The Dartboard
I studied string bass with Mike Richmond, an extraordinary bassist and teacher, for two years in college. During our first year together I was struggling to develop swinging eighth note “lines.” He would improvise with a swing feel and sound amazing. I would try, and sound robotic at best. The situation seemed unsolvable. Then, one day, I got it. I can clearly remember suddenly being able to string eighth notes together in a more or less jazz style. There was a distinct “a-ha” moment when my brain started to think in jazz phrases and was able to express that thought through my body/bass interface. It was a breakthrough. How did this happen?
Something must have changed in my kinesthetic ability to allow such progress. This change is an example of improvement at the level of expressive microtiming. My poorly swinging eighths suffered from two faults; a lack of overall precision (so that “even eighths” were probably herky-jerky) and an inability to execute the subtle explorations of beat subdivision intrinsic to jazz improvisation (so that “swinging eighths” probably didn’t swing). Both these problems involve very small adjustments in timing: my bad swing feel was poor microtiming and my better swing feel is expressive microtiming.
This series of posts seeks to develop theories and concepts leading to development of an “organic drum machine” using Max/MSP/Jitter. So far I have theorized that players in groove based music often metronomize patterns that are syncopated in relation to their instruments’ primary rhythmic role, and proposed that the organic drum machine program should enact rhythm without strictly adhering to clock time, be capable of speeding up and slowing down as people do, and incorporate hypermetronomic rhythmic information as part of its design. Effective microtiming is key to all these rhythmic events: metronomization in relation to a contrasting pattern requires smooth adjustments to the location of that pattern, grooves that run on human time will shift a bit from clock time with each beat or metric grouping, and musical speeding up and slowing down involves matching a pattern to the surrounding musical texture. One great strength of computer rhythm machines is their ability to accurately execute rhythms at any time scale from micro to macro. How can this strength be brought to bear in computer generation of grooving patterns? A first step is to find an analytical method that describes the real time performative enactment of rhythm in groove based music, as this will provide data to incorporate into new rhythm algorithms. Phenomenon such as the swinging eighth note can be studied through sub-tactus level analysis of pitch onset events (Benadon 2007). Such analyses are illuminating but do not engage with preparatory microtiming, the musical processes leading up to moments of actual musical sounding, and as such represent a post-mortem accounting of real-time music making rather than an investigation into the generative elements of improvised performance.
Effective microtiming includes a such a multitude of deviations from idealized beat subdivisions that truly wonderful and truly lousy microtiming can live in the same temporal locations. Though it is possible for listeners to identify very small discrepancies in division of the tactus an analysis of such divisions does nothing to clarify the connections between a musician’s choice in timing and expressivity. Further, this kind of analysis will result in similar results for both great and mediocre improvisation; all musicians deviate from clock time to some extent in their playing. Correlation of phrase level gestures with analysis of note onset intervals (as in Benadon 2009 and many other papers) sheds some light on an improviser’s process and problem solving strategies but ignores the fact that before an audible note onset occurs the improviser has already made their choices regarding pitch and rhythmic placement.
So the organic drum machine, while it must incorporate extensive and flexible deviation from idealized beat locations into its algorithms, should not strictly rely on a model of effective microtiming provided by analysis of preexisting music. In reality the audio information we hear in live performance or recordings, though it is filled with remarkable rhythmic information at every temporal level, is the result of an inaudible process of musical enactment, one level of which is the preparatory microtiming mentioned above. Let’s call the inaudible processes involved in preparation for musical sounding “wet” and those we hear “dry.” Wet processes are internal, chemical, biomechanical, and personal. Dry processes involve active creation of sound; singing, striking, bowing, blowing. Each audible dry process has a pendant wet process that precedes it. While analysis of recorded artifacts reveals the microtiming of dry events, wet events are left out of the picture. The organic drum machine should base its choices in microtiming on some kind of wet process model. For the organic drum machine, as for the real performer, dry process is about release of sound. It’s the wet process that truly dictates when and how that sound is made.
Picture a dartboard and a person preparing to throw a dart at it. All the preparation for throwing is wet process; selection of target, grip, windup, etc. Once the throwing motion has started dry process begins, and by the time the dart leaves the thrower’s hand there is no way to change the dart’s course or destination; the flight of the dart is all dry process. Each step in this procedure takes some amount of time and is controlled by the thrower, but the time the dart is in flight and the short period of time the dart takes to penetrate the board represent the results of a wet process rather than the process itself. Production of musical sound is similar; though some instruments allow continued input after an onset event, the moment of audible sound onset occurs after the wet process of preparation and release and as such its timing is a result of preparation before the onset event. Is it possible to study preparatory microtiming? Current studies of music perception and performance show that music engages diverse areas of the brain. How do these brain regions function in the milliseconds before a musician triggers a note onset? How do individual musicians execute these rapid thoughts and movements in similar or different ways?
Looking again at our dartboard, while the dart player intends to land a dart in some specific location (such as the bulls-eye, or the outermost 20 wedge), the wet process may produce a different resulting dry process (a total or close miss). Further, each zone of the dartboard has its own subvidisions. The top left corner of the outermost 20 wedge and the exact center are different points, yet both are in the same zone. Then there are game strategies: given the same or similar score different dart players will choose different targets. Unless a player announces their intent observers and fellow participants can’t be 100% sure whether a thrown dart found its intended location. Finally, there are different rule sets and games that use the same dartboard. To understand a dart player requires knowledge of both his or her strategy and the rules of the game.
Like a thrown dart, the wet process leading to a note onset can (and often does, especially in improvisation) result in an event that is close to but not quite exactly at the temporal spot the performer intended. An onset event can be located at various times within a temporal “zone” like the dartboard’s bulls-eye or outermost 20 wedge, yet still have a similar effect (like swinging eighths, each pair slightly different in MS duration, but more-or-less tripletized). A musician’s strategy also comes into play; how does he or she intend a particular onset to work at the phrase level? How will he or she adjust if their wet process leads to unexpected results? Stylistic rules play a prominent and changeable role in the enactment of rhythm and phrase. At what level can the listener or fellow musician truly understand the rules governing a player’s rhythmic impetus? Is the performer even aware of how cultural, situational, personal, and stylistic rules are being integrated into their preparatory microtiming?
Programs that are designed to create groove based patterns, like the organic drum machine, will benefit from a perspective that incorporates the embodied aspect of rhythmic performance, recognizes the relationship between wet and dry process, and uses human data to inform timing at every metric level.
____________________
References
Benadon, Fernando. “Time Warps in Early Jazz.” Music Theory Spectrum, Vol. 31, No. 1 (Spring 2009).
Benadon, Fernando. “A Circular Plot for Rhythm Visualization and Analysis.” Music Theory Online, Vol. 12, No. 3 (September 2007).
Iyer, Vijay. “Embodied Mind, Situated Cognition, and Expressive Microtiming in African-American Music.” Music Perception, Vol. 19, No. 3 (Spring 2002).
Guy Madison, Lea Forsman, Örjan Blom, Anke Karabanov, Fredrik Ullén. “Correlations Between Intelligence and Components of Serial Timing Variability.” Intelligence, # 37 (2009)
