90 Psychoacoustics Transient and Rise Time A
transient corresponds to the excitation stage. It is often considerably louder
than the rest of the sound and is always short. Typical transients are between
5ms and 50ms long. It may contain sound from both the excited object and
the excitor, for example a drum stick or bell hammer. The
rise occurs while
vibrational energy is still building up in the system. As a rule it is often shorter
than one period of the primary mode, or the length of the object divided by the
speed of sound in the material. Often it reaches its maximum before the tran-
sient has finished; thus the two seem inseparable. A good example where this
is not the case is a gong. When hitting a gong the transient of the beater on
metal precedes a short period where the initial excitation wave spreads through
the body. The beater impact creates short-lived high frequencies locally, but
the spread of the displacement is a fairly low-frequency event after which the
major vibrational modes emerge and the gong bursts into life.
Slow Attacks The attack may also be very long. In a gently bowed string the amplitude con-
tinues to increase until the restoring force from the string and the force applied
to the bow reach a maximal dynamic equilibrium limited by friction, which may
take several seconds. A sound that might be considered to have only attack is
an object sliding off an inclined surface into the air. In this case it continues to
accelerate with frictional excitation making a louder and louder sound until it
falls off the edge of the surface. Another example is a simple fireworks rocket.
Ignoring that it usually fades into the distance, the sound output rises in inten-
sity until it burns out or explodes, since an increasing surface area of fuel burns.
Decay Decay applies to systems where energy continues to be supplied after the tran-
sient stage. Such sounds are usually frictional or turbulent, such as dragging
a sack over the ground, or blowing a trumpet. Decay happens after the ini-
tial energy input to the system overshoots the sustainable level. After this the
system reaches some kind of steady dynamic equilibrium called the
sustain .
Sustain During this period energy input to the system equals energy output (minus
internal loss to heat), so the portion of energy that makes up the sound remains
steady. A good example is a bowed string. Schaeffer (1977) refers to this as
ener- getic maintenance in his sonic taxonomy. The sustain stage usually produces
a steady sound output, but this is not always true. Again the bowed string
provides a good example, because periodically the bow must change direction,
so while the overall level of energy transfer is quite even there are points where
it waxes and wanes. An interesting counterexample is a marble rolling down
a long uniform slope. Strictly the system is in a
release stage, since stored
energy is being given up at a constant rate. However, the sound energy pro-
duced is constant and seems sustained. Water flowing over a waterfall is a