DIY Canons

Pogus 21036

DIY Canons


Curated by Simon Wickham-Smith

bo bell, philip corner/mike winter, ross craig, kyoko kobayashi

drew krause, masaki kube, giuliano lombardo, steven m. miller

bruno roviano, miks swinchoski, stefan tomic

simon wickham-smith, mike winter, george zelenz


Entrei Pelo Canon 1a uses four types of screams: A) glissando up, B) glissando down, C) glissando up then down, and D) free form. Four male “screamers” performed different versions of each scream. The order of the recorded screams follows the permutation rules of FVC#13. The length of each part depended upon the variety of screams; during the recording session, some screamers were more prolific than others. The title is a reference to the Brazilian idiom entrar pelo cano (lit. "to enter through the pipe"), which means to find yourself in trouble, or refers to someone who has been deceived. — Bruno Ruviaro


Canon No 1 features manipulations of the several musical dimensions: notes and note octaves, instrumental arrangement, dynamics, note velocity, and vibration rates. A simple pattern of the pitches C, B, A and D make up the permutation list, which is repeated in all four voices. The list is realized, beginning as a different starting point for each voice: first to last, last to first, element 19 through element 18, and element 18 going backwards to element 19. Lists in other dimensions use variations on this pattern. This includes associating the notes with various octaves (2, 3, 4, 5). The octave pattern executes an order separate from the pitch list. The notes are assigned the MIDI loudnesses, 4-note velocity, and permutation of vibration rates.

Each of the voices uses 4 different instruments:

Voice 1: jazz guitar, contrabass, vibraphone, French horn

Voice 2: acoustic guitar, cello, tubular bell, English horn

Voice 3: clean guitar, violin, piano, clarinet

Voice 4: guitar harmonics, harp, electric piano, flute

The variations are according to the pattern listed above. All the voices come from General MIDI. — Mike Swinchoski


Transmissions uses a set of four General MIDI instrumental voices, each playing their permutation lists in different tempi. A clavichord begins the piece playing at 120 beats per minute (bpm). An acoustic guitar begins playing 5 seconds into the piece and starts at 95 bpm, followed 38 seconds later by an electric guitar playing at 87 bpm. A harpsichord is introduced 71 seconds into the piece and begins with a tempo of 78 bpm. The 2nd, 3rd and 4th voices accelerate at various rates until they join the first clavichord playing at 120 bpm. In order for all voices to complete the permutation list, they were scaled to fit the time frames that were selected, after which the tempo accelerations were applied. — Mike Swinchoski


Freeze Dried Canon reconsiders the function of time in a DIY canon. I wanted to subtract the intervals of time between notes (vacuum them out) and section them into separate modules so that no space intervals would be left between the notes. This can be visualized as a flat or two-dimensional structure. Each element in each voice runs through permutation lists and occurs simultaneously, but in its own time frame. Each element starts at a different point in the list and runs through the cycle in a different sequence from the others. The 1st voice is four seconds long, the 2nd three seconds, the 3rd two seconds and the 4th, one second. Elements of the permutation list are assigned to different generic synth patches or electric guitar. — Mike Swinchoski



Filter IV – P.I.X.L. Study no. I implements spectral filters for each voice of the canon. The information for the filters is read from a picture. The image seen on the cover of this CD is somewhat deceiving, the one read by the computer program is actually 633,880 pixels on the x-axis (time), and 233 pixels on the y-axis (frequency) — a print of the picture at 600 pixels per inch would be just over 88 feet long and less than a half an inch high. At full resolution, the colors come in and out of blackness with a quasi-Gaussian function to produce the gradients. The red, green, or blue values from 0 to 255 are scaled from 0–1 and determine the amount of attenuation for each passband, represented by the rows. Another picture determines the filters’ output in space.

Each canonic voice consists of a set of instruments passed through a real-time spectral filter that reads one color of the picture: the cellos’ (voices 1 and 4) filter reads the red values; the bowed cymbals’ (voice 2) the green values; vocalists’ (voice 3), the blue values. Only the filtered sounds are heard.

Function generators were used to determine the picture's parametric changes over time. As spectral granulation is implemented, pitch areas, granular durations, and granular densities constantly change and evolve. The rate of change of these parameters is faster for each incoming voice of the canon.

As the picture indicates, the pitch areas descend and ascend before the final descent that brings all voices together at the end. The form eventually reveals all parts of the instruments' spectra. Sometimes parts of the instruments' spectra sound separately, sometimes together, the latter occurring at the end when the live signal of the instruments is mixed in with the filtered signals. There are correlations between pitch areas, granular density, and duration. The lower and wider the pitch area, the longer the granular durations and the greater the granular densities.

Over time, there is also a general trend from inharmonic to harmonic exposures. For example, since the canon begins with the upper part of a cello’s spectrum, the revealed spectrum consists of close and tightly packed harmonics with small intervals, revealing unusual spectral details of the instrument. As the pitch areas descend towards the fundamental, the harmonics are more widely spaced. An unfiltered tam-tam is played softly as an ambient background for the work.

Thanks to Larry Polansky for encouraging me to write this piece. If it were not for his generous and continuous support as well as his prolific contribution to music, I would be a lost soul. — Mike Winter


Barbie’s Phone Canon uses a Barbie Phone I bought at a Goodwill store to add to my collection of electronic children’s' toys that synthesize music, sound effects or voice. One of the features of the phone is when the "message" button is pushed the phone randomly generates messages from four or five lists: a greeting (Wow, Great…), a suggestion (You should, We could…), an activity (get ice cream, practice singing…), sometimes a companion (with Ken, with Midge…) and a time (on Friday, after school…). The same week I got the phone, I heard Larry Polansky talk with Sarah Cahill at the Composer Portraits series in Berkeley. He discussed the Do-It-Yourself Canons and played several examples. I thought this was a great idea and set about making a canon with the Barbie Phone. I recorded 27 random messages and assembled them in groups of three so that all the groups were the same length. I strung together nine of these groups with a pause between each group. Each of the four voices of the canon starts two groups of three messages after the one before. The first voice has nine groups totaling 21 messages, etc. All start with the same group and proceed in the same order but end on a different group. I added some dial tone, a couple of rings, a nice voice notifying us of the number of messages, and there I got my I-Did-It-Myself Four Voice Canon. — Ross Craig


Kullankeltaiset Päivät (Larry-lle) (Golden Days for Larry) was composed in late 2003 and is a mensuration canon featuring samples of traditional Finnish music and the sounds of a wooden flute played by a lover of all things Finnish, the extraordinary Khan Gorlewski. — Simon Wickham-Smith


Twin Canon is a simultaneous four-voice mensuration canon and crab canon. The crab canon voices are two octaves below the monophonic main voices, with doubling of the crab canon voices one octave down, resulting in up to 12 voices simultaneously. Each of the four main voices ascends one octave of a 19-tone just scale from the tonic to the octave, totaling 20 notes per voice. The crab canon voices are the main voices in simultaneous retrograde. All pitches are calculated from a set of harmonic proportion ratios applied to a base frequency of 440 Hz. Total time, start time, and note durations for each voice are calculated from the same ratios applied to a base time of 20 seconds. FM synthesis parameters are also calculated from these same ratios. The C:M ratio is identical with the pitch ratio for all notes in all voices. The modulation index is calculated from the square of selected ratios on a per-voice basis. The panning rate for spatial distribution is derived from the same selected ratios per-voice.

Twin Canon is an exploration of harmonically proportioned pitch, timbre, and time; it grows progressively thicker in overall vertical and horizontal density. Harmonically, the piece is in an asymmetrical arch form, with increasingly dense pitch relationships finally converging into a resolution chord of tonic, major second, perfect fifth and minor seventh reinforced in 3 octaves. The timbres grow progressively brighter due to the higher index of modulation of successive voices. As the harmonics generated by the FM synthesis change as a function of the C:M ratio changing in direct relationship to the scale degree of the note, timbral development is related to the overall harmonic motion as well as to the increasing temporal density.

I am indebted to the work of James Tenney and Larry Polansky, separately as well as together, in formal perception and experimental tuning. — Steven M. Miller


Pulse Canon uses a subset of the same set of 20 superparticular ratios (and their inversions) as Twin Canon, though in an entirely different way. The pitch of the four voices is calculated 1/1, 9/8, 4/3 and 7/4 applied to 220 HZ to determine the pitch for each of the four voices. The same subset is used to apply maximum limits to a random panning algorithm, with successive voices having a greater degree of latitude in the stereo field. Each voice is a simple pulse.

A simple FM drum or wood block articulates a logarithmically decreasing time interval between successive notes. Each voice begins with its total duration divided into 10 pulses; by the end the total duration is divided into 10,000 pulses. This happens over the course of 200 seconds for voice 1, 177.7778 for voice 2, 150 for voice 3, and 114.2857 for voice 4. All voices end together.

Starting from a rather sparse texture of one pulse in 20 seconds, by the end of the piece the four voices articulate 50, 56.25, 66.67, 87.5 pulses per second, totaling approximately 260 pulses per second in the final moment. The overall effect is of a gradually thickening texture that takes on a gradually rising pitch. The piece is indebted to the work of James Tenney and Conlon Nancarrow, as well as Larry Polansky. — Steven M. Miller


Canon for LP starts with a self-similar melody in a mensuration canon at the rates of 4, 7 and 9. After some algorithmic dabbling, the fourth and fifth voices join in at the rates of 16 and 25. Completed and recorded in August 2004, the piece is dedicated to Larry Polansky. — Drew Krause


Thermohaline was composed on November 22nd, 2004. It is dedicated to my mother, Joyce Zelenz, who passed away the day before. John Whooley, plays bass clarinet; the tempo ratios are 16:17:19.

In the world’s oceans, cooling surface waters with increasing salinity sink down along the continental shelves towards the lowest depths of the oceanic body. Once fully descended they move as very slow currents under tremendous pressure. This process is known as thermohaline. —George Zelenz


Ringtone Canon is a remix of a piece I composed in 2003, originally titled Cellphone Canon. The piece follows the outline of Larry Polansky's DIY Canon #13, incorporating a few key concepts of the DIY canons, including the use of permutations for the ordering of compositional elements and ratios to determine the tempi of voices. Four samples are played in succession in the four voices of the piece. Resampling is used to play the samples at variable speeds, resulting in a transposition in pitch and a change in tempo. The ratio 15:12:8:6 is used to determine the rate at which the voices are initially resampled. The first voice begins without resampling, equivalent to a 1/1 resampling rate. The second voice is resampled at 8/6 (or 4/3) of that rate. The third and fourth voices are resampled at a rate of 12/6 (or 2/1) and 15/6 respectively. The resampling rate of each voice then increases towards the middle of the piece, and decreases back to the original rate. Artifacts of the resampling process can also be heard, especially in the middle of the piece. Each successive voice has a shorter overall duration due to its higher resampling rate. Instead of having the voices end at the same time, as demonstrated by the DIY canons, I chose to align them in the middle of the piece. This produces a symmetry in the piece as the voices return to their original tempi and pitches. — Stefan Tomic


Metrocard Canon (2002) is a four voice canon for performers using Metrocards (pass cards for the New York City subway system). The piece is made up of a sequential combination of five possible sounds: a swipe of the card through the turnstile (which produces a "beep"), a double-swipe (making a double "beep"), a "ping" produced by striking the card's edge upon the turnstile bar, a "clicking" produced by turning the bar one rotation in the proper direction and a "bang" produced by turning the bar in the wrong direction.

Performed in February 2002 by Bo Bell, Laura Roberts and Sai Sriskandarajah in the Clark Street subway station, Brooklyn, NY. Mobile recording by Bo Bell and Laura Roberts. — Bo Bell


Philip Corner/Mike Winter MENS 5/4

Philip Corner/Mike Winter MENS 9/8

When I first saw gamelan Mens I was excited to see that Philip Corner had written as a text score a very calculated, almost mathematical work that touched upon many of my own interests in music. In Corner’s canon each voice constructs an ascending melody that accelerates “by the smallest practicable degree.” Successive voices enter one scale-degree of the melody higher at a time, “with, a bit before, or after” the second tone of the first voice, and ascends in exact transposition. Each voice of the canon accelerates so all voices reach the peak of the melody at the same time. This process is then reversed, which concludes the first part of the piece. The second part flips the canon’s pitch and tempo around. The melody is inverted and voices decelerate to reach the last tone of the melody simultaneously before the retrograde.

I wanted the form to be the most important aspect of the realization without making it the most perceptible. I used an equal-tempered scale, each voice having the same timbre, to shift the perceptual focus to a slowly spreading chord. The chord departs from unison only when a voice reaches a higher tone before the other voices. As the voices change in speed and surpass each other, what Philip calls a gradually changing “harmonic blur” begins to evolve. Finally, we hear the four-tone, equally spaced chord.

The melody consists of 30 equal steps in a just major third on track 7 (a frequency ratio of 5/4 from the starting tone) and a just major second on track 8 (a frequency of 9/8 from the starting tone). I used a bell-tone with an imposed Gaussian envelope so each tone enters and exits imperceptibly. Since the spacing is small and all tones are created from the same sample, different beating patterns emerge to suggest the chord is spreading. Eventually all voices line up and the chord is revealed. — Mike Winter

The pieces on this CD are all based on the ideas in Larry Polansky’s four voice canons, a series of pieces he began in 1975. These canons are usually “mensuration canons,” which means that the tempi of successive voices is proportional to their start times, so that the voices end together. They also use simple ideas of moving through a list of permutations, and applying the elements of those permutations to various musical parameters. A set of Polansky’s canons was produced on Cold Blue Records as four voice canons (CB0011).

Polansky described these ideas in Four Voice Canon #13 ("DIY Canon”), a kind of open-source meta-canon score. He distributed this at talks and on the web as an open invitation to others who wanted to make four voice canons of their own. Composers responded with a wide variety of variations on the basic idea; a selection of these works is presented here. — Simon Wickham-Smith, curator