Synthesizing a Clarinet Nicole Bennett
Overview Frequency modulation Using FM to model instrument signals Generating envelopes Producing a clarinet note A-440 note
Frequency Modulation Used to reproduce signals with frequencies that vary with time General formula: x(t) = A*cos(ψ(t)) Oscillations of ψ(t) provide changes in instantaneous frequency - (ψ′(t))
Producing Instrument Sounds ψ(t) must be sinusoidal in order to reproduce both the fundamental frequency and the overtones of an instrument General equation: x(t) = A(t)*cos(2Π*f c *t + I(t)*cos(2Π*f m *t + Φ m ) + Φ c ) ٭٭ ٭٭John M. Chowning, “The Synthesis of Complex Audio Spectra by Means of Frequency Modulation,” Journal of the Audio Engineering Society, vol.21, no. 7, Sept. 1973, pp
x(t) = A(t)*cos(2Π*f c *t + I(t)*cos(2Π*f m *t + Φ m ) + Φ c ) A(t): amplitude envelope Function of time Allows sound to fade slowly or be cut off quickly f c : carrier frequency Frequency without any modulation f m : modulating frequency Rate of modulation of the instantaneous frequency
x(t) = A(t)*cos(2Π*f c *t + I(t)*cos(2Π*f m *t + Φ m ) + Φ c ) Φ c and Φ m : phase constants Set to Π/2 for this project I(t): modulation index envelope Used to vary the harmonic content of the sound Produces the overtones
Generating A(t) and I(t) WOODWENV2.m WOODWENV2.m
Scaling the I(t) Function A(t) and I(t) are normalized by the WOODWENV function I(t) must be scaled in order to produce a clarinet note scale.mscale.m
Synthesizing a Note Now have most of the pieces of x(t): x(t) = A(t)*cos(2Π*f c *t + I(t)*cos(2Π*f m *t + Φ m ) + Φ c ) f c and f m : ratio is 2:3 for the clarinet f 0 : frequency of the note – will be greatest common divisor of f c and f m
Clarinet Function clarinet.m clarinet.m
Playing a 440 Hz note Play the A note Play the A note Limitations of the equation
Conclusion Modeling instrument signals Generating a clarinet note Problems with modeling an instrument using a mathematical equation