In this section we will discuss activities in the particular case of sound synthesis. We call them sound activities. Like any activity they have a start time s, a duration d, a start program p_s, and an end program p_e. A sound activity represents a sound that starts playing at time s and lasts for d time. Note that it can be any sound: short, long, pure, complex, simple, composed, etc...

In this section we will make the following assumptions about the sound activities:

• they have one parameter that describes the frequency. This frequency is measured in Hertz.

• they have one parameter that describes amplitude. The amplitude is measured on a linear scale from 0 to 1.

• their start program schedules a synthesis process, their stop program kills the same synthesis process.

• they have an associated synthesis voice. The synthesis voice is charged with the creation of the synthesis process.

The start program of an activity takes two steps. First, it creates the synthesis process by calling its associated synthesis voice. It passes the frequency, the amplitude, and any additional parameters to the synthesis voice (see section 5.3.3). Second, it sends an add-event to add the synthesis process to the synthesizer. The end program of a sound activity kills the synthesis process.

We include a remark here. Initially we accepted timbre as the essential concept to describe a sound and considered the classical notions of pitch and intensity as attributes of the timbre. We therefore did not want to ``hard code'' frequency and amplitude in our environment. Studies in the field of musical perception and cognition seem to suggest, however, that the pitch and intensity of sound do occupy a special role in the sound perception (see for example [SD93] and [PPK99]). In addition, not introducing frequency and amplitude would have made most musical transformations cumbersome and loaded the problem of ``timbre representation'' on our backs. We therefore accept the frequency and amplitude as fundamental parameters for sound. Sounds that do not fall in this category (noises of all trades, for example) get this interface for free. They benefit from it, if possible, and can use, for example, the amplitude parameter while neglecting the frequency parameter.

We also define the interface PitchedSound. This interface defines the methods getFreq, setFreq, getAmp, and setAmp. Any object that can handle frequency and amplitude implements this interface and can participate in musical transformations. We use this interface to define sound patterns and sound motifs. They specialize the Pattern and Motif class, respectively. An amplitude envelope and frequency curve can be set for both objects. Associated modifiers can use these curves to change the pattern. We will give examples in the next section.

 

Figure 6.12: The classes and inheritance tree for sound activities. We used a solid line for class inheritance, and a dashed line for Java's interface inheritance.

We have also defined the class Note and Chord for convenience. Both classes inherit from SoundActivity. The note class allows the creation of sound activities with the following parameters:

• A start location and duration.

• A synthesis voice.

• A constant pitch value in midi-cents¹. This value is converted into a constant controller returning the value in Hertz.

• An amplitude value in decibels. This value is converted into a constant controller returning the value in linear amplitude.

The chord class is very similar, except that it accepts a list of pitch and amplitude values. For reasons of efficiency we choose to model a chord as a single activity instead of a pattern. An overview of the musical classes is given in figure 6.12. In the next section we will give some examples of modifiers in the context of sound activities.