Exercise in Styles:
On Developing Multiple Versions of Debussy’s Syrinx

The departure point for the next four versions is the manipulation of time. The aspect of temporality is especially significant to performance, as all the performance parameters are shaped in relation to the axis of time. In other words, time is a key element not only in relation to the tempo and time inflections, but also to the shaping of the different components of the musical gesture, such as vibrato, tone colour, and articulation mentioned earlier. Exploring temporality in the following versions, I distinguish between tempo and timing: Tempo as relating to the duration of basic regular beat, assuming that there is one, while timing refers to the moment-to-moment fluctuations within the beat.


Most scores include a verbal indication for tempo that are sometimes also accompanied by a metronome mark, which is one of the most precise indications provided by the score.

However, as widely documented in recordings, the extent of which performers relate to a specific tempo indication depends on performance practice, and in many cases, tempo marks are often treated in a personal way and as an indication for a tempo zone, rather than to be followed to the letter. In the case of Syrinx, the indication given is Tres modéré (very moderated). Although commonly used, it is not actually clear what is being moderated. Is it the tempo? If so, does it means moderating a fast or a slow tempo? Or is it the character or mood that is being moderated? And how slow or fast Tres modéré can be?

Listening to a wide range of recordings reveals that Tres modéré is being read at the faster range of larghetto, with a basic tempo ranging between 58 and 66 per quaver. The focal point in the next version was a tempo that is slower than the one usually taken.

In this version, the short notes, which, in a faster tempo are perceived as passing notes, present an expressive potential, due to their chromatic nature. Furthermore, taking the opportunities presented in the slower tempo in shaping the short note passages sheds light on an alternative formal reading, in which the piece is read in the context of whole tone mode. Briefly, the piece is usually read as written in a ABA form floating between related tonal centres of B-flat minor, and a middle section based around E-flat. The ‘arabesque like’ chromatic passages are considered in this context as a momentary chromatic instability. A formal reading that results form what John Rink calls ‘a performer’s analysis’ (2002), the slower tempo and the resultant emphasis on the short notes affords an alternative reading of the piece as a process of move from a B-flat to B-natural whole-tone scales as a reference.

In the next two versions I explored taking two opposite extremes of a slow and a fast tempo. The fast tempo version actually meant as a joke; paradoxically, as it is so fast, its character seems quite moderated!


There are endless possibilities available to performers in shaping timing following their personal taste, and to a large extent, as a result of the performance trends of their time. Using technology it is possible to demonstrate how Syrinx may sound like if mechanically executed as written. I have here two versions: the first is of a raw midi score, in which the sound is mechanical.


The second example is made using sample sounds of the Vienna Sound Library. Briefly, for those who are not familiar with this technology, the library includes a vast number of very short samples, each of only few seconds long. These short samples are recorded by players of the Vienna Symphony, who played all notes in different dynamics, articulation, tone colour etc., creating a huge library of sample sounds. Unlike the first example, which is quite horrible, this one provides some interesting food for thought.

I’ve recorded several versions focusing on timing with different degrees of freedom, resulting from different emotional states as departure point. In the first version I was aiming for being free while remaining within the range of what may still be considered as a mainstream approach to rubato. The version that follows takes a step further in allowing the narrative to lead the shaping of timing. The last version wasn’t planned in the recording session and was made at the end of the end of the recording as a joke. To my surprise, this version is preferred by many, and especially by those who listened to the recording without having strong views regarding what to expect, and how Syrinx should sound like. Since the recording was made I’ve been experimenting with how far can I go taking this an improvisatory approach, never knowing what would be the outcome.

Timing 2 (…that was such a long time ago)