CODEWITHSUNDEEP
pythonfilterscipysignal-processingsympy
dvc
dvc

Reputation: 124

Constructing SymPy rational functions from Scipy filter design

Why aren't num_den_to_sympy(b, a) and sy.simplify(sy.expand(tf)) the same? The expanded version contains an additional s ** 4 and I don't know where it came from.

import numpy as np
import sympy as sy
from scipy.signal import *
from IPython.display import display

sy.init_printing()

def num_den_to_sympy(num, den, symplify=True):
    s = sy.Symbol('s')
    G = sy.Poly(num, s) / sy.Poly(den, s)
    return sy.simplify(G) if symplify else G

b, a = iirdesign(wp=2 * np.pi * 2.5e3, ws=2 * np.pi * 4.5e3, gpass=1, gstop=26,
                        analog=True, ftype='cheby1', output='ba')

tf = 1
for sos in tf2sos(b, a):
    tf *= num_den_to_sympy(sos[0:3], sos[3:6])

display(num_den_to_sympy(b, a))
display(sy.simplify(sy.expand(tf)))

Upvotes: 2

Views: 196

Answers (1)

user6655984
user6655984

Reputation:

The issue is the inconsistent order of coefficients. SciPy orders them from lowest to highest degrees. SymPy's Poly constructor expects them to be from highest to lowest: for example, Poly([1, 0, 0], s) is s**2

If you reverse the coefficients when using Poly, the two results are identical.

G = sy.Poly(num[::-1], s) / sy.Poly(den[::-1], s)

Upvotes: 1

Related Questions