playback does not necessarily capture the true spirit of the tempo of Debussy’s original performance, it is possible to at least generally relate the different tempos to each other. In other words, we can hear if a section is played faster or slower in relation to the general tempo. Based on these aspects, I shall roughly examine if traces of the Viñes’s performance style can be found also in the Debussy reproducing piano roll recording – or to reverse this question: If Debussy’s performance style is heard in the recording of Viñes.

Listening to the Debussy recording, we can immediately establish that this is slower than Viñes’s.[8] Whether this is Debussy’s original tempo or if it has been modified in the piano roll production or playback is open for discussion, but I will leave it with that in the context of this research project. What is interesting in the beginning though, is that we hear that the melody in bars 7-16 is clearly separated from the other voices because it has its individual rubato timing. Also, Debussy “arpeggiates” the C sharp octaves and thus “dislocates” the habanera rhythm from the high note. Both aspects are also heard in Viñes’s performance. In the Tempo giusto section, Debussy greatly rushes between bars 17-18 and 19-20, just like Viñes. In 23-28, we clearly hear how he “dislocates” the chords from the habanera rhythm. The second Tempo giusto section is not rushed like the first, which is contrasting Viñes’s performance. The rubato in bars 33-36 however, is remarkably similar to Viñes’s approach. In bars 38-60, Debussy has a similar way of balancing the rushing/slowing as well as pacing the general tempo compared to Viñes. He does also slightly “dislocate” the melody from the chords. Debussy does not however “arpeggiate” the chords on the first beat in bars 38-43, nor does he omit the thirds (chord notes) on the second beat of bars 45-47, or add thirds to the second last semiquaver octaves in bars 48-49. I find it very interesting that Debussy seems to play bars 67-68 by “dislocating” the chords from the bass line and then continuing to play bars 69-73 with the melody octaves “dislocated” from the middle voice. This is exactly how I have marked it in my annotated score, according to Viñes’s performance. Debussy does also play with an extensive use of rubato, balancing slowing and rushing in this section, much like Viñes. In bars 78-89, Debussy also uses “dislocation” to emphasize the polyphony. The third Tempo giusto-section in bars 92-95 is rushed more in the manner we heard the first time. In 96-97, Debussy adds arpeggios to the first beat, just like Viñes and Gaillard. Unlike Viñes, he plays both low C sharps in the bass. In bars 98-106, Debussy has “polyphonic rubato” in a similar manner as Viñes in his playing, although he generally plays this section a little slower. What is mostly different between the two recordings is that Debussy “arpeggiates” the habanera rhythm chords in the high register. The tendency seems to be that he clearly “arpeggiates” the chords on the first beat and then varies the practice with more subtlety for the rest of the bar. Bars 109-121 are remarkably similar to the Viñes recording. It seems like the crescendo markings in bars 110 and 116 are interpreted with excessive rushing in the performances of both Viñes, Gaillard and Debussy. The polyphonic playing through “dislocating” the octaves in bars 113-114 and 119-121 is very much in line with my markings in the annotated score. Regardless of the original main tempo, it seems that Debussy plays both the beginning and the ending of the piece coherently much slower than Viñes in relation to the other tempos in his performance. In the last bars of the piece, we hear a very subtle use of “dislocation” in varying degree. It is difficult for me to say whether they vary because of Debussy’s touch or if they vary as a result of unprecise reproduction from the Welte system. The most important aspect is nonetheless that the practice is evident. 

As a final note in this comparison, it is interesting to note that in the 2006 Durand edition of the piece, Roy Howat indicated four cases where the Welte recording of Debussy deviated from the score.[9] These were notated in auxiliary staves labelled “R”. In all four cases of this, Viñes does not play as Debussy, rather he plays what is written in the score.

Final thoughts  

My recreated performances, the accompanying reflection texts, as well as the annotated scores all serve to suggest that there is indeed a fundamental presence of romantic performance practice in the Viñes’s recordings of the four Debussy pieces. In my personal opinion though, rather than regarding these performances as “colored” by the romantic performance traditions of Viñes’s playing, the practices should instead be seen as artistic tools, applied by Viñes in his playing to bring out the essence of the compositions.

Previous                                                                                                                                                                              Next