Measuring onset timing with a waveform graph
When music has a steady pulse and each of the onsets falls squarely on one of the pulses, an attentive listener can usually tell exactly when the onsets occur and notate the onset timing accurately by ear. This is harder to do when the music either doesn’t have a pulse, or has a pulse that the performer treats flexibly, for instance by placing onsets slightly before or after the pulses or varying the rate at which the pulses flow (both very common expressive techniques in music). In these cases, if we want to specify onset timing precisely, sound analysis software can help.
Let’s return to our first global notation score, an excerpt from a piece by the Korean percussion quartet Samul-Nori, played on a single drum without a regular pulse (see Pulseless onset timing).
For this score, the onset timing was determined using a computer-generated waveform graph—essentially, a graph of loudness against time. Any number of programs can produce a waveform graph from a recording; in this case, the program used was Audacity (see Using sound analysis software). In the graph below, the blue waveform represents the amount of sound at each moment: the taller the waveform, the louder the sound. Time is measured in seconds as marked by the figures along the top of the graph.
The dotted lines indicate how the waveform was converted into global notation. For each drum stroke, a wedge was placed above the layer line, with its left-hand edge vertically above one of the “peaks” in the waveform. (For the first few strokes, dotted vertical lines have been drawn in to show the alignment.) The time scale in seconds is taken from the top of the waveform graph, with the bracket shifted diagonally so that its left-hand edge aligns with the first onset.
Note that the waveform also indicates dynamics: the peaks get smaller as the drum strokes get softer. The global notation score does not reflect this as all the wedges are the same size. However, dynamics can be specified by adding further markings if required.
To some extent, waveform graphs can help establish onset timing for sounds of definite pitch as well, provided each sound has a sufficiently distinct “attack” (as, for instance, with most “impulsive” sounds; see Unspecified duration). But in most cases where definite pitch is involved, it will be more helpful to have a graph that shows pitch as well as timing. Fortunately, sound analysis software can produce this too.
Source of audio:
“Woodo-kut” from Samul-Nori: Drums and Voices of Korea (Seoul: Oasis Record Co., ORC-1041, track 2.