tophu.bandpass_equiripple_filter

tophu.bandpass_equiripple_filter(bandwidth, shape, ripple, attenuation, centerfreq=0.0, samplerate=1.0, *, force_odd_length=False, maxiter=25, grid_density=16)

Design a bandpass FIR digital filter using the Parks-McClellan algorithm.

Form a linear-phase finite impulse response (FIR) discrete-time digital filter wth desired passband, stopband, and attenuation characteristics using the Parks-McClellan algorithm1. The algorithm produces a filter that is optimal in the sense that it minimizes the maximum weighted deviation from the ideal frequency response in the passband and stopband.

The resulting filter has equiripple frequency response in both the passband and stopband.

Parameters

bandwidthfloat

Width of the passband, in the same units as samplerate.

shapefloat

Shape factor – the ratio of the width of the combined passband and transition band to the passband width.

ripplefloat

Passband ripple – the maximum allowed ripple amplitude below unity gain in the passband, in decibels.

attenuationfloat

Stopband attenuation – the difference in amplitude, in decibels, between the ideal gain in the passband and the highest gain in the stopband.

centerfreqfloat, optional

Center frequency of the passband, in the same units as samplerate. Defaults to 0.

sampleratefloat, optional

Sampling frequency of the signal. Defaults to 1.

force_odd_lengthbool, optional

Whether to force the filter length to be odd-valued. Defaults to False.

maxiterint, optional

Maximum number of iterations of the algorithm. Defaults to 25.

grid_densityint, optional

Density of Lagrange interpolation points used in the algorithm. Defaults to 16.

Returns

coeffsnumpy.ndarray

Filter coefficients.

References

1

J. McClellan and T. Parks. A unified approach to the design of optimum FIR linear-phase digital filters. IEEE Transactions on Circuit Theory, 20(6):697–701, November 1973. URL: https://doi.org/10.1109/tct.1973.1083764, doi:10.1109/tct.1973.1083764.