DF

class pylayers.signal.DF.DF(b=array([1]), a=array([1]))

Bases: pylayers.util.project.PyLayers

Digital Filter Class

aarray

transfer function coefficients denominator

barray

transfer function coefficients numerator

parray dtype=complex

transfer function poles

zarray dtype=complex

transfer function zeros

filter : filter a signal freqz : Display transfer function ellip_bp : elliptic bandpath filter remez : FIR design

Attributes Summary

a
b
p
z

Methods Summary

butter([order, w, typ])	Butterwoth digital filter design
ellip_bp(wp, ws[, gpass, gstop])	Elliptic Bandpath filter
factorize()	factorize the filter into a minimal phase and maximal phase
filter(s[, causal])	filter s
flipc()	flip coefficient
freqz(**kwargs)	freqz : evaluation of filter transfer function
ir([N, show])	returns impulse response
match()	return a match filter
remez([numtaps, bands, desired])	FIR design Remez algorithm
simplify([tol])	simplify transfer function
update()
wnxcorr(**kwargs)	filtered noise output autocorrelation
zplane([title, fontsize])	Display filter in the complex plane

Attributes Documentation

a

b

p

z

Methods Documentation

butter(order=5, w=0.25, typ='low')

Butterwoth digital filter design

orderint

filter order

warray_like

a scalar or length-2 sequence (relative frequency 0<1 fN <=> 1)

typstring

‘lowpass’ | ‘highpass’ | ‘bandpass’ | ‘bandstop’

>>> from pylayers.signal.DF import *
>>> flt = DF()
>>> flt.butter(order=3,w=0.25,typ='low')
>>> flt.freqz()

ellip_bp(wp, ws, gpass=0.5, gstop=20)

Elliptic Bandpath filter

wp : ws : gpass gstop :

iirdesign

factorize()

factorize the filter into a minimal phase and maximal phase

(Fmin,Fmax)tuple of filter

Fmin : minimal phase filter Fmax : maximal phase filter

filter(s, causal=True)

filter s

snp.array |TUsignal

input signal

ynp.array | TUsignal

output signal

>>> import matplotlib.pyplot as plt
>>> df = DF(b=np.array([1,1],a=np.array([1,-1]))
>>> N = 100 
>>> s = np.zeros(N)
>>> s[0] = 1
>>> y = df.filter(s)
>>> plt.stem(np.arange(N),x)
>>> plt.stem(np.arange(N),y)
>>> plt.show()

(Source code)

flipc()

flip coefficient

This is equivalent to the transformation

z -> 1/z

freqz(**kwargs)

freqz : evaluation of filter transfer function

The function evaluates an FUsignal

displayboolean

True

fsGHzfloat

if set to 0 (relative frequency)

ir(N=100, show=False)

returns impulse response

irnp.array

impulse response

match()

return a match filter

DF :

if $$H(z)=\frac{1+b_1z^1+…+b_Nz^{-N}}{1+a_1z^1+…+a_Mz^{-M}}$$

it returns

$$H_m(z)=\frac{b_N+b_{N-1}z^1+…+z^{-N}}{1+a_1z^1+…+a_Mz^{-M}}$$

remez(numtaps=401, bands=(0, 0.1, 0.11, 0.2, 0.21, 0.5), desired=(0.0001, 1, 0.0001))

FIR design Remez algorithm

numtaps : int bands : tuple of N floats desired : tuple of N/2 floats

flt.remez(numtaps=401,bands=(0,0.1,0.11,0.2,0.21,0.5),desired=(0.0001,1,0.0001)):

numtaps = 401 bands = (0,0.1,0.11,0.2,0.21,0.5) desired = (0.0001,1,0.0001))

flt.remez( numtaps , bands , desired)

simplify(tol=1e-16)

simplify transfer function

tolfloat

tolerance for simplification

update()

wnxcorr(**kwargs)

filtered noise output autocorrelation

varfloat

input noise variance (default = 1)

zplane(title='', fontsize=18)

Display filter in the complex plane