[ top image ]

Back to Compositions

Fether lyre

This composition was synthesized with physical modelling of mass vibrations (difference equation resonator array). The sound-producing objects were modelled as consisting of parts that each have a spring constant and a mass. However, to make the model more realistic, the spring constants are mutually modulated by the neighbour part vibrations (tension change under spring length variations when the parts are pushing or pulling each other).

Listen Fether lyre (2M - 2:06).

For curiosity, below sound material of the piece (before they were mixed together).

Please, note that some of the material go into chaotic noise. [However parts of the material were used in the composition because of their unique fascination!]

Listen to the sound material of the piece:
test31g.mp3 (300k - 0:19) Tape 1 (bell-like sound, goes chaotic)
test30y.mp3 (40k - 0:02) Tape 2 (low struck-wood-like sound)
test33a.mp3 (40k - 0:02) Tape 3 (muffled bass-pizzicato-like sound)
test30t.mp3 (40k - 0:02) Tape 4 (high struck-wood-like sound)
test30f.mp3 (200k - 0:11) Tape 5 (another bell-like sound, goes chaotic)
lyre_test2f.mp3 (500k - 0:31) Tape 6 (lower Lute sound)
lyre_test2e.mp3 (150k - 0:10) Tape 7 (lower Lute sound)
lyre_test2g.mp3 (300k - 0:19) Tape 8 (lower Lute sound)
lyre_test2d.mp3 (300k - 0:19) Tape 9 (lower Lute sound)
lyre_test2b.mp3 (300k - 0:19) Tape 10 (Lute sound, fluctuating noise)
lyre_test2a.mp3 (400k - 0:25) Tape 11 (Lute, some fluctuating noise)
lyre_test1v.mp3 (500k - 0:34) Tape 12 (lower Lute sound)
Lyre_Electric07a.mp3 (600k - 0:40) special lute sounds
Lyre_Electric08a.mp3 (600k - 0:40) special lute sounds
Lyre_Electric09a.mp3 (600k - 0:40) special lute sounds
metalRvb3b.mp3 (900k - 1:00) background
metalRvb3e.mp3 (900k - 1:00) background
metalRvb4l.mp3 (900k - 1:00) background
Lyre104a.mp3 (2M - 2:19) dry foreground channels
Lyre105c_REVERBs.mp3 (2M - 2:19) wet foreground channels
Lyre99c.mp3 (2M - 2:14) lute channels

Listening Score / Performance Score

Here a listening score (sheet music) in PDF format FetherLyre_PDF (1.5M - Adobe Acrobat Document); and also in Coda Notation File format FetherLyre_MUS.zip (60k - Finale2004 score zipped) for your convenience (those of you having Finale 2004 or higher, I mean).

The original purpose of the score, for the composer, was to make a draft for an electro-acoustic composition - even though the sound would be generated by a separate computer music synthesis system (see below).

One day, perhaps, somebody might even want to use the score as a performance score for human player(s), (with or without optional live electronics), in a consert?

Music Synthesis Environment

The Music Synthesis System for this piece was a C++ program (MAL-d) written by the composer.

The acronym MAL-d stands for Modular Audio Laboratory (development environment), and the system was primarily built for evaluating the feasibility of use of difference equation resonators for audio signal generation. (For basics about digital signal processing, see e.g. [1].) The advantage of a difference equation resonator is small memory consumption, but with the disadvantages of need for a separate resonator element for each partial (harmonic or non-harmonic) plus high tendency toward instability (see e.g. [2]). The difference equation approach is one way of doing so-called physical modelling. The additional benefit of physical modelling in general is that it is easier to produce sounds that are more lively resembling real sounds (compared to the "dead-and-dull" sound of simple wave table synthesis). So, one mandatory requirement for the system was ability to apply frequency modulation (FM) to individual resonator elements, for simulating variations in the resulting spectra.

(Note: Another branch of physical modelling, Digital Waveguides ([3]), might achieve quite similar results when applying variable filters in the feedback path(s).)

At a late phase of the project, also an additional feature was added for ability to provide input events in text file (much like the score file format in the Csound language [4]), though that was not an original intention (nor a requirement either). As an example of the MAL-d event file format, here is a file lyre99_eve from the Fether lyre composition project.

Here is also one article draft LyrePaperDraft_PDF, written by me (unpublished - you are allowed to refer to the general ideas presented in the draft but direct quotes are prohibited), concerning the composition project and the used environment.


[1] Spjut, E. : An Introduction to Signal Processing with Csound; in The Csound Book, edited by Boulanger, R. - The MIT Press, 2000.
[2] Gordon, J.W. : System Architectures for Computer Music; article in Computing Surveys, Vol 17, No. 2, June 1985.
[3] Mikelson, H. : Mathematical Modeling with Csound: From Waveguides to Chaos; in The Csound Book, edited by Boulanger, R. - The MIT Press, 2000.
[4] Boulanger, R., [et al.]: The Csound Book. - The MIT Press, 2000.

Copyright © 2005 kari väkevä
Updated: 04.06.2005