next up previous
Next: TrikTraks Up: Wigout Previous: Waveform Specification

Multiple cycles

The ability of specifying duration changes independently from amplitude changes allows for these two variables to have distinct cycle lengths. Since a state can have up to 64 segments, up to 128 different cycle lengths can be explored, with the segment type wiggle.

As is written above, the cycle length of a variable and its increment of change are inversely related. So if we know what the cycle length is to be, and we know the range of the variable, its increment of change can be determined by inc=2*(max-min)/numCycles.

Here is the input file for 4 segments, whose variables all have cycle lengths that are relatively prime. The ranges for each variable are given in parentheses following the hash mark. The ranges for the samples are written in standard pitch/octave notation, and assume a sampling rate of 44100 samples per second. The ranges for the amplitudes are in decibels, and assume 16-bit sound samples. The last number on each line is the variable's cycle length.

# 2nd input datafile for wigout
w2 wiggle
15 45 15 0.134529              # (b5 f#6) 223
-5107 5107 -5107 65.057325     # (80 dB) 157
w3 wiggle
42 42 16 0.114537              # (c5 f6) 227
2016 2016 -2016 24.736196      # (72 dB) 163
w4 wiggle
17 40 17 0.100437              # (c#5 e6) 229
-796 796 -796 9.532934         # (64 dB) 167
w5 wiggle
38 38 18 0.085837              # (d5 eb6) 233
20588 20588 -20588 238.011561  # (92 dB) 173

And this is a plot of the changing base frequency for the waveform, using only the segments defined above, over a duration of 60 seconds.

\scalebox {0.6}{\includegraphics*{graphics/freqChanges23314.ps}}

Note that gradually, over 60 seconds, the base frequency's range changes significantly. This occurs gradually because the duration's increment for each segment is very small, meaning that upon every iteration of the state, the number of samples for each segment will barely change.

The timbre is changing faster than the base frequency, as is shown by the shorter cycle lengths for the amplitudes (157, 163, 167 and 173). Unfortunately, the change in timbre is more noticeable when heard than when seen as a sequence of FFTs, so no plots of the changing timbres are included in this report.


next up previous
Next: TrikTraks Up: Wigout Previous: Waveform Specification
Arun Chandra
arunc@evergreen.edu