Joel Grant Taylor's Post on Spectral Techniques for Polytempic Polymicrotonal Music, Part XVI

 A Few Thoughts On Spectral Techniques for Polytempic, Polymicrotonal Music

For the sake of discussion, let's imagine that we have a perfect ability to realize or perform any sounds that we can imagine.

We can create any sound and in any tuning or temperament, laid out in time in any fashion that we desire. We can shape the timbre and texture of our music in such a way as to cause people to hear voices where there are none, to transform the sound of rain into a percussion trio, into a crowd of people shouting.

This can be done today with computers using a plethora of techniques. It can also be done with acoustic instruments by using “instrumental synthesis” techniques, with the help of computer-aided audio analysis and composition tools.

In instrumental synthesis, the composer uses the computer to aid the composing process. She does a spectral analysis of a recording of a stretch of sound, and asks the computer to produce a score to simulate that, using a set of acoustic instruments that the composer selects.

The computer does this with linear approximation or genetic programming techniques. It has access to a huge database of spectrally codified instrumental sounds for all of the orchestra instruments and many ethnic, Baroque, and Renaissance instruments besides. This database includes all the normal articulations plus many special instrumental techniques. The program used to do this is called Orchid; it and the database that it uses are the result of years of research at IRCAM in Paris, France.

For example, if you were the composer, you might ask Orchid to analyze a recording of a riot in a busy city and to produce a score to simulate that with wind and percussion instruments. It not only produces a score, but it also produces an 1^st draft audio realization of that score, so that you can judge the score by ear. Once you have the score, you can change it and extend it. So you might impose a tuning on the pitch content, or several tunings in different octaves or different instruments. Or gradually cause a texture with random rhythmic qualities to take on a meter and a tempo.

One of Orchid's limitations is that it only produces scores in standard 12 tET or 24 tET. Perhaps a later version will be completely retunable. The other limitation is that it is slow and expensive, computationally. You need a fast computer and some time and patience to use it. Even as it is, it is an amazingly powerful tool. If I want to take a score produced by Orchid and rework it in Just Intonation or 17 tET, nothing is stopping me.

In computer music, and when using contemporary modular synthesizers, the composer has a huge choice of tools with which to work. There are many synthesis methods and processes to choose from. Using additive synthesis, you can create sounds that have a pitch spectra that mirrors the microtonal scales that you are using. You can use physical modeling and/or sampling-based instruments to simulate acoustic sounds. Once you have generated some sounds to work with, you can impose spectral processes to change the timbres using pitch shifting, time stretching, filters, EQ, ring modulation, harmonic exciters, dynamic multi-band compressors, etc. You can add reverberation and locate the sound anywhere in the stereo field. If you have access to a multi-speaker system, you can virtually place a sound anywhere in the listening space.

Hopefully, the above gives us a glimpse of what the potentials are in contemporary music production tools. The chief impediments at this point in history are monetary. Gear is expensive, time is expensive, and the fact that many music tools, whether computer music languages, modular synthesizers, or software plugins, have steep learning curves makes time commitment necessary. Some math may be involved in the use of some tools. But the contemporary composer is not afraid of math; she goes where angels fear to tread.

One of the biggest draws of doing music that is polytempic and polymicrotonal is that it enables you to create music of unparalleled complexity that is also comprehensible, given a listener with an athletic ear.

The ear more easily separates and attends to several streams of music at once if the streams are separated in pitch height, if the streams have tempos (and rhythms) that are not similar, and if the streams have different tunings.

Additional separation of musical streams can be effected by also using different timbres for the different streams, and a decisive separation of musical streams can often be affected by placing the source of one stream in a different location from the other streams.

By contrast, using unified, or similar, timbres across a number of musical streams, as in a string quartet, will promote fusion and bring the vertical/harmonic element more clearly into play. This will be especially true if the pitch height level of the different streams overlaps or aligns smoothly with each other.

Conjecture Suite, composed 2011 – 2012 https://joeltaylor.bandcamp.com/album/conjecture-suite

This is my first extended polymicrotonal, polytempic piece; about 25 minutes long, it has five parts, and follows a plan of increasing microtonal density over the duration of the work. The intervallic content used becomes smaller as the work progresses. The work is programmatic even though this is concert music. It raises up for examination the situation humanity finds itself in now, and questions whether or not we have the ability to survive our own doings – Global warming, endless war, loss of species, etc. The piece was realized by multi-tracking “keyboard” improvisations done to click tracks. I used different tempos for different tracks, and the click tracks that I listened to while realizing the graphic score I used to guide my improvising are derived from a tri-nome I designed.

The graphic scores I used to guide my improvisations consist of pitch-time rectangles, with time being represented on the horizontal, and pitch being indicated by vertical placement of squiggly lines, blocks for clusters, stars for dissonant chords, some written instructions, etc.

I am purposely bringing the old device of using click tracks to coordinate improvisations to the foreground here because it is a very simple way of producing polytempic music. No sequencers or computer music languages need to be involved.

The first section of Conjecture Suite uses only one tuning, Wendy Carlos' Alpha Scale, which has a smallest stepsize of 77.965 cents, or 15.3915 steps per octave. The 2^nd section uses the Carlos Alpha and Beta Scales. The Beta Scale has the smallest stepsize 63.833 cents. The 3^rd section uses the Carlos Alpha and Beta Scale, and Ben Johnston's Septimal Just Intonation Blues Scale. The Johnston scale is a 12-tone scale, with a 7/6 for a minor 3^rd, a 5/4 for a major 3^rd, and a harmonic seventh, 7/4, for the minor seventh tone. I view it as a “grounding” scale by comparison to the Carlos scales, which are kind of “floaty” because they don't repeat at the octave. It also provides some JI “buzz.” The 4^th section uses all three of these scales plus the Carlos Gamma Scale. The Gamma Scale has a smallest step size of 35.1 cents, and there are approximately 34.19 steps per octave. The 5^th section, which is a kind of extended meditation, again uses all four scales simultaneously, each in a different track and tempo.

Four Tunings/Scales:

• https://en.wikipedia.org/wiki/Alpha_scale

• https://en.wikipedia.org/wiki/Beta_scale

• https://en.wikipedia.org/wiki/Gamma_scale

• and Ben Johnston's JI Blues Scale

Instrumentation: 4 tracks in the Reaper DAW. Any DAW will do.

Instrument/FX Chain for each voice:

Pianoteq > PitchSift > SpiralStretch > Compression&EQ>OUT

Pianoteq Instruments: miscellaneous physical modeling piano models and harpsichord models. Different models have different spectra and give different results when granular analysis/resynthesis is performed.

What comes out of this effects chain does not sound like piano sounds, instead one hears clouds of strings, voices, brass instruments...

Here are the controls for Tom Erbe's SpiralStretch plugin:

This plugin produces multiple streams of sound in response to a single stream of input. The output streams are stretched out in time, so the sharp attacks of the piano are stretched out to the point where the amplitude envelopes are gentler, hence the strings, brass, and voice sounds. The grainsize controls the size of the sound grains that are stretched. The rate controls the size of the FFT window used to do the stretching, and this produces pulsations, which sometimes sound like attacks, and sometimes like echoes.

Here are the controls for the Pitchsift plugin:

This plugin takes its input and transposes it without affecting the speed. You can set the number of harmonics in the output with the “harmonics” dial, the “pitchshift” and “cents” dials let you choose an interval of transposition using either just intonation or cents. The “octave” control lets you move the output frequencies up or down by octaves. The “partialgate” control lets you tell the plugin to ignore frequency content that is below a given level in dB. The number of bands shows how many frequency bands are involved in the analysis. More bands, more frequency resolution, smoother sound. Sometimes using a few bands gives better, more interesting, and grittier output.

You can find these two very useful plugins at https://www.soundhack.com. $25/ea. There is some freeware on this site as well.

Ok. The point of this overview of the plugin operation is to make sure that you understand that when I am realizing a track in Conjecture Suite using this setup, what comes out of the speakers when I am laying down a track is very different from playing a physical modeling-based piano. Experimentation with the settings on the two plugins is essential to success. The graphic scores for each track ended up having notes in them about changing settings during performance. Settings have an effect on the perception of tempo, timbre, pitch height, and more.

In the end, we have a polyrhythmic, polymicrotonal piece that has considerable textural and timbral/spectral complexity, and I hope, a lot of emotional, intellectual nuance to it. Yet the process of realizing the work was relatively simple and straightforward, and supported an intuitive, improvisational approach to the composing process.

So, dear reader, those of you who have taken the trouble to also listen, what do you think of the idea of using Orchid to produce an entirely instrumental version of Conjecture Suite? Is this even possible?

I hope you've found this blog entry interesting, perhaps even useful, and stimulating. If Dr. Thoegersen is willing and there is interest, I may follow up this entry with a discussion of some more of my music, and speculation and thoughts on polyrhythmic, polymicrotonal, polytempic music in general. I'm also open to discussing electronic and computer music techniques and compositional approaches in future blog entries.

JGT/8-12-25, Staunton