This Bohlen-Pierce arrangement of Terry Riley’s “In C” was realized on the musical hardware described in Rehearsal Room. It was performed by 13 identical algorithmic musicians known collectively as The Technical Academy. A characteristic feature of the ensemble is an indiscriminate fader controlled by a Markov chain on each voice. This is the third recording of the completed algorithm.
The notes of Terry Riley’s original score were remapped to Bohlen-Pierce notes in a slightly modified A Moll I mode; the B note would have mapped to BP’s J, but it was changed to Jb simply because Jb sounded better. Riley’s C D E F F# G A Bb B becomes A B C D E F G H Jb. The work is titled “Not In C” because it’s based on this different scale.
Developing Ruby code to drive antique synths with MIDI pitch bend messages. This is a Roland D-110 coupled to a Roland U-110. Each MIDI channel can only play one note because MIDI pitch bend affects the whole channel. The audio track is monotonous, but useful for ringing out the instruments.
This table demonstrates some of the logic behind the Lambda scale. It shows the 17-limit just intonation (JI) intervals that are closest to the equally tempered Bohlen Pierce scale. Notice that there are no close JI intervals for the F#, H#, and A# notes. Also, melody fragments in the D E F G range are virtually equal tempered with respect to one another.
The original JI scale has accidentals that are beyond the 17 limit: C# is 25/21, F# is 75/49, H# is 49/25, and A# is 63/25. Using those ratios would bring the accidentals within 10 cents of the ET notes, but the ear gravitates towards ratios with lower denominators. In any case, it’s good to avoid H# because of its proximity to the octave. 49/25 can’t compete with 2/1. Same with A# against a strong C – most listeners will hear it as a too-sharp 10th.
Bohlen Pierce ET to JI intervals
best match, 17 limit
Using a sampled A=440Hz piano note, I squished and stretched it to play a BP chromatic scale. Here it is juxtaposed against random MIDI BP notes using the Sonic Pi’s FM synth.
# Bohlen-Pierce notes
tritave = 19.02 # a 3/2 tritave is 1902 cents, or 19.02 in MIDI
half = tritave / 13 # size of the BP half step in MIDI
whole = half * 2
a = 69 # MIDI 69 is A440
b = a + whole
c = b + half
d = c + whole
e = d + half
f = e + half
g = f + whole
h = g + half
j = h + whole
a_ = a + tritave
bpScale = [a,b,c,d,e,f,g,h,j,a_]
exp = 3 ** (1.0/13)
# play random notes with FM synth
divs = [1.0/exp,exp,1] # divisors add random BP flavor for FM synth
for i in 0..48
highN = bpScale.choose
lowN = highN - tritave
play highN, divisor: divs.choose, depth: 0.5, pan: 49/highN, amp: rrand(0.1,0.2), attack: rrand(0,0.01), sustain: 0.2, release: rrand(0.1,0.5)
play lowN, divisor: divs.choose, depth: 0.7, pan: -34/lowN, amp: rrand(0.1,0.2), attack: rrand(0,0.01), sustain: 0.2, release: rrand(0.1,0.5)
# play chromatic BP scale with sampled piano
speed = 5.0/9 # initial sample speed yields BP E note (JI) calculated from A=440Hz
for i in 1..36
sample "~/Sonic Pi/piano-a440.wav", rate: speed, pan: speed - 1.5, amp: rrand(1.0/3,2.0/3)
speed = i < 18 ? speed * exp : speed / exp
Way back in 1991, I produced an album of algorithmic music. It has recently been re-released by Fixture Records. Renewed interest in the project sent me on a quest for a comparable modern programming environment, one that would also support BP compositions.
I’m pleased to report that I’ve found such a platform in a program called Sonic Pi. Originally written for the Raspberry Pi miniature computer, the program is now also available for Mac OS X and Windows. My first quick attempt at producing BP tones was incredibly easy. I’m very excited at the possibilities here.
Here is a screen shot of the program, followed by an MP3 of the sounds it produced.
This is a 12 bar blues I wrote using the Bohlen Pierce scale C D E F G H Jb A B C. The lead part is pedal steel and the other parts (except drums) were played on the Roland Handsonic. Recently remixed with real drums by my good friend Oscar Calderon.