Prev Next gainz

Complex Gain of a Rational Digital Transfer Function
Syntax y = gainz(bz,az,f,fs)
Include: include spt\gainz.oms
See Also Other Window Functions

      bz = COLUMN VECTOR, numerator   polynomial in z^(-1).
      az = COLUMN VECTOR, denominator polynomial in z^(-1).
      f  = MATRIX, any type, evaluation frequencies [Hz].
      fs = SCALAR, any type, sampling rate, [Samples/sec].
   RETURN: MATRIX, COMPLEX, gain at input frequencies.


Calculate the complex gain of a rational digital transfer function.

Arguments 'bz' and 'az' are column vectors representing the numerator and denominator polynomials, respectively, of a digital transfer function in z-transform form. The represented function has the form:

   bz(z)   b(1) + b(2)*z^(-1) + b(3)*z^(-2) + ... + b(Nb)*z^(-N)
   ----- = -----------------------------------------------------
   az(z)   a(1) + a(2)*z^(-1) + a(3)*z^(-2) + ... + a(Ma)*z^(-M)

   where: N  = numerator   polynomial order
          M  = denominator polynomial order
          Nb = rowdim(bz), N = Nb-1
          Ma = rowdim(az), M = Ma-1
          z  = exp(-j*2*PI*f*t/fs), j=sqrt(-1)

The transfer function is evaluated at the frequencies in [Hz] contained in input matrix 'f'. The system sampling rate is specified through argument 'fs' [Hz]. These two arguments are coerced to DOUBLE for local processing.

The function returns a COMPLEX matrix of the same dimensions as 'f'. Absolute value of this return gives frequency response magnitude.

Make a 3rd order Butterworth digital filter and evaluate it at three frequencies.

include spt\gainz.oms            # include gains() function
b  = novalue;                    # Declare numerator polynomial
a  = novalue;                    # Declare denominator polynomial
fnbut(3,b,a);                    # Make an order = 3 Butterworth function
fs = 1d3;                        # Set the sampling frequency
fc = 1d0;                        # Set the lowpass cutoff frequency
fn2dlp(2d0*PI*fc,1d0/fs,b,a,b,a) # Convert to a digital transfer function
f  = {fc/10d0, fc, fc*10d0};     # Make a frequency evaluation vector
H  = gainz(b,a,f,fs)             # Find the complex gain
format complex "f10.6"
H                                # Print the gain results

The results are:

(  0.979999, -0.198999)
( -0.500000, -0.500000)
( -0.000199,  0.000979)