In this installation the Sonatina in F# minor by Ravel goes through hundreds of transformations, or musical 'morphings'. It becomes a fixed point to be "distorted in intonation and rhythm".1 In most cases the original cannot be recognized. Open Music became important for these transformations.2
The original Ravel fragment is tuned to the 'Franck chorale' tuning used in the microtonal organ.3 A temperament not too distant in time from Ravel can contribute to sonorities of the music.
Some versions go directly to the next processes. After some time I added a musical 'granulation' to the versions. 4 Instead of processing the chronological original, random extracts of Ravel were superposed with individual delays, to create textures which are more dense, inpredictable, and variable in lengths. The 'granulated' quotations often have more than the orginal 5 parts.
The whole range and pitch space is distorted through pitch shifting.5
The chronology of the already disintegrated music is altered through a chaotic time pointer curve.6
The music is scaled by a curve between original and inversion.7 In sketches and patches, transformations by time pointer and inversion curve were named "MQ flux". Modus Quaternion offers four variations of a material, through static changes in pitch and time domains. With curve shapes as controllers, there are no longer just four, but an infinite number of versions.
A tempo curve adjusts the speed.8
The now thoroughly distorted pitches are approximated to a number of possible tuning principles. Multiple versions of a tuning type are joined to form denser and more microtonal approximations. Tunings systems can give the music more modal and resonant colours.
Random scales of javanese slendro intervals, with 5 transpositions by the same slendro intervals. Similar slendro randoms are available in the microtonal organ, but orders of intervals will here be unique for each fragment.9
Random scales of javanese pelog intervals, with 7 transpositions by pelog intervals.10
The 'grains tuning expansion' from the microtonal organ, on 2, 3 or 5 parallel concert pitches.11
A random note from the input is transposed down as a very deep virtual fundamental.12 All other notes are approximated to this overtone series.
A curve of dynamics is applied to the music. The manually entered Ravel score did not have the same dynamic variation as tam-tam analyses.
The music is time scaled again, to adjust the overall speed.
Notes are kept if they are within reasonable distance to available samples. The Ravel manipulations are performed by sampled glockenspiel, vibraphone, marimba13 and gamelan ensemble.14 The total range of these percussion instruments is comparable to the range of a piano. The western percussion instruments are equal tempered, while the gamelan instruments are tuned in slendro or pelog. When sampled pitches are retuned for the fragments, this important intonation difference is no longer relevant, all instruments can take part in a homogenous virtual ensemble. Samples are normalized through a csound score.15 Some times this can cause problems balancing the samples. A low register flute at the same dynamic level as a high register flute will not sound realistic. The gamelan kempuls gongs and slentem16 got very dominant in the first csound synthesis attempts. I solved this by dividing amplitudes within a certain range, and in some cases randomly reduce the amount of notes within a certain register.17 These are the samples:
Played mf with hard sticks (secco, short and long notes).
Played mp with hard sticks.
Played with soft sticks (1 second and 1/2 second notes).
Played with a soft stick with a metal stick is resting on the keys. The result is a metal buzzing sound.
Saron panerus (peking).
Kempul, gong ageng, gong suwakan.
At early stages, Ravel transformations were tried out "reproduced as drops of water",18 or morphing between percussion and flute sounds. None of these worked as well as I had imagined, and I ended up with a pure ensemble of percussion.
Each note has it's own spatial position, with a fade between several different reverbs. To keep a balance, notes were placed along the sides of the virtual room, not in the centre, and not too distant. A few percussionists performing would come from a more narrow locations than a string orchestra. These percussion fragments have the spatial behaviour of rain, and are thus not merely realistic virtual ensembles.
The first Ravel percussion versions are similar to the Gamelan-Ravel fragments, but kept in a higher registers on only the glockenspiel, vibraphone and marimba samples. These sounds were easier to balance than the gamelan samples, and filtering or balancing of registers were not that necessary. There are more elegant technologies for this, which I will look into for future projects.
look at some fragments with 1/8-tone or 1/16-tone approximations.
Intonations for synthesis are more presice. First the original Ravel
quotation, which is a source for all the transformations. The amount
of activity within each part can be recognized even in
transformations far removed from the original: Usually, the
accompanying second part will be the most active.
Example 9: Ravel Sonatina fragment for virtual percussion ensemble, in the 'Franck chorale' temperament.
Example 10: Ravel-percussion 63.
Example 11: Ravel-percussion 192.
Example 12: Gamelan-Ravel 4.
Time pointers can make the music stutter at a particular point of the music, and they introduce rhythmic irregularities I have found interesting.
The actual cent values of the first part reveal more detailed intonations:
(3560 3712 3716 3716 3769 3712 3716 4246 4020 3712 4728 4616 4178 4204 4460 3450 3450 3450 3790 3765 4020 3898 3171 4247 2646 3873 3535 3504 3560 3530 3441 3504 3560 3560 3560 3504 3560 3560 3560 3796 3796 3796 3796 3796 3796 3796 3796 3796 3796 3873 3873 3878 3873 3820 3873 3790 3716 3712 3931 4020)
Example 13: Gamelan-Ravel 5.
Example 14: Gamelan-Ravel 51 (in a random pelog tuning).
1 Ruben Sverre Gjertsen, 2011, Between instrument and everyday sound, p. 3.
4 Ruben Sverre Gjertsen, 2013, The Ruben-OM patch library, p. 61.
5 The Ruben-OM function pitchshift-multiseq.
6 The Ruben-OM function multiseq-pointer.
7 The Ruben-OM function scale2inv-multiseq.
8 The Ruben-OM function time-scaler.
9 Ruben Sverre Gjertsen, 2013, 3 Manual Microtonal Organ, p. 9.
10 Ibid. p. 9.
11 Ibid. p. 15.
12 The Ruben-OM function spectralize-rand-chordnote.
13 Recording sessions with Sjøforsvarets Musikkkorps, Bergen.
16 Deep metallophone.
17 The Ruben-OM function bandpass-percent-multiseq.
18 Ruben Sverre Gjertsen, 2011, Between instrument and everyday sound, p.3.
19 Fast Fourier Transform.
< Contents >