Owner`s manual

CHAPTER 3 -
23
CHAPTER 3 -
The Craft of audio Synthesis
3.6.4.3
Cutoff Slope
This is the rate at which a  lter attenuates spectral components, as a function of their
frequency. It is usually a multiple of 6dB/octave.
3.6.4.4
Feedback and Resonance
It is usually possible, with any signal-processing device or system of devices, to mix some
of its output signal back into the input signal. This may be intentional, or it can happen
by accident; everybody who has ever worked with a PA system has experienced the
terrible screech of a system “in feedback”.
In  lter modules for audio synthesis, it is common to provide controllable feedback. With
just a little, the  lter response begins to peak around its cutoff frequency; as the feedback
level increases, the peak gets stronger. Eventually – just as happens with a PA system
– the  lter falls into
oscillation.
Regardless of the input signal, it “screams” a sine-wave at
its cutoff frequency. In this state it is no longer behaving as a  lter at all; it has become
an oscillator.
Systems in feedback have been very well studied in physics. Their behavior can be
described mathematically. Whereas feedback in a PA system can be unpredictable
and uncontrollable, feedback in a  lter module for audio synthesis can be – and is – a
controllable and useful feature.
3.6.4.5
Lag Processors
A
lag processor
is a low-pass  lter intended speci cally for processing subaudio signals. It
lag processor is a low-pass  lter intended speci cally for processing subaudio signals. It lag processor
introduces a “lag” in the output signal wherever the input shows a sharp change in value.
How much of a lag depends on the “cutoff frequency” of the processor.
3.6.5
Modulation Methods
The simplest possible signal, we said above, has exactly three attributes and no more:
frequency, amplitude, and phase. If, starting from a steady-state signal, we systematically
modify any of these characteristics, we are said to be modulating the signal. And so,
based on these three signal attributes, there are three possible forms of modulation:
Amplitude Modulation (AM), Frequency Modulation (FM), and Phase Modulation (PM).
The  rst two are more commonly used in audio synthesis than the third; we won’t say
anything here about phase modulation.
Using AM and FM methods, it is possible to generate waveforms and spectra that are far
more complicated – and interesting to the ear – than anything that can be produced by
merely mixing and  ltering signals. Mainly that is because of sidebands.