scipy 1.16.2__cp311-cp311-win_arm64.whl

scipy/signal/tests/test_filter_design.py

@@ -0,0 +1,4744 @@
+import math
+import cmath
+import warnings
+
+from itertools import product
+
+from scipy._lib import _pep440
+import numpy as np
+from numpy.testing import (
+    assert_array_almost_equal_nulp, assert_warns, suppress_warnings
+)
+import pytest
+from pytest import raises as assert_raises
+from scipy._lib._array_api import (
+    xp_assert_close, xp_assert_equal,
+    assert_array_almost_equal, xp_size, xp_default_dtype, is_numpy
+)
+
+from numpy import array, spacing, sin, pi
+from scipy.signal import (argrelextrema, BadCoefficients, bessel, besselap, bilinear,
+                          buttap, butter, buttord, cheb1ap, cheb1ord, cheb2ap,
+                          cheb2ord, cheby1, cheby2, ellip, ellipap, ellipord,
+                          firwin, freqs_zpk, freqs, freqz, freqz_zpk,
+                          gammatone, group_delay, iircomb, iirdesign, iirfilter,
+                          iirnotch, iirpeak, lp2bp, lp2bs, lp2hp, lp2lp, normalize,
+                          sos2tf, sos2zpk, sosfreqz, freqz_sos, tf2sos, tf2zpk, zpk2sos,
+                          zpk2tf, bilinear_zpk, lp2lp_zpk, lp2hp_zpk, lp2bp_zpk,
+                          lp2bs_zpk)
+from scipy.signal._filter_design import (_cplxreal, _cplxpair, _norm_factor,
+                                        _bessel_poly, _bessel_zeros)
+from scipy.signal._filter_design import _logspace
+from scipy.signal import _polyutils as _pu
+from scipy.signal._polyutils import _sort_cmplx
+
+skip_xp_backends = pytest.mark.skip_xp_backends
+xfail_xp_backends = pytest.mark.xfail_xp_backends
+
+
+try:
+    import mpmath
+except ImportError:
+    mpmath = None
+
+
+def mpmath_check(min_ver):
+    return pytest.mark.skipif(
+        mpmath is None
+        or _pep440.parse(mpmath.__version__) < _pep440.Version(min_ver),
+        reason=f"mpmath version >= {min_ver} required",
+    )
+
+
+class TestCplxPair:
+
+    def test_trivial_input(self):
+        assert _cplxpair([]).size == 0
+        assert _cplxpair(1) == 1
+
+    def test_output_order(self):
+        xp_assert_close(_cplxpair([1+1j, 1-1j]), [1-1j, 1+1j])
+
+        a = [1+1j, 1+1j, 1, 1-1j, 1-1j, 2]
+        b = [1-1j, 1+1j, 1-1j, 1+1j, 1, 2]
+        xp_assert_close(_cplxpair(a), b)
+
+        # points spaced around the unit circle
+        z = np.exp(2j*pi*array([4, 3, 5, 2, 6, 1, 0])/7)
+        z1 = np.copy(z)
+        np.random.shuffle(z)
+        xp_assert_close(_cplxpair(z), z1)
+        np.random.shuffle(z)
+        xp_assert_close(_cplxpair(z), z1)
+        np.random.shuffle(z)
+        xp_assert_close(_cplxpair(z), z1)
+
+        # Should be able to pair up all the conjugates
+        x = np.random.rand(10000) + 1j * np.random.rand(10000)
+        y = x.conj()
+        z = np.random.rand(10000)
+        x = np.concatenate((x, y, z))
+        np.random.shuffle(x)
+        c = _cplxpair(x)
+
+        # Every other element of head should be conjugates:
+        xp_assert_close(c[0:20000:2], np.conj(c[1:20000:2]))
+        # Real parts of head should be in sorted order:
+        xp_assert_close(c[0:20000:2].real, np.sort(c[0:20000:2].real))
+        # Tail should be sorted real numbers:
+        xp_assert_close(c[20000:], np.sort(c[20000:]))
+
+    def test_real_integer_input(self):
+        xp_assert_equal(_cplxpair([2, 0, 1]), [0, 1, 2])
+
+    def test_tolerances(self):
+        eps = spacing(1)
+        xp_assert_close(_cplxpair([1j, -1j, 1+1j*eps], tol=2*eps),
+                        [-1j, 1j, 1+1j*eps])
+
+        # sorting close to 0
+        xp_assert_close(_cplxpair([-eps+1j, +eps-1j]), [-1j, +1j])
+        xp_assert_close(_cplxpair([+eps+1j, -eps-1j]), [-1j, +1j])
+        xp_assert_close(_cplxpair([+1j, -1j]), [-1j, +1j])
+
+    def test_unmatched_conjugates(self):
+        # 1+2j is unmatched
+        assert_raises(ValueError, _cplxpair, [1+3j, 1-3j, 1+2j])
+
+        # 1+2j and 1-3j are unmatched
+        assert_raises(ValueError, _cplxpair, [1+3j, 1-3j, 1+2j, 1-3j])
+
+        # 1+3j is unmatched
+        assert_raises(ValueError, _cplxpair, [1+3j, 1-3j, 1+3j])
+
+        # Not conjugates
+        assert_raises(ValueError, _cplxpair, [4+5j, 4+5j])
+        assert_raises(ValueError, _cplxpair, [1-7j, 1-7j])
+
+        # No pairs
+        assert_raises(ValueError, _cplxpair, [1+3j])
+        assert_raises(ValueError, _cplxpair, [1-3j])
+
+
+class TestCplxReal:
+
+    def test_trivial_input(self):
+        assert all(x.size == 0 for x in _cplxreal([]))
+
+        x = _cplxreal(1)
+        assert x[0].size == 0
+        xp_assert_equal(x[1], np.asarray([1]))
+
+
+    def test_output_order(self):
+        zc, zr = _cplxreal(np.roots(array([1, 0, 0, 1])))
+        xp_assert_close(np.append(zc, zr), [1/2 + 1j*sin(pi/3), -1])
+
+        eps = spacing(1)
+
+        a = [0+1j, 0-1j, eps + 1j, eps - 1j, -eps + 1j, -eps - 1j,
+             1, 4, 2, 3, 0, 0,
+             2+3j, 2-3j,
+             1-eps + 1j, 1+2j, 1-2j, 1+eps - 1j,  # sorts out of order
+             3+1j, 3+1j, 3+1j, 3-1j, 3-1j, 3-1j,
+             2-3j, 2+3j]
+        zc, zr = _cplxreal(a)
+        xp_assert_close(zc, [1j, 1j, 1j, 1+1j, 1+2j, 2+3j, 2+3j, 3+1j, 3+1j,
+                             3+1j])
+        xp_assert_close(zr, [0.0, 0, 1, 2, 3, 4])
+
+        z = array([1-eps + 1j, 1+2j, 1-2j, 1+eps - 1j, 1+eps+3j, 1-2*eps-3j,
+                   0+1j, 0-1j, 2+4j, 2-4j, 2+3j, 2-3j, 3+7j, 3-7j, 4-eps+1j,
+                   4+eps-2j, 4-1j, 4-eps+2j])
+
+        zc, zr = _cplxreal(z)
+        xp_assert_close(zc, [1j, 1+1j, 1+2j, 1+3j, 2+3j, 2+4j, 3+7j, 4+1j,
+                             4+2j])
+        xp_assert_equal(zr, np.asarray([]))
+
+    def test_unmatched_conjugates(self):
+        # 1+2j is unmatched
+        assert_raises(ValueError, _cplxreal, [1+3j, 1-3j, 1+2j])
+
+        # 1+2j and 1-3j are unmatched
+        assert_raises(ValueError, _cplxreal, [1+3j, 1-3j, 1+2j, 1-3j])
+
+        # 1+3j is unmatched
+        assert_raises(ValueError, _cplxreal, [1+3j, 1-3j, 1+3j])
+
+        # No pairs
+        assert_raises(ValueError, _cplxreal, [1+3j])
+        assert_raises(ValueError, _cplxreal, [1-3j])
+
+    def test_real_integer_input(self):
+        zc, zr = _cplxreal([2, 0, 1, 4])
+        xp_assert_equal(zc, [])
+        xp_assert_equal(zr, [0, 1, 2, 4])
+
+
+class TestTf2zpk:
+
+    @skip_xp_backends(
+        cpu_only=True, reason="XXX zpk2sos is numpy-only", exceptions=['cupy']
+    )
+    @skip_xp_backends("dask.array", reason="https://github.com/dask/dask/issues/11883")
+    @pytest.mark.parametrize('dt', ('float64', 'complex128'))
+    def test_simple(self, dt, xp):
+        dtyp = getattr(xp, dt)
+
+        z_r = xp.asarray([0.5, -0.5])
+        p_r = xp.asarray([1.j / math.sqrt(2), -1.j / math.sqrt(2)])
+        # Sort the zeros/poles so that we don't fail the test if the order
+        # changes
+        z_r = _sort_cmplx(z_r, xp=xp)
+        p_r = _sort_cmplx(p_r, xp=xp)
+
+        b = xp.astype(_pu.poly(z_r, xp=xp), dtyp)
+        a = xp.astype(_pu.poly(p_r, xp=xp), dtyp)
+
+        z, p, k = tf2zpk(b, a)
+        z = _sort_cmplx(z, xp=xp)
+        # The real part of `p` is ~0.0, so sort by imaginary part
+        p = p[xp.argsort(xp.imag(p))]
+
+        assert_array_almost_equal(z, z_r)
+        assert_array_almost_equal(p, p_r)
+        assert math.isclose(xp.real(k), 1.)
+        assert k.dtype == dtyp
+
+    def test_bad_filter(self):
+        # Regression test for #651: better handling of badly conditioned
+        # filter coefficients.
+        with suppress_warnings():
+            warnings.simplefilter("error", BadCoefficients)
+            assert_raises(BadCoefficients, tf2zpk, [1e-15], [1.0, 1.0])
+
+
+class TestZpk2Tf:
+
+    def test_identity(self, xp):
+        """Test the identity transfer function."""
+        z = xp.asarray([])
+        p = xp.asarray([])
+        k = 1.
+        b, a = zpk2tf(z, p, k)
+        b_r = xp.asarray([1.])  # desired result
+        a_r = xp.asarray([1.])  # desired result
+        # The test for the *type* of the return values is a regression
+        # test for ticket #1095. In the case p=[], zpk2tf used to
+        # return the scalar 1.0 instead of array([1.0]).
+        xp_assert_equal(b, b_r)
+        xp_assert_equal(a, a_r)
+        if is_numpy(xp):
+            assert isinstance(b, np.ndarray)
+            assert isinstance(a, np.ndarray)
+
+    @skip_xp_backends("dask.array", reason="https://github.com/dask/dask/issues/11883")
+    @skip_xp_backends(cpu_only=True, reason="XXX zpk2sos is numpy-only")
+    def test_conj_pair(self, xp):
+        # conjugate pairs give real-coeff num & den
+        z = xp.asarray([1j, -1j, 2j, -2j])
+        # shouldn't need elements of pairs to be adjacent
+        p = xp.asarray([1+1j, 3-100j, 3+100j, 1-1j])
+        k = 23
+
+        # np.poly should do the right thing, but be explicit about
+        # taking real part
+        z_np, p_np = map(np.asarray, (z, p))
+        b_np = k * np.poly(z_np).real
+        a_np = np.poly(p_np).real
+        b, a = map(xp.asarray, (b_np, a_np))
+
+        bp, ap = zpk2tf(z, p, k)
+
+        xp_assert_close(b, bp)
+        xp_assert_close(a, ap)
+
+        assert xp.isdtype(bp.dtype, 'real floating')
+        assert xp.isdtype(ap.dtype, 'real floating')
+
+    @skip_xp_backends("dask.array", reason="https://github.com/dask/dask/issues/11883")
+    @skip_xp_backends(
+        cpu_only=True, reason="XXX zpk2sos is numpy-only", exceptions=['cupy']
+    )
+    def test_complexk(self, xp):
+        # regression: z, p real, k complex k gave real b, a
+        b, a = xp.asarray([1j, 1j]), xp.asarray([1.0, 2])
+        z, p, k = tf2zpk(b, a)
+        xp_assert_close(k, xp.asarray(1j), check_0d=False)
+        bp, ap = zpk2tf(z, p, k)
+        xp_assert_close(b, bp)
+        xp_assert_close(a, ap)
+
+
+@skip_xp_backends("jax.numpy", reason='no eig in JAX on GPU.')
+class TestSos2Zpk:
+
+    @skip_xp_backends("dask.array", reason="it https://github.com/dask/dask/issues/11883")
+    def test_basic(self, xp):
+        sos = [[1, 0, 1, 1, 0, -0.81],
+               [1, 0, 0, 1, 0, +0.49]]
+        sos = xp.asarray(sos)
+        z, p, k = sos2zpk(sos)
+        z2 = xp.asarray([1j, -1j, 0, 0])
+        p2 = xp.asarray([0.9, -0.9, 0.7j, -0.7j])
+        k2 = 1.
+        assert_array_almost_equal(_sort_cmplx(z, xp), _sort_cmplx(z2, xp), decimal=4)
+        assert_array_almost_equal(_sort_cmplx(p, xp), _sort_cmplx(p2, xp), decimal=4)
+        assert math.isclose(k, k2)
+
+        sos = [[1.00000, +0.61803, 1.0000, 1.00000, +0.60515, 0.95873],
+               [1.00000, -1.61803, 1.0000, 1.00000, -1.58430, 0.95873],
+               [1.00000, +1.00000, 0.0000, 1.00000, +0.97915, 0.00000]]
+        sos = xp.asarray(sos)
+        z, p, k = sos2zpk(sos)
+        z2 = [-0.3090 + 0.9511j, -0.3090 - 0.9511j, 0.8090 + 0.5878j,
+              0.8090 - 0.5878j, -1.0000 + 0.0000j, 0]
+        p2 = [-0.3026 + 0.9312j, -0.3026 - 0.9312j, 0.7922 + 0.5755j,
+              0.7922 - 0.5755j, -0.9791 + 0.0000j, 0]
+        z2 = xp.asarray(z2)
+        p2 = xp.asarray(p2)
+        k2 = 1
+        assert_array_almost_equal(_sort_cmplx(z, xp), _sort_cmplx(z2, xp), decimal=4)
+        assert_array_almost_equal(_sort_cmplx(p, xp), _sort_cmplx(p2, xp), decimal=4)
+
+        sos = array([[1, 2, 3, 1, 0.2, 0.3],
+                     [4, 5, 6, 1, 0.4, 0.5]])
+        z = array([-1 - 1.41421356237310j, -1 + 1.41421356237310j,
+                  -0.625 - 1.05326872164704j, -0.625 + 1.05326872164704j])
+        p = array([-0.2 - 0.678232998312527j, -0.2 + 0.678232998312527j,
+                  -0.1 - 0.538516480713450j, -0.1 + 0.538516480713450j])
+        sos, z, p = map(xp.asarray, (sos, z, p))
+        k = 4
+        z2, p2, k2 = sos2zpk(sos)
+
+        xp_assert_close(_sort_cmplx(z2, xp=xp), _sort_cmplx(z, xp=xp))
+        xp_assert_close(_sort_cmplx(p2, xp=xp), _sort_cmplx(p, xp=xp))
+        assert k2 == k
+
+    @pytest.mark.thread_unsafe
+    def test_fewer_zeros(self, xp):
+        """Test not the expected number of p/z (effectively at origin)."""
+        sos = butter(3, 0.1, output='sos')
+        sos = xp.asarray(sos)   # XXX convert butter
+        z, p, k = sos2zpk(sos)
+        assert z.shape[0] == 4
+        assert p.shape[0] == 4
+
+        sos = butter(12, [5., 30.], 'bandpass', fs=1200., analog=False,
+                    output='sos')
+        xp = xp.asarray(sos)
+        with pytest.warns(BadCoefficients, match='Badly conditioned'):
+            z, p, k = sos2zpk(sos)
+        assert z.shape[0] == 24
+        assert p.shape[0] == 24
+
+
+@skip_xp_backends(
+    cpu_only=True, reason="XXX zpk2sos is numpy-only", exceptions=['cupy']
+)
+class TestSos2Tf:
+
+    def test_basic(self, xp):
+        sos = [[1.0, 1, 1, 1, 0, -1],
+               [-2, 3, 1, 1, 10, 1]]
+        sos = xp.asarray(sos)
+        b, a = sos2tf(sos)
+        assert_array_almost_equal(b, xp.asarray([-2.0, 1, 2, 4, 1]))
+        assert_array_almost_equal(a, xp.asarray([1.0, 10, 0, -10, -1]))
+
+
+@skip_xp_backends(cpu_only=True, reason="XXX zpk2sos is numpy-only")
+class TestTf2Sos:
+
+    def test_basic(self, xp):
+        num = xp.asarray([2., 16, 44, 56, 32])
+        den = xp.asarray([3., 3, -15, 18, -12])
+        sos = tf2sos(num, den)
+        sos2 = [[0.6667, 4.0000, 5.3333, 1.0000, +2.0000, -4.0000],
+                [1.0000, 2.0000, 2.0000, 1.0000, -1.0000, +1.0000]]
+        sos2 = xp.asarray(sos2)
+        assert_array_almost_equal(sos, sos2, decimal=4)
+
+        b = xp.asarray([1.0, -3, 11, -27, 18])
+        a = xp.asarray([16.0, 12, 2, -4, -1])
+        sos = tf2sos(b, a)
+        sos2 = [[0.0625, -0.1875, 0.1250, 1.0000, -0.2500, -0.1250],
+                [1.0000, +0.0000, 9.0000, 1.0000, +1.0000, +0.5000]]
+        sos2 = xp.asarray(sos2)
+        #assert_array_almost_equal(sos, sos2, decimal=4)
+
+    @pytest.mark.parametrize('b, a, analog, sos',
+                             [([1.0], [1.0], False, [[1., 0., 0., 1., 0., 0.]]),
+                              ([1.0], [1.0], True, [[0., 0., 1., 0., 0., 1.]]),
+                              ([1.0], [1., 0., -1.01, 0, 0.01], False,
+                               [[1., 0., 0., 1., 0., -0.01],
+                                [1., 0., 0., 1., 0., -1]]),
+                              ([1.0], [1., 0., -1.01, 0, 0.01], True,
+                               [[0., 0., 1., 1., 0., -1],
+                                [0., 0., 1., 1., 0., -0.01]])])
+    def test_analog(self, b, a, analog, sos, xp):
+        b, a, sos = map(xp.asarray, (b, a, sos))
+        sos2 = tf2sos(b, a, analog=analog)
+        assert_array_almost_equal(sos, sos2, decimal=4)
+
+    def test_gh_23221(self):
+        # Tests that this tf2sos call below does not produce ComplexWarnings
+        # This test is specific for scipy==1.16.0: later scipy versions do not produce
+        # the warning.
+        with suppress_warnings():
+            warnings.simplefilter("error")
+            tf2sos([0.21860986786301265, -0.4372197357260253, -0.2186098678630126,
+                    0.8744394714520509,  -0.21860986786301248, -0.4372197357260253,
+                    0.21860986786301265],
+                   [1., -4.18323041786553, 6.829924151626914, -5.407777865686045,
+                    2.0773105450802336,  -0.33482732571537893,  0.0186009178695853 ]
+            )
+
+
+@skip_xp_backends(
+    cpu_only=True, reason="XXX zpk2sos is numpy-only", exceptions=['cupy']
+)
+class TestZpk2Sos:
+
+#    @pytest.mark.parametrize('dt', 'fdgFDG')
+    # XXX: quietly remove float128 and complex256
+    @pytest.mark.parametrize('dt', ['float32', 'float64', 'complex64', 'complex128'])
+    @pytest.mark.parametrize('pairing, analog',
+                             [('nearest', False),
+                              ('keep_odd', False),
+                              ('minimal', False),
+                              ('minimal', True)])
+    def test_dtypes(self, dt, pairing, analog, xp):
+        dtype = getattr(xp, dt)
+        # the poles have to be complex
+        cdtype = (xp.empty(1, dtype=dtype) + 1j*xp.empty(1, dtype=dtype)).dtype
+
+        z = xp.asarray([-1, -1], dtype=dtype)
+        p = xp.asarray([0.57149 + 0.29360j, 0.57149 - 0.29360j], dtype=cdtype)
+        k = xp.asarray(1, dtype=dtype)
+        sos = zpk2sos(z, p, k, pairing=pairing, analog=analog)
+        # octave & MATLAB
+        sos2 = xp.asarray([[1, 2, 1, 1, -1.14298, 0.41280]], dtype=dtype)
+        assert_array_almost_equal(sos, sos2, decimal=4)
+
+    def test_basic(self, xp):
+        for pairing in ('nearest', 'keep_odd'):
+            #
+            # Cases that match octave
+            #
+
+            z = xp.asarray([-1.0, -1.0])
+            p = xp.asarray([0.57149 + 0.29360j, 0.57149 - 0.29360j])
+            k = 1
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            sos2 = xp.asarray([[1, 2, 1, 1, -1.14298, 0.41280]])  # octave & MATLAB
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+            z = xp.asarray([1j, -1j])
+            p = xp.asarray([0.9, -0.9, 0.7j, -0.7j])
+            k = 1
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            sos2 = [[1, 0, 1, 1, 0, +0.49],
+                    [1, 0, 0, 1, 0, -0.81]]  # octave
+            sos2 = xp.asarray(sos2)
+            # sos2 = [[0, 0, 1, 1, -0.9, 0],
+            #         [1, 0, 1, 1, 0.9, 0]]  # MATLAB
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+            z = xp.asarray([])
+            p = xp.asarray([0.8, -0.5+0.25j, -0.5-0.25j])
+            k = 1.
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            sos2 = [[1., 0., 0., 1., 1., 0.3125],
+                    [1., 0., 0., 1., -0.8, 0.]]  # octave, MATLAB fails
+            sos2 = xp.asarray(sos2)
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+            z = xp.asarray([1., 1., 0.9j, -0.9j])
+            p = xp.asarray([0.99+0.01j, 0.99-0.01j, 0.1+0.9j, 0.1-0.9j])
+            k = 1
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            sos2 = [[1, 0, 0.81, 1, -0.2, 0.82],
+                    [1, -2, 1, 1, -1.98, 0.9802]]  # octave
+            sos2 = xp.asarray(sos2)
+            # sos2 = [[1, -2, 1, 1,  -0.2, 0.82],
+            #         [1, 0, 0.81, 1, -1.98, 0.9802]]  # MATLAB
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+            z = xp.asarray([0.9+0.1j, 0.9-0.1j, -0.9])
+            p = xp.asarray([0.75+0.25j, 0.75-0.25j, 0.9])
+            k = 1
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            if pairing == 'keep_odd':
+                sos2 = [[1, -1.8, 0.82, 1, -1.5, 0.625],
+                        [1, 0.9, 0, 1, -0.9, 0]]  # octave; MATLAB fails
+                sos2 = xp.asarray(sos2)
+                assert_array_almost_equal(sos, sos2, decimal=4)
+            else:  # pairing == 'nearest'
+                sos2 = [[1, 0.9, 0, 1, -1.5, 0.625],
+                        [1, -1.8, 0.82, 1, -0.9, 0]]  # our algorithm
+                sos2 = xp.asarray(sos2)
+                assert_array_almost_equal(sos, sos2, decimal=4)
+
+            #
+            # Cases that differ from octave:
+            #
+
+            z = [-0.3090 + 0.9511j, -0.3090 - 0.9511j, 0.8090 + 0.5878j,
+                 +0.8090 - 0.5878j, -1.0000 + 0.0000j]
+            p = [-0.3026 + 0.9312j, -0.3026 - 0.9312j, 0.7922 + 0.5755j,
+                 +0.7922 - 0.5755j, -0.9791 + 0.0000j]
+            z = xp.asarray(z)
+            p = xp.asarray(p)
+            k = 1
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            # sos2 = [[1, 0.618, 1, 1, 0.6052, 0.95870],
+            #         [1, -1.618, 1, 1, -1.5844, 0.95878],
+            #         [1, 1, 0, 1, 0.9791, 0]]  # octave, MATLAB fails
+            sos2 = [[1, 1, 0, 1, +0.97915, 0],
+                    [1, 0.61803, 1, 1, +0.60515, 0.95873],
+                    [1, -1.61803, 1, 1, -1.58430, 0.95873]]
+            sos2 = xp.asarray(sos2)
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+            z = [-1 - 1.4142j, -1 + 1.4142j,
+                 -0.625 - 1.0533j, -0.625 + 1.0533j]
+            p = [-0.2 - 0.6782j, -0.2 + 0.6782j,
+                 -0.1 - 0.5385j, -0.1 + 0.5385j]
+            z = xp.asarray(z)
+            p = xp.asarray(p)
+            k = 4
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            sos2 = [[4, 8, 12, 1, 0.2, 0.3],
+                    [1, 1.25, 1.5, 1, 0.4, 0.5]]  # MATLAB
+            sos2 = xp.asarray(sos2, dtype=xp.float64)
+            # sos2 = [[4, 8, 12, 1, 0.4, 0.5],
+            #         [1, 1.25, 1.5, 1, 0.2, 0.3]]  # octave
+            xp_assert_close(sos, sos2, rtol=1e-4, atol=1e-4)
+
+            z = xp.asarray([])
+            p = xp.asarray([0.2, -0.5+0.25j, -0.5-0.25j])
+            k = 1.
+            sos = zpk2sos(z, p, k, pairing=pairing)
+            sos2 = [[1., 0., 0., 1., -0.2, 0.],
+                    [1., 0., 0., 1., 1., 0.3125]]
+            sos2 = xp.asarray(sos2)
+            # sos2 = [[1., 0., 0., 1., 1., 0.3125],
+            #         [1., 0., 0., 1., -0.2, 0]]  # octave, MATLAB fails
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+            # The next two examples are adapted from Leland B. Jackson,
+            # "Digital Filters and Signal Processing (1995) p.400:
+            # http://books.google.com/books?id=VZ8uabI1pNMC&lpg=PA400&ots=gRD9pi8Jua&dq=Pole%2Fzero%20pairing%20for%20minimum%20roundoff%20noise%20in%20BSF.&pg=PA400#v=onepage&q=Pole%2Fzero%20pairing%20for%20minimum%20roundoff%20noise%20in%20BSF.&f=false
+
+            deg2rad = xp.pi / 180.
+            k = 1.
+
+            # first example
+            thetas = xp.asarray([22.5, 45, 77.5])
+            mags = xp.asarray([0.8, 0.6, 0.9])
+            z = xp.exp(1j * deg2rad * thetas)
+            z = xp.concat((z, xp.conj(z)))
+            p = xp.exp(1j * deg2rad * thetas) * mags
+            p = xp.concat((p, xp.conj(p)))
+            sos = zpk2sos(z, p, k)
+            # sos2 = [[1, -0.43288, 1, 1, -0.38959, 0.81],  # octave,
+            #         [1, -1.41421, 1, 1, -0.84853, 0.36],  # MATLAB fails
+            #         [1, -1.84776, 1, 1, -1.47821, 0.64]]
+            # Note that pole-zero pairing matches, but ordering is different
+            sos2 = [[1, -1.41421, 1, 1, -0.84853, 0.36],
+                    [1, -1.84776, 1, 1, -1.47821, 0.64],
+                    [1, -0.43288, 1, 1, -0.38959, 0.81]]
+            sos2 = xp.asarray(sos2)
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+            # second example
+            thetas = xp.asarray([85., 10.])
+            z = xp.exp(1j * deg2rad * thetas)
+            z = xp.concat((z, xp.conj(z), xp.asarray([1.0, -1.0])))
+            sos = zpk2sos(z, p, k)
+
+            # sos2 = [[1, -0.17431, 1, 1, -0.38959, 0.81],  # octave "wrong",
+            #         [1, -1.96962, 1, 1, -0.84853, 0.36],  # MATLAB fails
+            #         [1, 0, -1, 1, -1.47821, 0.64000]]
+            # Our pole-zero pairing matches the text, Octave does not
+            sos2 = [[1, 0, -1, 1, -0.84853, 0.36],
+                    [1, -1.96962, 1, 1, -1.47821, 0.64],
+                    [1, -0.17431, 1, 1, -0.38959, 0.81]]
+            sos2 = xp.asarray(sos2)
+            assert_array_almost_equal(sos, sos2, decimal=4)
+
+    # these examples are taken from the doc string, and show the
+    # effect of the 'pairing' argument
+    @pytest.mark.parametrize('pairing, sos',
+                             [('nearest',
+                               np.array([[1., 1., 0.5, 1., -0.75, 0.],
+                                         [1., 1., 0., 1., -1.6, 0.65]])),
+                              ('keep_odd',
+                               np.array([[1., 1., 0, 1., -0.75, 0.],
+                                         [1., 1., 0.5, 1., -1.6, 0.65]])),
+                              ('minimal',
+                               np.array([[0., 1., 1., 0., 1., -0.75],
+                                         [1., 1., 0.5, 1., -1.6, 0.65]]))])
+    def test_pairing(self, pairing, sos, xp):
+        sos = xp.asarray(sos)
+        z1 = xp.asarray([-1, -0.5-0.5j, -0.5+0.5j])
+        p1 = xp.asarray([0.75, 0.8+0.1j, 0.8-0.1j])
+        sos2 = zpk2sos(z1, p1, 1, pairing=pairing)
+        assert_array_almost_equal(sos, sos2, decimal=4)
+
+    @pytest.mark.parametrize('p, sos_dt',
+                             [([-1, 1, -0.1, 0.1],
+                               [[0., 0., 1., 1., 0., -0.01],
+                                [0., 0., 1., 1., 0., -1]]),
+                              ([-0.7071+0.7071j, -0.7071-0.7071j, -0.1j, 0.1j],
+                               [[0., 0., 1., 1., 0., 0.01],
+                                [0., 0., 1., 1., 1.4142, 1.]])])
+    def test_analog(self, p, sos_dt, xp):
+        # test `analog` argument
+        # for discrete time, poles closest to unit circle should appear last
+        # for cont. time, poles closest to imaginary axis should appear last
+        z, p = xp.asarray([]), xp.asarray(p)
+        sos_dt = xp.asarray(sos_dt)
+        sos2_dt = zpk2sos(z, p, 1, pairing='minimal', analog=False)
+        sos2_ct = zpk2sos(z, p, 1, pairing='minimal', analog=True)
+        assert_array_almost_equal(sos_dt, sos2_dt, decimal=4)
+        assert_array_almost_equal(xp.flip(sos_dt, axis=0), sos2_ct, decimal=4)
+
+    def test_bad_args(self):
+        with pytest.raises(ValueError, match=r'pairing must be one of'):
+            zpk2sos([1], [2], 1, pairing='no_such_pairing')
+
+        with pytest.raises(ValueError, match=r'.*pairing must be "minimal"'):
+            zpk2sos([1], [2], 1, pairing='keep_odd', analog=True)
+
+        with pytest.raises(ValueError,
+                           match=r'.*must have len\(p\)>=len\(z\)'):
+            zpk2sos([1, 1], [2], 1, analog=True)
+
+        with pytest.raises(ValueError, match=r'k must be real'):
+            zpk2sos([1], [2], k=1j)
+
+
+class TestFreqs:
+
+    def test_basic(self, xp):
+        _, h = freqs(xp.asarray([1.0]), xp.asarray([1.0]), worN=8)
+        assert_array_almost_equal(h, xp.ones(8))
+
+    def test_output(self, xp):
+        # 1st order low-pass filter: H(s) = 1 / (s + 1)
+        w = xp.asarray([0.1, 1, 10, 100])
+        num = xp.asarray([1.])
+        den = xp.asarray([1, 1.])
+        w, H = freqs(num, den, worN=w)
+        s = w * 1j
+        expected = 1 / (s + 1)
+        assert_array_almost_equal(xp.real(H), xp.real(expected))
+        assert_array_almost_equal(xp.imag(H), xp.imag(expected))
+
+    @skip_xp_backends("jax.numpy", reason="eigvals not available on CUDA")
+    def test_freq_range(self, xp):
+        # Test that freqresp() finds a reasonable frequency range.
+        # 1st order low-pass filter: H(s) = 1 / (s + 1)
+        # Expected range is from 0.01 to 10.
+        num = xp.asarray([1.])
+        den = xp.asarray([1, 1.])
+        n = 10
+        expected_w = _logspace(-2, 1, n, xp=xp)
+        w, H = freqs(num, den, worN=n)
+        assert_array_almost_equal(w, expected_w)
+
+    def test_plot(self, xp):
+
+        def plot(w, h):
+            assert_array_almost_equal(h, xp.ones(8))
+
+        with assert_raises(ZeroDivisionError):
+            freqs([1.0], [1.0], worN=8, plot=lambda w, h: 1 / 0)
+
+        freqs(xp.asarray([1.0]), xp.asarray([1.0]), worN=8, plot=plot)
+
+    def test_backward_compat(self, xp):
+        # For backward compatibility, test if None act as a wrapper for default
+        w1, h1 = freqs(xp.asarray([1.0]), xp.asarray([1.0]))
+        w2, h2 = freqs(xp.asarray([1.0]), xp.asarray([1.0]), None)
+        assert_array_almost_equal(w1, w2)
+        assert_array_almost_equal(h1, h2)
+
+    def test_w_or_N_types(self):
+        # Measure at 8 equally-spaced points
+        for N in (8, np.int8(8), np.int16(8), np.int32(8), np.int64(8),
+                  np.array(8)):
+            w, h = freqs([1.0], [1.0], worN=N)
+            assert len(w) == 8
+            assert_array_almost_equal(h, np.ones(8))
+
+        # Measure at frequency 8 rad/sec
+        for w in (8.0, 8.0+0j):
+            w_out, h = freqs([1.0], [1.0], worN=w)
+            assert_array_almost_equal(w_out, [8])
+            assert_array_almost_equal(h, [1])
+
+
+class TestFreqs_zpk:
+
+    def test_basic(self, xp):
+        _, h = freqs_zpk(
+            xp.asarray([1.0]), xp.asarray([1.0]), xp.asarray([1.0]), worN=8
+        )
+        assert_array_almost_equal(h, xp.ones(8))
+
+    def test_output(self, xp):
+        # 1st order low-pass filter: H(s) = 1 / (s + 1)
+        w = xp.asarray([0.1, 1, 10, 100])
+        z = xp.asarray([])
+        p = xp.asarray([-1.0])
+        k = 1
+        w, H = freqs_zpk(z, p, k, worN=w)
+        s = w * 1j
+        expected = 1 / (s + 1)
+        assert_array_almost_equal(xp.real(H), xp.real(expected))
+        assert_array_almost_equal(xp.imag(H), xp.imag(expected))
+
+    def test_freq_range(self, xp):
+        # Test that freqresp() finds a reasonable frequency range.
+        # 1st order low-pass filter: H(s) = 1 / (s + 1)
+        # Expected range is from 0.01 to 10.
+        z = xp.asarray([])
+        p = xp.asarray([-1.])
+        k = 1
+        n = 10
+        expected_w = _logspace(-2, 1, n, xp=xp)
+        w, H = freqs_zpk(z, p, k, worN=n)
+        assert_array_almost_equal(w, expected_w)
+
+    @skip_xp_backends("jax.numpy", reason="eigvals not available on CUDA")
+    def test_vs_freqs(self, xp):
+        b, a = cheby1(4, 5, 100, analog=True, output='ba')
+        z, p, k = cheby1(4, 5, 100, analog=True, output='zpk')
+
+        b, a = map(xp.asarray, (b, a))    # XXX convert cheby1
+        z, p = map(xp.asarray, (z, p))
+
+        w1, h1 = freqs(b, a)
+        w2, h2 = freqs_zpk(z, p, k)
+        xp_assert_close(w1, w2)
+        xp_assert_close(h1, h2, rtol=1e-6)
+
+    def test_backward_compat(self, xp):
+        # For backward compatibility, test if None act as a wrapper for default
+        # Also, keep testing `k` a length-one list: it is documented as a scalar,
+        # but the implementation was allowing for a one-element array-likes
+        w1, h1 = freqs_zpk(xp.asarray([1.0]), xp.asarray([1.0]), [1.0])
+        w2, h2 = freqs_zpk(xp.asarray([1.0]), xp.asarray([1.0]), [1.0], None)
+        assert_array_almost_equal(w1, w2)
+        assert_array_almost_equal(h1, h2)
+
+    def test_w_or_N_types(self):
+        # Measure at 8 equally-spaced points
+        for N in (8, np.int8(8), np.int16(8), np.int32(8), np.int64(8),
+                  np.array(8)):
+            w, h = freqs_zpk([], [], 1, worN=N)
+            assert len(w) == 8
+            assert_array_almost_equal(h, np.ones(8))
+
+        # Measure at frequency 8 rad/sec
+        for w in (8.0, 8.0+0j):
+            w_out, h = freqs_zpk([], [], 1, worN=w)
+            assert_array_almost_equal(w_out, [8])
+            assert_array_almost_equal(h, [1])
+
+
+class TestFreqz:
+
+    def test_ticket1441(self, xp):
+        """Regression test for ticket 1441."""
+        # Because freqz previously used arange instead of linspace,
+        # when N was large, it would return one more point than
+        # requested.
+        N = 100000
+        w, h = freqz(xp.asarray([1.0]), worN=N)
+        assert w.shape == (N,)
+
+    def test_gh_22886(self, xp):
+        w, h = freqz(xp.asarray([1.]), worN=xp.asarray([0., 0.1]))
+        xp_assert_equal(w, xp.asarray([0. , 0.1]))
+        xp_assert_equal(h, xp.asarray([1.+0.j, 1.+0.j]))
+
+    def test_gh_23277(self):
+        # backwards compatibility: `w` array must be real, not complex
+        filt = [0.5 + 0.0j, 0.5 + 0.0j]
+        w, _ = freqz(filt, worN=8)
+        assert w.dtype == np.float64
+
+    def test_basic(self, xp):
+        w, h = freqz(xp.asarray([1.0]), worN=8)
+        assert_array_almost_equal(w, xp.pi * xp.arange(8, dtype=w.dtype) / 8.)
+        assert_array_almost_equal(h, xp.ones(8))
+        w, h = freqz(xp.asarray([1.0]), worN=9)
+        assert_array_almost_equal(w, xp.pi * xp.arange(9, dtype=w.dtype) / 9.)
+        assert_array_almost_equal(h, xp.ones(9))
+
+        for a in [1, xp.ones(2)]:
+            w, h = freqz(xp.ones(2), a, worN=0)
+            assert w.shape == (0,)
+            assert h.shape == (0,)
+            hdt = xp.complex128 if xp_default_dtype(xp) == xp.float64 else xp.complex64
+            assert h.dtype == hdt
+
+    def test_basic2(self, xp):
+        t = xp.linspace(0, 1, 4, endpoint=False)
+        for b, a, h_whole in zip(
+                (xp.asarray([1., 0, 0, 0]), xp.sin(2 * xp.pi * t)),
+                (xp.asarray([1., 0, 0, 0]), xp.asarray([0.5, 0, 0, 0])),
+                (xp.asarray([1., 1., 1., 1.]), xp.asarray([0, -4j, 0, 4j]))
+        ):
+
+            w, h = freqz(b, a, worN=4, whole=True)
+            expected_w = xp.linspace(0, 2 * xp.pi, 4, endpoint=False)
+            assert_array_almost_equal(w, expected_w)
+            assert_array_almost_equal(h, h_whole)
+
+            # simultaneously check int-like support
+            w, h = freqz(b, a, worN=np.int32(4), whole=True)
+            assert_array_almost_equal(w, expected_w)
+            assert_array_almost_equal(h, h_whole)
+
+            w, h = freqz(b, a, worN=w, whole=True)
+            assert_array_almost_equal(w, expected_w)
+            assert_array_almost_equal(h, h_whole)
+
+    def test_basic3(self):
+        t = np.linspace(0, 1, 4, endpoint=False)
+        expected_w = np.linspace(0, 2 * np.pi, 4, endpoint=False)
+        for b, a, h_whole in zip(
+                (np.asarray([1., 0, 0, 0]), np.sin(2 * np.pi * t)),
+                (np.asarray([1., 0, 0, 0]), np.asarray([0.5, 0, 0, 0])),
+                (np.asarray([1., 1., 1., 1.]), np.asarray([0, -4j, 0, 4j]))
+        ):
+
+            w, h = freqz(b, a, worN=np.int32(4), whole=True)
+            assert_array_almost_equal(w, expected_w)
+            assert_array_almost_equal(h, h_whole)
+
+            w, h = freqz(b, a, worN=w, whole=True)
+            assert_array_almost_equal(w, expected_w)
+            assert_array_almost_equal(h, h_whole)
+
+    def test_basic_whole(self, xp):
+        w, h = freqz(xp.asarray([1.0]), worN=8, whole=True)
+        assert_array_almost_equal(w, 2 * xp.pi * xp.arange(8.0) / 8)
+        assert_array_almost_equal(h, xp.ones(8))
+
+    def test_plot(self, xp):
+
+        def plot(w, h):
+            assert_array_almost_equal(w, xp.pi * xp.arange(8.0) / 8)
+            assert_array_almost_equal(h, xp.ones(8))
+
+        with assert_raises(ZeroDivisionError):
+            freqz(xp.asarray([1.0]), worN=8, plot=lambda w, h: 1 / 0)
+
+        freqz(xp.asarray([1.0]), worN=8, plot=plot)
+
+    def test_fft_wrapping(self, xp):
+        # Some simple real FIR filters
+        bs = list()  # filters
+        as_ = list()
+        hs_whole = list()
+        hs_half = list()
+        # 3 taps
+        t = xp.linspace(0, 1, 3, endpoint=False)
+        bs.append(xp.sin(2 * xp.pi * t))
+        as_.append(3.)
+        hs_whole.append(xp.asarray([0, -0.5j, 0.5j]))
+        hs_half.append(xp.asarray([0, math.sqrt(1./12.), -0.5j]))
+        # 4 taps
+        t = xp.linspace(0, 1, 4, endpoint=False)
+        bs.append(xp.sin(2 * xp.pi * t))
+        as_.append(0.5)
+        hs_whole.append(xp.asarray([0, -4j, 0, 4j]))
+        hs_half.append(xp.asarray([0, math.sqrt(8), -4j, -math.sqrt(8)]))
+        del t
+        for ii, b in enumerate(bs):
+            # whole
+            a = as_[ii]
+            expected_w = xp.linspace(0, 2 * xp.pi, b.shape[0], endpoint=False)
+            w, h = freqz(b, a, worN=expected_w, whole=True)  # polyval
+            err_msg = f'b = {b}, a={a}'
+            assert_array_almost_equal(w, expected_w, err_msg=err_msg)
+            assert_array_almost_equal(h, hs_whole[ii], err_msg=err_msg)
+
+            w, h = freqz(b, a, worN=b.shape[0], whole=True)  # FFT
+            assert_array_almost_equal(w, expected_w, err_msg=err_msg)
+            assert_array_almost_equal(h, hs_whole[ii], err_msg=err_msg)
+
+            # non-whole
+            expected_w = xp.linspace(0, xp.pi, b.shape[0], endpoint=False)
+            w, h = freqz(b, a, worN=expected_w, whole=False)  # polyval
+            assert_array_almost_equal(w, expected_w, err_msg=err_msg)
+            assert_array_almost_equal(h, hs_half[ii], err_msg=err_msg)
+
+            w, h = freqz(b, a, worN=b.shape[0], whole=False)  # FFT
+            assert_array_almost_equal(w, expected_w, err_msg=err_msg)
+            assert_array_almost_equal(h, hs_half[ii], err_msg=err_msg)
+
+        # some random FIR filters (real + complex)
+        # assume polyval is accurate
+        rng = np.random.RandomState(0)
+        for ii in range(2, 10):  # number of taps
+            b = xp.asarray(rng.randn(ii))
+            for kk in range(2):
+                a = xp.asarray(rng.randn(1) if kk == 0 else rng.randn(3))
+                for jj in range(2):
+                    if jj == 1:
+                        b = b + xp.asarray(rng.randn(ii)) * 1j
+
+                    # whole
+                    expected_w = xp.linspace(0, 2 * xp.pi, ii, endpoint=False)
+                    w, expected_h = freqz(b, a, worN=expected_w, whole=True)
+                    assert_array_almost_equal(w, expected_w)
+                    w, h = freqz(b, a, worN=ii, whole=True)
+                    assert_array_almost_equal(w, expected_w)
+                    assert_array_almost_equal(h, expected_h, decimal=4)
+
+                    # half
+                    expected_w = xp.linspace(0, xp.pi, ii, endpoint=False)
+                    w, expected_h = freqz(b, a, worN=expected_w, whole=False)
+                    assert_array_almost_equal(w, expected_w)
+                    w, h = freqz(b, a, worN=ii, whole=False)
+                    assert_array_almost_equal(w, expected_w)
+                    assert_array_almost_equal(h, expected_h, decimal=4)
+
+    def test_broadcasting1(self, xp):
+        # Test broadcasting with worN an integer or a 1-D array,
+        # b and a are n-dimensional arrays.
+        np.random.seed(123)
+        b = np.random.rand(3, 5, 1)
+        a = np.random.rand(2, 1)
+        b, a = map(xp.asarray, (b, a))
+
+        for whole in [False, True]:
+            # Test with worN being integers (one fast for FFT and one not),
+            # a 1-D array, and an empty array.
+            for worN in [16, 17, xp.linspace(0, 1, 10), xp.asarray([])]:
+                w, h = freqz(b, a, worN=worN, whole=whole)
+                for k in range(b.shape[1]):
+                    bk = b[:, k, 0]
+                    ak = a[:, 0]
+                    ww, hh = freqz(bk, ak, worN=worN, whole=whole)
+                    xp_assert_close(ww, w)
+                    xp_assert_close(hh, h[k, ...])
+
+    def test_broadcasting2(self, xp):
+        # Test broadcasting with worN an integer or a 1-D array,
+        # b is an n-dimensional array, and a is left at the default value.
+        np.random.seed(123)
+        b = np.random.rand(3, 5, 1)
+        b = xp.asarray(b)
+        for whole in [False, True]:
+            for worN in [16, 17, xp.linspace(0, 1, 10)]:
+                w, h = freqz(b, worN=worN, whole=whole)
+                for k in range(b.shape[1]):
+                    bk = b[:, k, 0]
+                    ww, hh = freqz(bk, worN=worN, whole=whole)
+                    xp_assert_close(ww, w)
+                    xp_assert_close(hh, h[k, :])
+
+    def test_broadcasting3(self, xp):
+        # Test broadcasting where b.shape[-1] is the same length
+        # as worN, and a is left at the default value.
+        np.random.seed(123)
+        N = 16
+        b = np.random.rand(3, N)
+        b = xp.asarray(b)
+        for whole in [False, True]:
+            for worN in [N, xp.linspace(0, 1, N)]:
+                w, h = freqz(b, worN=worN, whole=whole)
+                assert xp_size(w) == N
+                for k in range(N):
+                    bk = b[:, k]
+                    ww, hh = freqz(bk, worN=w[k], whole=whole)
+                    xp_assert_close(ww, xp.asarray(w[k])[None])
+                    xp_assert_close(hh, xp.asarray(h[k])[None])
+
+    def test_broadcasting4(self, xp):
+        # Test broadcasting with worN a 2-D array.
+        np.random.seed(123)
+        b = np.random.rand(4, 2, 1, 1)
+        a = np.random.rand(5, 2, 1, 1)
+        b, a = map(xp.asarray, (b, a))
+
+        for whole in [False, True]:
+            for worN in [np.random.rand(6, 7), np.empty((6, 0))]:
+                worN = xp.asarray(worN)
+                w, h = freqz(b, a, worN=worN, whole=whole)
+                xp_assert_close(w, worN, rtol=1e-14)
+                assert h.shape == (2,) + worN.shape
+                for k in range(2):
+                    ww, hh = freqz(b[:, k, 0, 0], a[:, k, 0, 0],
+                                   worN=xp.reshape(worN, (-1,)),
+                                   whole=whole)
+                    xp_assert_close(ww, xp.reshape(worN, (-1,)), rtol=1e-14)
+                    xp_assert_close(hh, xp.reshape(h[k, :, :], (-1,)))
+
+    def test_backward_compat(self):
+        # For backward compatibility, test if None act as a wrapper for default
+        w1, h1 = freqz([1.0], 1)
+        w2, h2 = freqz([1.0], 1, None)
+        assert_array_almost_equal(w1, w2)
+        assert_array_almost_equal(h1, h2)
+
+    def test_fs_param(self, xp):
+        fs = 900
+        b = xp.asarray([0.039479155677484369, 0.11843746703245311, 0.11843746703245311,
+                        0.039479155677484369])
+        a = xp.asarray([1.0, -1.3199152021838287, 0.80341991081938424,
+                        -0.16767146321568049])
+
+        # N = None, whole=False
+        w1, h1 = freqz(b, a, fs=fs)
+        w2, h2 = freqz(b, a)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs/2, 512, endpoint=False))
+
+        # N = None, whole=True
+        w1, h1 = freqz(b, a, whole=True, fs=fs)
+        w2, h2 = freqz(b, a, whole=True)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs, 512, endpoint=False))
+
+        # N = 5, whole=False
+        w1, h1 = freqz(b, a, 5, fs=fs)
+        w2, h2 = freqz(b, a, 5)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs/2, 5, endpoint=False))
+
+        # N = 5, whole=True
+        w1, h1 = freqz(b, a, 5, whole=True, fs=fs)
+        w2, h2 = freqz(b, a, 5, whole=True)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs, 5, endpoint=False))
+
+        # w is an array_like
+        for w in ([123], (123,), xp.asarray([123]), (50, 123, 230),
+                  xp.asarray([50, 123, 230])):
+            w1, h1 = freqz(b, a, w, fs=fs)
+            w2, h2 = freqz(b, a, 2*pi*xp.asarray(w, dtype=xp.float64)/ fs)
+            xp_assert_close(h1, h2)
+            xp_assert_close(w1, xp.asarray(w), check_dtype=False)
+
+    def test_w_or_N_types(self):
+        # Measure at 7 (polyval) or 8 (fft) equally-spaced points
+        for N in (7, np.int8(7), np.int16(7), np.int32(7), np.int64(7),
+                  np.array(7),
+                  8, np.int8(8), np.int16(8), np.int32(8), np.int64(8),
+                  np.array(8)):
+
+            w, h = freqz([1.0], worN=N)
+            assert_array_almost_equal(w, np.pi * np.arange(N) / N)
+            assert_array_almost_equal(h, np.ones(N))
+
+            w, h = freqz([1.0], worN=N, fs=100)
+            assert_array_almost_equal(w, np.linspace(0, 50, N, endpoint=False))
+            assert_array_almost_equal(h, np.ones(N))
+
+        # Measure at frequency 8 Hz
+        for w in (8.0, 8.0+0j):
+            # Only makes sense when fs is specified
+            w_out, h = freqz(np.asarray([1.0]), worN=w, fs=100)
+            assert_array_almost_equal(w_out, np.asarray([8]))
+            assert_array_almost_equal(h, np.asarray(1.), check_0d=False)
+
+    def test_nyquist(self, xp):
+        w, h = freqz(xp.asarray([1.0]), worN=8, include_nyquist=True)
+        assert_array_almost_equal(w, xp.pi * xp.arange(8, dtype=w.dtype) / 7.)
+        assert_array_almost_equal(h, xp.ones(8))
+        w, h = freqz(xp.asarray([1.0]), worN=9, include_nyquist=True)
+        assert_array_almost_equal(w, xp.pi * xp.arange(9, dtype=w.dtype) / 8.)
+        assert_array_almost_equal(h, xp.ones(9))
+
+        for a in [1, xp.ones(2)]:
+            w, h = freqz(xp.ones(2), a, worN=0, include_nyquist=True)
+            assert w.shape == (0,)
+            assert h.shape == (0,)
+            hdt = xp.complex128 if xp_default_dtype(xp) == xp.float64 else xp.complex64
+            assert h.dtype == hdt
+
+        w1, h1 = freqz(xp.asarray([1.0]), worN=8, whole = True, include_nyquist=True)
+        w2, h2 = freqz(xp.asarray([1.0]), worN=8, whole = True, include_nyquist=False)
+        assert_array_almost_equal(w1, w2)
+        assert_array_almost_equal(h1, h2)
+
+    # https://github.com/scipy/scipy/issues/17289
+    # https://github.com/scipy/scipy/issues/15273
+    @pytest.mark.parametrize('whole,nyquist,worN',
+                             [(False, False, 32),
+                              (False, True, 32),
+                              (True, False, 32),
+                              (True, True, 32),
+                              (False, False, 257),
+                              (False, True, 257),
+                              (True, False, 257),
+                              (True, True, 257)])
+    def test_17289(self, whole, nyquist, worN, xp):
+        d = xp.asarray([0.0, 1.0])
+        w, Drfft = freqz(d, worN=32, whole=whole, include_nyquist=nyquist)
+        _, Dpoly = freqz(d, worN=w)
+        xp_assert_close(Drfft, Dpoly)
+
+    def test_fs_validation(self):
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            freqz([1.0], fs=np.array([10, 20]))
+
+        with pytest.raises(ValueError, match="Sampling.*be none."):
+            freqz([1.0], fs=None)
+
+
+class TestFreqz_sos:
+
+    def test_freqz_sos_basic(self, xp):
+        # Compare the results of freqz and freqz_sos for a low order
+        # Butterworth filter.
+
+        N = 500
+
+        b, a = butter(4, 0.2)
+        sos = butter(4, 0.2, output='sos')
+        b, a, sos = map(xp.asarray, (b, a, sos))   # XXX until butter is converted
+
+        w, h = freqz(b, a, worN=N)
+        w2, h2 = freqz_sos(sos, worN=N)
+        xp_assert_equal(w2, w)
+        xp_assert_close(h2, h, rtol=1e-10, atol=1e-14)
+
+        b, a = ellip(3, 1, 30, (0.2, 0.3), btype='bandpass')
+        sos = ellip(3, 1, 30, (0.2, 0.3), btype='bandpass', output='sos')
+        b, a, sos = map(xp.asarray, (b, a, sos))   # XXX until ellip is converted
+
+        w, h = freqz(b, a, worN=N)
+        w2, h2 = freqz_sos(sos, worN=N)
+        xp_assert_equal(w2, w)
+        xp_assert_close(h2, h, rtol=1e-10, atol=1e-14)
+
+        # must have at least one section
+        with assert_raises(ValueError):
+            freqz_sos(sos[:0, ...])
+
+    def test_backward_compat(self, xp):
+        # For backward compatibility, test if None act as a wrapper for default
+        N = 500
+
+        sos = butter(4, 0.2, output='sos')
+        sos = xp.asarray(sos)   # XXX until butter is converted
+        w1, h1 = freqz_sos(sos, worN=N)
+        w2, h2 = sosfreqz(sos, worN=N)
+        assert_array_almost_equal(w1, w2)
+        assert_array_almost_equal(h1, h2)
+
+    @skip_xp_backends("dask.array", reason="float cannot be interpreted as in integer")
+    def test_freqz_sos_design(self, xp):
+        # Compare freqz_sos output against expected values for different
+        # filter types
+
+        # from cheb2ord
+        N, Wn = cheb2ord([0.1, 0.6], [0.2, 0.5], 3, 60)
+        sos = cheby2(N, 60, Wn, 'stop', output='sos')
+        sos = xp.asarray(sos)  # XXX
+        zero = xp.asarray(0., dtype=xp.float64)
+
+        w, h = freqz_sos(sos)
+        h = xp.abs(h)
+        w = w / xp.pi
+        xp_assert_close(20 * xp.log10(h[w <= 0.1]),
+                        zero, atol=3.01,
+                        check_shape=False)
+        xp_assert_close(20 * xp.log10(h[w >= 0.6]),
+                        zero, atol=3.01,
+                        check_shape=False)
+        xp_assert_close(h[(w >= 0.2) & (w <= 0.5)],
+                        zero, atol=1e-3,
+                        check_shape=False)  # <= -60 dB
+
+        N, Wn = cheb2ord([0.1, 0.6], [0.2, 0.5], 3, 150)
+        sos = cheby2(N, 150, Wn, 'stop', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        dB = 20*xp.log10(xp.abs(h))
+        w = w / xp.pi
+        xp_assert_close(dB[w <= 0.1], zero, atol=3.01, check_shape=False)
+        xp_assert_close(dB[w >= 0.6], zero, atol=3.01, check_shape=False)
+        assert xp.all(dB[(w >= 0.2) & (w <= 0.5)] < -149.9)
+
+        # from cheb1ord
+        N, Wn = cheb1ord(0.2, 0.3, 3, 40)
+        sos = cheby1(N, 3, Wn, 'low', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        h = xp.abs(h)
+        w = w / xp.pi
+        xp_assert_close(20 * xp.log10(h[w <= 0.2]), zero, atol=3.01,
+                        check_shape=False)
+        xp_assert_close(h[w >= 0.3], zero, atol=1e-2,
+                        check_shape=False)  # <= -40 dB
+
+        N, Wn = cheb1ord(0.2, 0.3, 1, 150)
+        sos = cheby1(N, 1, Wn, 'low', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        dB = 20*xp.log10(xp.abs(h))
+        w /= np.pi
+        xp_assert_close(dB[w <= 0.2], zero, atol=1.01, check_shape=False)
+        assert xp.all(dB[w >= 0.3] < -149.9)
+
+        # adapted from ellipord
+        N, Wn = ellipord(0.3, 0.2, 3, 60)
+        sos = ellip(N, 0.3, 60, Wn, 'high', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        h = xp.abs(h)
+        w = w / xp.pi
+        xp_assert_close(20 * xp.log10(h[w >= 0.3]), zero, atol=3.01,
+                        check_shape=False)
+        xp_assert_close(h[w <= 0.1], zero, atol=1.5e-3,
+                        check_shape=False)  # <= -60 dB (approx)
+
+        # adapted from buttord
+        N, Wn = buttord([0.2, 0.5], [0.14, 0.6], 3, 40)
+        sos = butter(N, Wn, 'band', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        h = xp.abs(h)
+        w = w / xp.pi
+
+        h014 = h[w <= 0.14]
+        xp_assert_close(h014, xp.zeros_like(h014), atol=1e-2)  # <= -40 dB
+        h06 = h[w >= 0.6]
+        xp_assert_close(h06, xp.zeros_like(h06), atol=1e-2)  # <= -40 dB
+        h0205 = 20 * xp.log10(h[(w >= 0.2) & (w <= 0.5)])
+        xp_assert_close(h0205, xp.zeros_like(h0205), atol=3.01)
+
+        N, Wn = buttord([0.2, 0.5], [0.14, 0.6], 3, 100)
+        sos = butter(N, Wn, 'band', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        dB = 20*xp.log10(xp.maximum(xp.abs(h), xp.asarray(1e-10)))
+        w = w / xp.pi
+
+        assert xp.all(dB[(w > 0) & (w <= 0.14)] < -99.9)
+        assert xp.all(dB[w >= 0.6] < -99.9)
+        db0205 = dB[(w >= 0.2) & (w <= 0.5)]
+        xp_assert_close(db0205, xp.zeros_like(db0205), atol=3.01)
+
+    def test_freqz_sos_design_ellip(self, xp):
+        N, Wn = ellipord(0.3, 0.1, 3, 60)
+        sos = ellip(N, 0.3, 60, Wn, 'high', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        h = xp.abs(h)
+        w = w / xp.pi
+
+        h03 = 20 * xp.log10(h[w >= 0.3])
+        xp_assert_close(h03, xp.zeros_like(h03), atol=3.01)
+        h01 = h[w <= 0.1]
+        xp_assert_close(h01, xp.zeros_like(h01), atol=1.5e-3)  # <= -60 dB (approx)
+
+        N, Wn = ellipord(0.3, 0.2, .5, 150)
+        sos = ellip(N, .5, 150, Wn, 'high', output='sos')
+        sos = xp.asarray(sos)
+
+        w, h = freqz_sos(sos)
+        dB = 20*xp.log10(xp.maximum(xp.abs(h), xp.asarray(1e-10)))
+        w = w / xp.pi
+
+        db03 = dB[w >= 0.3]
+        xp_assert_close(db03, xp.zeros_like(db03), atol=.55)
+        # Allow some numerical slop in the upper bound -150, so this is
+        # a check that dB[w <= 0.2] is less than or almost equal to -150.
+        assert xp.max(dB[w <= 0.2]) < -150*(1 - 1e-12)
+
+    @mpmath_check("0.10")
+    def test_freqz_sos_against_mp(self, xp):
+        # Compare the result of freqz_sos applied to a high order Butterworth
+        # filter against the result computed using mpmath.  (signal.freqz fails
+        # miserably with such high order filters.)
+        from . import mpsig
+        N = 500
+        order = 25
+        Wn = 0.15
+        with mpmath.workdps(80):
+            z_mp, p_mp, k_mp = mpsig.butter_lp(order, Wn)
+            w_mp, h_mp = mpsig.zpkfreqz(z_mp, p_mp, k_mp, N)
+        w_mp = xp.asarray([float(x) for x in w_mp], dtype=xp.float64)
+        h_mp = xp.asarray([complex(x) for x in h_mp], dtype=xp.complex128)
+
+        sos = butter(order, Wn, output='sos')
+        sos = xp.asarray(sos, dtype=xp.float64)
+        w, h = freqz_sos(sos, worN=N)
+        xp_assert_close(w, w_mp, rtol=1e-12, atol=1e-14)
+        xp_assert_close(h, h_mp, rtol=1e-12, atol=1e-14)
+
+    def test_fs_param(self, xp):
+        fs = 900
+        sos = xp.asarray(
+              [[0.03934683014103762, 0.07869366028207524, 0.03934683014103762,
+                1.0, -0.37256600288916636, 0.0],
+               [1.0, 1.0, 0.0, 1.0, -0.9495739996946778, 0.45125966317124144]]
+        )
+
+        # N = None, whole=False
+        w1, h1 = freqz_sos(sos, fs=fs)
+        w2, h2 = freqz_sos(sos)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs/2, 512, endpoint=False))
+
+        # N = None, whole=True
+        w1, h1 = freqz_sos(sos, whole=True, fs=fs)
+        w2, h2 = freqz_sos(sos, whole=True)
+        xp_assert_close(h1, h2, atol=1e-27)
+        xp_assert_close(w1, xp.linspace(0, fs, 512, endpoint=False))
+
+        # N = 5, whole=False
+        w1, h1 = freqz_sos(sos, 5, fs=fs)
+        w2, h2 = freqz_sos(sos, 5)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs/2, 5, endpoint=False))
+
+        # N = 5, whole=True
+        w1, h1 = freqz_sos(sos, 5, whole=True, fs=fs)
+        w2, h2 = freqz_sos(sos, 5, whole=True)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs, 5, endpoint=False))
+
+    @skip_xp_backends(np_only=True, reason="array-likes")
+    def test_fs_param2(self, xp):
+        fs = 900
+        sos = xp.asarray(
+              [[0.03934683014103762, 0.07869366028207524, 0.03934683014103762,
+                1.0, -0.37256600288916636, 0.0],
+               [1.0, 1.0, 0.0, 1.0, -0.9495739996946778, 0.45125966317124144]]
+        )
+
+        # w is an array_like
+        for w in ([123], (123,), xp.asarray([123]), (50, 123, 230),
+                  xp.asarray([50, 123, 230])):
+            w1, h1 = freqz_sos(sos, w, fs=fs)
+            w1, h1 = map(xp.asarray, (w1, h1))
+
+            w2, h2 = freqz_sos(sos, 2*pi*xp.asarray(w, dtype=sos.dtype)/fs)
+            xp_assert_close(h1, h2)
+            xp_assert_close(w, w1, check_dtype=False)
+
+    def test_w_or_N_types(self):
+        # Measure at 7 (polyval) or 8 (fft) equally-spaced points
+        for N in (7, np.int8(7), np.int16(7), np.int32(7), np.int64(7),
+                  np.array(7),
+                  8, np.int8(8), np.int16(8), np.int32(8), np.int64(8),
+                  np.array(8)):
+
+            w, h = freqz_sos([1, 0, 0, 1, 0, 0], worN=N)
+            assert_array_almost_equal(w, np.pi * np.arange(N) / N)
+            assert_array_almost_equal(h, np.ones(N))
+
+            w, h = freqz_sos([1, 0, 0, 1, 0, 0], worN=N, fs=100)
+            assert_array_almost_equal(w, np.linspace(0, 50, N, endpoint=False))
+            assert_array_almost_equal(h, np.ones(N))
+
+        # Measure at frequency 8 Hz
+        for w in (8.0, 8.0+0j):
+            # Only makes sense when fs is specified
+            w_out, h = freqz_sos([1, 0, 0, 1, 0, 0], worN=w, fs=100)
+            assert_array_almost_equal(w_out, [8])
+            assert_array_almost_equal(h, [1])
+
+    def test_fs_validation(self):
+        sos = butter(4, 0.2, output='sos')
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            freqz_sos(sos, fs=np.array([10, 20]))
+
+
+class TestFreqz_zpk:
+
+    def test_ticket1441(self, xp):
+        """Regression test for ticket 1441."""
+        # Because freqz previously used arange instead of linspace,
+        # when N was large, it would return one more point than
+        # requested.
+        N = 100000
+        w, h = freqz_zpk(xp.asarray([0.5]), xp.asarray([0.5]), 1.0, worN=N)
+        assert w.shape == (N,)
+
+    def test_basic(self, xp):
+        w, h = freqz_zpk(xp.asarray([0.5]), xp.asarray([0.5]), 1.0, worN=8)
+        assert_array_almost_equal(w, xp.pi * xp.arange(8.0) / 8)
+        assert_array_almost_equal(h, xp.ones(8))
+
+    def test_basic_whole(self, xp):
+        w, h = freqz_zpk(xp.asarray([0.5]), xp.asarray([0.5]), 1.0, worN=8, whole=True)
+        assert_array_almost_equal(w, 2 * xp.pi * xp.arange(8.0) / 8)
+        assert_array_almost_equal(h, xp.ones(8))
+
+    def test_vs_freqz(self, xp):
+        b, a = cheby1(4, 5, 0.5, analog=False, output='ba')
+        z, p, k = cheby1(4, 5, 0.5, analog=False, output='zpk')
+
+        b, a = map(xp.asarray, (b, a))  # XXX convert cheby1
+        z, p = map(xp.asarray, (z, p))
+
+        w1, h1 = freqz(b, a)
+        w2, h2 = freqz_zpk(z, p, k)
+        xp_assert_close(w1, w2)
+        xp_assert_close(h1, h2, rtol=1e-6)
+
+    def test_backward_compat(self, xp):
+        # For backward compatibility, test if None act as a wrapper for default
+        w1, h1 = freqz_zpk(xp.asarray([0.5]), xp.asarray([0.5]), 1.0)
+        w2, h2 = freqz_zpk(xp.asarray([0.5]), xp.asarray([0.5]), 1.0, None)
+        assert_array_almost_equal(w1, w2)
+        assert_array_almost_equal(h1, h2)
+
+    def test_fs_param(self, xp):
+        fs = 900
+        z = xp.asarray([-1, -1, -1.0])
+        p = xp.asarray(
+            [0.4747869998473389 + 0.4752230717749344j,
+             0.37256600288916636,
+             0.4747869998473389 - 0.4752230717749344j]
+        )
+        k = 0.03934683014103762
+
+        # N = None, whole=False
+        w1, h1 = freqz_zpk(z, p, k, whole=False, fs=fs)
+        w2, h2 = freqz_zpk(z, p, k, whole=False)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs/2, 512, endpoint=False))
+
+        # N = None, whole=True
+        w1, h1 = freqz_zpk(z, p, k, whole=True, fs=fs)
+        w2, h2 = freqz_zpk(z, p, k, whole=True)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs, 512, endpoint=False))
+
+        # N = 5, whole=False
+        w1, h1 = freqz_zpk(z, p, k, 5, fs=fs)
+        w2, h2 = freqz_zpk(z, p, k, 5)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs/2, 5, endpoint=False))
+
+        # N = 5, whole=True
+        w1, h1 = freqz_zpk(z, p, k, 5, whole=True, fs=fs)
+        w2, h2 = freqz_zpk(z, p, k, 5, whole=True)
+        xp_assert_close(h1, h2)
+        xp_assert_close(w1, xp.linspace(0, fs, 5, endpoint=False))
+
+    @skip_xp_backends(np_only=True, reason="array_likes")
+    def test_fs_param2(self, xp):
+        fs = 900
+        z = xp.asarray([-1, -1, -1.0])
+        p = xp.asarray(
+            [0.4747869998473389 + 0.4752230717749344j,
+             0.37256600288916636,
+             0.4747869998473389 - 0.4752230717749344j]
+        )
+        k = 0.03934683014103762
+
+        # w is an array_like
+        for w in ([123], (123,), xp.asarray([123]), (50, 123, 230),
+                  xp.asarray([50, 123, 230])):
+            w1, h1 = freqz_zpk(z, p, k, w, fs=fs)
+            w2, h2 = freqz_zpk(z, p, k, 2*pi*xp.asarray(w)/fs)
+            xp_assert_close(h1, h2)
+            xp_assert_close(w, w1, check_dtype=False)
+
+    def test_w_or_N_types(self):
+        # Measure at 8 equally-spaced points
+        for N in (8, np.int8(8), np.int16(8), np.int32(8), np.int64(8),
+                  np.array(8)):
+
+            w, h = freqz_zpk([], [], 1, worN=N)
+            assert_array_almost_equal(w, np.pi * np.arange(8) / 8.)
+            assert_array_almost_equal(h, np.ones(8))
+
+            w, h = freqz_zpk([], [], 1, worN=N, fs=100)
+            assert_array_almost_equal(w, np.linspace(0, 50, 8, endpoint=False))
+            assert_array_almost_equal(h, np.ones(8))
+
+        # Measure at frequency 8 Hz
+        for w in (8.0, 8.0+0j):
+            # Only makes sense when fs is specified
+            w_out, h = freqz_zpk([], [], 1, worN=w, fs=100)
+            assert_array_almost_equal(w_out, [8])
+            assert_array_almost_equal(h, [1])
+
+    def test_fs_validation(self):
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            freqz_zpk([1.0], [1.0], [1.0], fs=np.array([10., 20]))
+
+        with pytest.raises(ValueError, match="Sampling.*be none."):
+            freqz_zpk([1.0], [1.0], [1.0], fs=None)
+
+
+class TestNormalize:
+
+    def test_allclose(self):
+        """Test for false positive on allclose in normalize() in
+        filter_design.py"""
+        # Test to make sure the allclose call within signal.normalize does not
+        # choose false positives. Then check against a known output from MATLAB
+        # to make sure the fix doesn't break anything.
+
+        # These are the coefficients returned from
+        #   `[b,a] = cheby1(8, 0.5, 0.048)'
+        # in MATLAB. There are at least 15 significant figures in each
+        # coefficient, so it makes sense to test for errors on the order of
+        # 1e-13 (this can always be relaxed if different platforms have
+        # different rounding errors)
+        b_matlab = np.array([2.150733144728282e-11, 1.720586515782626e-10,
+                             6.022052805239190e-10, 1.204410561047838e-09,
+                             1.505513201309798e-09, 1.204410561047838e-09,
+                             6.022052805239190e-10, 1.720586515782626e-10,
+                             2.150733144728282e-11])
+        a_matlab = np.array([1.000000000000000e+00, -7.782402035027959e+00,
+                             2.654354569747454e+01, -5.182182531666387e+01,
+                             6.334127355102684e+01, -4.963358186631157e+01,
+                             2.434862182949389e+01, -6.836925348604676e+00,
+                             8.412934944449140e-01])
+
+        # This is the input to signal.normalize after passing through the
+        # equivalent steps in signal.iirfilter as was done for MATLAB
+        b_norm_in = np.array([1.5543135865293012e-06, 1.2434508692234413e-05,
+                              4.3520780422820447e-05, 8.7041560845640893e-05,
+                              1.0880195105705122e-04, 8.7041560845640975e-05,
+                              4.3520780422820447e-05, 1.2434508692234413e-05,
+                              1.5543135865293012e-06])
+        a_norm_in = np.array([7.2269025909127173e+04, -5.6242661430467968e+05,
+                              1.9182761917308895e+06, -3.7451128364682454e+06,
+                              4.5776121393762771e+06, -3.5869706138592605e+06,
+                              1.7596511818472347e+06, -4.9409793515707983e+05,
+                              6.0799461347219651e+04])
+
+        b_output, a_output = normalize(b_norm_in, a_norm_in)
+
+        # The test on b works for decimal=14 but the one for a does not. For
+        # the sake of consistency, both of these are decimal=13. If something
+        # breaks on another platform, it is probably fine to relax this lower.
+        assert_array_almost_equal(b_matlab, b_output, decimal=13)
+        assert_array_almost_equal(a_matlab, a_output, decimal=13)
+
+    def test_errors(self):
+        """Test the error cases."""
+        # all zero denominator
+        assert_raises(ValueError, normalize, [1, 2], 0)
+
+        # denominator not 1 dimensional
+        assert_raises(ValueError, normalize, [1, 2], [[1]])
+
+        # numerator too many dimensions
+        assert_raises(ValueError, normalize, [[[1, 2]]], 1)
+
+
+class TestLp2lp:
+
+    def test_basic(self, xp):
+        b = xp.asarray([1])
+        a = xp.asarray([1, math.sqrt(2), 1])
+        b_lp, a_lp = lp2lp(b, a, 0.38574256627112119)
+        assert_array_almost_equal(b_lp, xp.asarray([0.1488]), decimal=4)
+        assert_array_almost_equal(a_lp, xp.asarray([1, 0.5455, 0.1488]), decimal=4)
+
+
+class TestLp2hp:
+
+    @skip_xp_backends(eager_only=True, reason="in-place item assignment")
+    def test_basic(self, xp):
+        b = xp.asarray([0.25059432325190018])
+        a = xp.asarray(
+            [1, 0.59724041654134863, 0.92834805757524175, 0.25059432325190018]
+        )
+        b_hp, a_hp = lp2hp(b, a, 2*math.pi*5000)
+        xp_assert_close(b_hp, xp.asarray([1.0, 0, 0, 0]))
+        xp_assert_close(
+            a_hp, xp.asarray([1, 1.1638e5, 2.3522e9, 1.2373e14]), rtol=1e-4
+        )
+
+
+class TestLp2bp:
+
+    @skip_xp_backends(eager_only=True, reason="in-place item assignment")
+    def test_basic(self, xp):
+        b = xp.asarray([1])
+        a = xp.asarray([1, 2, 2, 1])
+        b_bp, a_bp = lp2bp(b, a, 2*math.pi*4000, 2*math.pi*2000)
+        xp_assert_close(b_bp, xp.asarray([1.9844e12, 0, 0, 0]), rtol=1e-6)
+        xp_assert_close(
+            a_bp,
+            xp.asarray([1, 2.5133e4, 2.2108e9, 3.3735e13,
+                        1.3965e18, 1.0028e22, 2.5202e26]), rtol=1e-4
+        )
+
+
+class TestLp2bs:
+
+    @skip_xp_backends(eager_only=True, reason="in-place item assignment")
+    def test_basic(self, xp):
+        b = xp.asarray([1])
+        a = xp.asarray([1, 1])
+        b_bs, a_bs = lp2bs(b, a, 0.41722257286366754, 0.18460575326152251)
+        assert_array_almost_equal(b_bs, xp.asarray([1, 0, 0.17407]), decimal=5)
+        assert_array_almost_equal(a_bs, xp.asarray([1, 0.18461, 0.17407]), decimal=5)
+
+
+class TestBilinear:
+    """Tests for function `signal.bilinear`. """
+
+    def test_exceptions(self):
+        """Raise all exceptions in `bilinear()`. """
+        with pytest.raises(ValueError, match="Parameter a is not .*"):
+            bilinear(1., np.array([[1, 2, 3]]))
+        with pytest.raises(ValueError, match="Parameter b is not .*"):
+            bilinear(np.ones((2,3)), 1. )
+
+    def test_basic(self):
+        # reference output values computed with sympy
+        b = [0.14879732743343033]
+        a = [1, 0.54552236880522209, 0.14879732743343033]
+        b_zref = [0.08782128175913713, 0.17564256351827426, 0.08782128175913713]
+        a_zref = [1.0, -1.0047722097030667, 0.3560573367396151]
+
+        b_z, a_z = bilinear(b, a, 0.5)
+
+        assert_array_almost_equal_nulp(b_z, b_zref)
+        assert_array_almost_equal_nulp(a_z, a_zref)
+
+        b = [1, 0, 0.17407467530697837]
+        a = [1, 0.18460575326152251, 0.17407467530697837]
+        b_zref = [0.8641286432189045, -1.2157757001964216, 0.8641286432189045]
+        a_zref = [1.0, -1.2157757001964216, 0.7282572864378091]
+
+        b_z, a_z = bilinear(b, a, 0.5)
+
+        assert_array_almost_equal_nulp(b_z, b_zref)
+        assert_array_almost_equal_nulp(a_z, a_zref)
+
+
+    def test_ignore_leading_zeros(self):
+        # regression for gh-6606
+        # results shouldn't change when leading zeros are added to
+        # input numerator or denominator
+        b = [0.14879732743343033]
+        a = [1, 0.54552236880522209, 0.14879732743343033]
+
+        b_zref = [0.08782128175913713, 0.17564256351827426, 0.08782128175913713]
+        a_zref = [1.0, -1.0047722097030667, 0.3560573367396151]
+
+        for lzn, lzd in product(range(4), range(4)):
+            b_z, a_z = bilinear(np.pad(b, (lzn, 0)),
+                                np.pad(a, (lzd, 0)),
+                                0.5)
+            assert_array_almost_equal_nulp(b_z, b_zref)
+            assert_array_almost_equal_nulp(a_z, a_zref)
+
+
+    def test_complex(self):
+        # reference output values computed with sympy
+        # this is an elliptical filter, 5Hz width, centered at +50Hz:
+        #     z, p, k = signal.ellip(2, 0.5, 20, 2*np.pi*5/2, output='zpk', analog=True)
+        #     z = z.astype(complex) + 2j * np.pi * 50
+        #     p = p.astype(complex) + 2j * np.pi * 50
+        #     b, a = signal.zpk2tf(z, p, k)
+        b = [(0.09999999999999991+0j),
+             -62.831853071795805j,
+             (-9505.857007071314+0j)]
+        a = [(1+0j),
+             (21.09511000343942-628.3185307179587j),
+             (-98310.74322875646-6627.2242613473845j)]
+        # sample frequency
+        fs = 1000
+        b_zref = [(0.09905575106715676-0.00013441423112828688j),
+                  (-0.18834281923181084-0.06032810039049478j),
+                  (0.08054306669414343+0.05766172295523972j)]
+        a_zref = [(1+0j),
+                  (-1.8839476369292854-0.606808151331815j),
+                  (0.7954687330018285+0.5717459398142481j)]
+
+        b_z, a_z = bilinear(b, a, fs)
+
+        # the 3 ulp difference determined from testing
+        assert_array_almost_equal_nulp(b_z, b_zref, 3)
+        assert_array_almost_equal_nulp(a_z, a_zref, 3)
+
+
+    def test_fs_validation(self):
+        b = [0.14879732743343033]
+        a = [1, 0.54552236880522209, 0.14879732743343033]
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            bilinear(b, a, fs=np.array([10, 20]))
+
+        with pytest.raises(ValueError, match="Sampling.*be none"):
+            bilinear(b, a, fs=None)
+
+
+class TestLp2lp_zpk:
+
+    @xfail_xp_backends(
+        'dask.array', reason='https://github.com/dask/dask/issues/11883'
+    )
+    def test_basic(self, xp):
+        z = xp.asarray([])
+        p = xp.asarray([(-1+1j) / math.sqrt(2), (-1-1j) / math.sqrt(2)])
+        k = 1
+        z_lp, p_lp, k_lp = lp2lp_zpk(z, p, k, 5)
+        xp_assert_equal(z_lp, xp.asarray([]))
+        xp_assert_close(_sort_cmplx(p_lp, xp=xp), _sort_cmplx(p, xp=xp) * 5)
+        assert k_lp == 25.
+
+        # Pseudo-Chebyshev with both poles and zeros
+        z = xp.asarray([-2j, +2j])
+        p = xp.asarray([-0.75, -0.5-0.5j, -0.5+0.5j])
+        k = 3
+        z_lp, p_lp, k_lp = lp2lp_zpk(z, p, k, 20)
+        xp_assert_close(
+            _sort_cmplx(z_lp, xp=xp), _sort_cmplx([-40j, +40j], xp=xp)
+        )
+        xp_assert_close(
+            _sort_cmplx(p_lp, xp=xp), _sort_cmplx([-15, -10-10j, -10+10j], xp=xp)
+        )
+        assert k_lp == 60.
+
+    def test_fs_validation(self):
+        z = [-2j, +2j]
+        p = [-0.75, -0.5 - 0.5j, -0.5 + 0.5j]
+        k = 3
+
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            bilinear_zpk(z, p, k, fs=np.array([10, 20]))
+
+        with pytest.raises(ValueError, match="Sampling.*be none"):
+            bilinear_zpk(z, p, k, fs=None)
+
+
+class TestLp2hp_zpk:
+
+    @xfail_xp_backends(
+        'dask.array', reason='https://github.com/dask/dask/issues/11883'
+    )
+    def test_basic(self, xp):
+        z = xp.asarray([])
+        p = xp.asarray([(-1+1j) / math.sqrt(2), (-1-1j) / math.sqrt(2)])
+        k = 1
+
+        z_hp, p_hp, k_hp = lp2hp_zpk(z, p, k, 5)
+        xp_assert_equal(z_hp, xp.asarray([0.0, 0.0], dtype=z_hp.dtype))
+        xp_assert_close(_sort_cmplx(p_hp, xp=xp), _sort_cmplx(p, xp=xp) * 5)
+        assert math.isclose(k_hp, 1.0, rel_tol=4e-7)
+
+        z = xp.asarray([-2j, +2j])
+        p = xp.asarray([-0.75, -0.5-0.5j, -0.5+0.5j])
+        k = 3
+        z_hp, p_hp, k_hp = lp2hp_zpk(z, p, k, 6)
+        xp_assert_close(
+            _sort_cmplx(z_hp, xp=xp), _sort_cmplx([-3j, 0, +3j], xp=xp)
+        )
+        xp_assert_close(
+            _sort_cmplx(p_hp, xp=xp), _sort_cmplx([-8, -6-6j, -6+6j], xp=xp)
+        )
+        assert k_hp == 32.0
+
+
+class TestLp2bp_zpk:
+
+    @xfail_xp_backends(
+        'dask.array', reason='https://github.com/dask/dask/issues/11883'
+    )
+    def test_basic(self, xp):
+        z = xp.asarray([-2j, +2j])
+        p = xp.asarray([-0.75, -0.5-0.5j, -0.5+0.5j])
+        k = 3
+        z_bp, p_bp, k_bp = lp2bp_zpk(z, p, k, 15, 8)
+        xp_assert_close(
+            _sort_cmplx(z_bp, xp=xp),
+            _sort_cmplx([-25j, -9j, 0, +9j, +25j], xp=xp), check_dtype=False
+        )
+        xp_assert_close(
+            _sort_cmplx(p_bp, xp=xp),
+            _sort_cmplx(
+                [-3 + 6j*math.sqrt(6), -3 - 6j*math.sqrt(6),
+                 +2j + cmath.sqrt(-8j - 225) - 2, -2j + cmath.sqrt(+8j - 225) - 2,
+                 +2j - cmath.sqrt(-8j - 225) - 2, -2j - cmath.sqrt(+8j - 225) - 2
+                ], xp=xp
+            ), check_dtype=False
+        )
+        assert math.isclose(k_bp, 24.0)
+
+
+class TestLp2bs_zpk:
+
+    @xfail_xp_backends(
+        'dask.array', reason='https://github.com/dask/dask/issues/11883'
+    )
+    def test_basic(self, xp):
+        z = xp.asarray([-2j, +2j])
+        p = xp.asarray([-0.75, -0.5-0.5j, -0.5+0.5j])
+        k = 3
+
+        z_bs, p_bs, k_bs = lp2bs_zpk(z, p, k, 35, 12)
+
+        xp_assert_close(
+            _sort_cmplx(z_bs, xp=xp),
+            _sort_cmplx([+35j, -35j,
+                         +3j + math.sqrt(1234)*1j,
+                         -3j + math.sqrt(1234)*1j,
+                         +3j - math.sqrt(1234)*1j,
+                         -3j - math.sqrt(1234)*1j], xp=xp), check_dtype=False
+        )
+        xp_assert_close(
+            _sort_cmplx(p_bs, xp=xp),
+            _sort_cmplx([+3j*math.sqrt(129) - 8,
+                         -3j*math.sqrt(129) - 8,
+                         (-6 + 6j) - cmath.sqrt(-1225 - 72j),
+                         (-6 - 6j) - cmath.sqrt(-1225 + 72j),
+                         (-6 + 6j) + cmath.sqrt(-1225 - 72j),
+                         (-6 - 6j) + cmath.sqrt(-1225 + 72j), ], xp=xp),
+            check_dtype=False
+        )
+        assert math.isclose(k_bs, 32.0)
+
+
+class TestBilinear_zpk:
+
+    @xfail_xp_backends(
+        'dask.array', reason='https://github.com/dask/dask/issues/11883'
+    )
+    def test_basic(self, xp):
+        z = xp.asarray([-2j, +2j])
+        p = xp.asarray([-0.75, -0.5-0.5j, -0.5+0.5j])
+        k = 3
+
+        z_d, p_d, k_d = bilinear_zpk(z, p, k, 10)
+
+        xp_assert_close(
+            _sort_cmplx(z_d, xp=xp),
+            _sort_cmplx([(20-2j) / (20+2j), (20+2j) / (20-2j), -1], xp=xp)
+        )
+        xp_assert_close(
+            _sort_cmplx(p_d, xp=xp),
+            _sort_cmplx(
+                [77/83, (1j/2 + 39/2) / (41/2 - 1j/2), (39/2 - 1j/2) / (1j/2 + 41/2)],
+                xp=xp
+            )
+        )
+        assert math.isclose(k_d, 9696/69803, rel_tol=4e-7)
+
+
+class TestPrototypeType:
+
+    def test_output_type(self):
+        # Prototypes should consistently output arrays, not lists
+        # https://github.com/scipy/scipy/pull/441
+        for func in (buttap,
+                     besselap,
+                     lambda N: cheb1ap(N, 1),
+                     lambda N: cheb2ap(N, 20),
+                     lambda N: ellipap(N, 1, 20)):
+            for N in range(7):
+                z, p, k = func(N)
+                assert isinstance(z, np.ndarray)
+                assert isinstance(p, np.ndarray)
+
+
+def dB(x):
+    # Return magnitude in decibels, avoiding divide-by-zero warnings
+    # (and deal with some "not less-ordered" errors when -inf shows up)
+    return 20 * np.log10(np.maximum(np.abs(x), np.finfo(np.float64).tiny))
+
+
+class TestButtord:
+
+    def test_lowpass(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+        N, Wn = buttord(wp, ws, rp, rs, False)
+        b, a = butter(N, Wn, 'lowpass', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp < dB(h[w <= wp]))
+        assert np.all(dB(h[ws <= w]) < -rs)
+
+        assert N == 16
+        xp_assert_close(Wn,
+                        2.0002776782743284e-01, rtol=1e-15)
+
+    def test_highpass(self):
+        wp = 0.3
+        ws = 0.2
+        rp = 3
+        rs = 70
+        N, Wn = buttord(wp, ws, rp, rs, False)
+        b, a = butter(N, Wn, 'highpass', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp < dB(h[wp <= w]))
+        assert np.all(dB(h[w <= ws]) < -rs)
+
+        assert N == 18
+        xp_assert_close(Wn,
+                        2.9996603079132672e-01, rtol=1e-15)
+
+    def test_bandpass(self):
+        wp = [0.2, 0.5]
+        ws = [0.1, 0.6]
+        rp = 3
+        rs = 80
+        N, Wn = buttord(wp, ws, rp, rs, False)
+        b, a = butter(N, Wn, 'bandpass', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+
+        assert np.all((-rp - 0.1) < dB(h[np.logical_and(wp[0] <= w, w <= wp[1])]))
+
+        assert np.all(dB(h[np.logical_or(w <= ws[0], ws[1] <= w)]) < (-rs + 0.1))
+
+        assert N == 18
+        xp_assert_close(Wn, [1.9998742411409134e-01, 5.0002139595676276e-01],
+                        rtol=1e-15)
+
+    def test_bandstop(self):
+        wp = [0.1, 0.6]
+        ws = [0.2, 0.5]
+        rp = 3
+        rs = 90
+        N, Wn = buttord(wp, ws, rp, rs, False)
+        b, a = butter(N, Wn, 'bandstop', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+
+        assert np.all(-rp < dB(h[np.logical_or(w <= wp[0], wp[1] <= w)]))
+        assert np.all(dB(h[np.logical_and(ws[0] <= w, w <= ws[1])]) < -rs)
+
+        assert N == 20
+        xp_assert_close(Wn, [1.4759432329294042e-01, 5.9997365985276407e-01],
+                        rtol=1e-6)
+
+    def test_analog(self):
+        wp = 200
+        ws = 600
+        rp = 3
+        rs = 60
+        N, Wn = buttord(wp, ws, rp, rs, True)
+        b, a = butter(N, Wn, 'lowpass', True)
+        w, h = freqs(b, a)
+        assert np.all(-rp < dB(h[w <= wp]))
+        assert np.all(dB(h[ws <= w]) < -rs)
+
+        assert N == 7
+        xp_assert_close(Wn, 2.0006785355671877e+02, rtol=1e-15)
+
+        n, Wn = buttord(1, 550/450, 1, 26, analog=True)
+        assert n == 19
+        xp_assert_close(Wn, 1.0361980524629517, rtol=1e-15)
+
+        xp_assert_equal(buttord(1, 1.2, 1, 80, analog=True)[0], 55)
+
+    def test_fs_param(self):
+        wp = [4410, 11025]
+        ws = [2205, 13230]
+        rp = 3
+        rs = 80
+        fs = 44100
+        N, Wn = buttord(wp, ws, rp, rs, False, fs=fs)
+        b, a = butter(N, Wn, 'bandpass', False, fs=fs)
+        w, h = freqz(b, a, fs=fs)
+        assert np.all(-rp - 0.1 < dB(h[np.logical_and(wp[0] <= w, w <= wp[1])]))
+        assert np.all(dB(h[np.logical_or(w <= ws[0], ws[1] <= w)]) < -rs + 0.1)
+
+        assert N == 18
+        xp_assert_close(Wn, [4409.722701715714, 11025.47178084662],
+                        rtol=1e-15)
+
+    def test_invalid_input(self):
+        with pytest.raises(ValueError) as exc_info:
+            buttord([20, 50], [14, 60], 3, 2)
+        assert "gpass should be smaller than gstop" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            buttord([20, 50], [14, 60], -1, 2)
+        assert "gpass should be larger than 0.0" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            buttord([20, 50], [14, 60], 1, -2)
+        assert "gstop should be larger than 0.0" in str(exc_info.value)
+
+    @pytest.mark.thread_unsafe
+    def test_runtime_warnings(self):
+        msg = "Order is zero.*|divide by zero encountered"
+        with pytest.warns(RuntimeWarning, match=msg):
+            buttord(0.0, 1.0, 3, 60)
+
+    def test_ellip_butter(self):
+        # The purpose of the test is to compare to some known output from past
+        # scipy versions. The values to compare to are generated with scipy
+        # 1.9.1 (there is nothing special about this particular version though)
+        n, wn = buttord([0.1, 0.6], [0.2, 0.5], 3, 60)
+        assert n == 14
+
+    def test_fs_validation(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            buttord(wp, ws, rp, rs, False, fs=np.array([10, 20]))
+
+
+class TestCheb1ord:
+
+    def test_lowpass(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+        N, Wn = cheb1ord(wp, ws, rp, rs, False)
+        b, a = cheby1(N, rp, Wn, 'low', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[w <= wp]))
+        assert np.all(dB(h[ws <= w]) < -rs + 0.1)
+
+        assert N == 8
+        xp_assert_close(Wn, 0.2, rtol=1e-15)
+
+    def test_highpass(self):
+        wp = 0.3
+        ws = 0.2
+        rp = 3
+        rs = 70
+        N, Wn = cheb1ord(wp, ws, rp, rs, False)
+        b, a = cheby1(N, rp, Wn, 'high', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[wp <= w]))
+        assert np.all(dB(h[w <= ws]) < -rs + 0.1)
+
+        assert N == 9
+        xp_assert_close(Wn, 0.3, rtol=1e-15)
+
+    def test_bandpass(self):
+        wp = [0.2, 0.5]
+        ws = [0.1, 0.6]
+        rp = 3
+        rs = 80
+        N, Wn = cheb1ord(wp, ws, rp, rs, False)
+        b, a = cheby1(N, rp, Wn, 'band', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[np.logical_and(wp[0] <= w, w <= wp[1])]))
+        assert np.all(dB(h[np.logical_or(w <= ws[0], ws[1] <= w)]) < -rs + 0.1)
+
+        assert N == 9
+        xp_assert_close(Wn, [0.2, 0.5], rtol=1e-15)
+
+    def test_bandstop(self):
+        wp = [0.1, 0.6]
+        ws = [0.2, 0.5]
+        rp = 3
+        rs = 90
+        N, Wn = cheb1ord(wp, ws, rp, rs, False)
+        b, a = cheby1(N, rp, Wn, 'stop', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[np.logical_or(w <= wp[0], wp[1] <= w)]))
+        assert np.all(dB(h[np.logical_and(ws[0] <= w, w <= ws[1])]) < -rs + 0.1)
+
+        assert N == 10
+        xp_assert_close(Wn, [0.14758232569947785, 0.6], rtol=1e-5)
+
+    def test_analog(self):
+        wp = 700
+        ws = 100
+        rp = 3
+        rs = 70
+        N, Wn = cheb1ord(wp, ws, rp, rs, True)
+        b, a = cheby1(N, rp, Wn, 'high', True)
+        w, h = freqs(b, a)
+        assert np.all(-rp - 0.1 < dB(h[wp <= w]))
+        assert np.all(dB(h[w <= ws]) < -rs + 0.1)
+
+        assert N == 4
+        xp_assert_close(Wn, 700.0, rtol=1e-15)
+
+        xp_assert_equal(cheb1ord(1, 1.2, 1, 80, analog=True)[0], 17)
+
+    def test_fs_param(self):
+        wp = 4800
+        ws = 7200
+        rp = 3
+        rs = 60
+        fs = 48000
+        N, Wn = cheb1ord(wp, ws, rp, rs, False, fs=fs)
+        b, a = cheby1(N, rp, Wn, 'low', False, fs=fs)
+        w, h = freqz(b, a, fs=fs)
+        assert np.all(-rp - 0.1 < dB(h[w <= wp]))
+        assert np.all(dB(h[ws <= w]) < -rs + 0.1)
+
+        assert N == 8
+        xp_assert_close(Wn, 4800.0, rtol=1e-15)
+
+    def test_invalid_input(self):
+        with pytest.raises(ValueError) as exc_info:
+            cheb1ord(0.2, 0.3, 3, 2)
+        assert "gpass should be smaller than gstop" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            cheb1ord(0.2, 0.3, -1, 2)
+        assert "gpass should be larger than 0.0" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            cheb1ord(0.2, 0.3, 1, -2)
+        assert "gstop should be larger than 0.0" in str(exc_info.value)
+
+    def test_ellip_cheb1(self):
+        # The purpose of the test is to compare to some known output from past
+        # scipy versions. The values to compare to are generated with scipy
+        # 1.9.1 (there is nothing special about this particular version though)
+        n, wn = cheb1ord([0.1, 0.6], [0.2, 0.5], 3, 60)
+        assert n == 7
+
+        n2, w2 = cheb2ord([0.1, 0.6], [0.2, 0.5], 3, 60)
+        assert not (wn == w2).all()
+
+    def test_fs_validation(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            cheb1ord(wp, ws, rp, rs, False, fs=np.array([10, 20]))
+
+
+class TestCheb2ord:
+
+    def test_lowpass(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+        N, Wn = cheb2ord(wp, ws, rp, rs, False)
+        b, a = cheby2(N, rs, Wn, 'lp', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[w <= wp]))
+        assert np.all(dB(h[ws <= w]) < -rs + 0.1)
+
+        assert N == 8
+        xp_assert_close(Wn, 0.28647639976553163, rtol=1e-15)
+
+    def test_highpass(self):
+        wp = 0.3
+        ws = 0.2
+        rp = 3
+        rs = 70
+        N, Wn = cheb2ord(wp, ws, rp, rs, False)
+        b, a = cheby2(N, rs, Wn, 'hp', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[wp <= w]))
+        assert np.all(dB(h[w <= ws]) < -rs + 0.1)
+
+        assert N == 9
+        xp_assert_close(Wn, 0.20697492182903282, rtol=1e-15)
+
+    def test_bandpass(self):
+        wp = [0.2, 0.5]
+        ws = [0.1, 0.6]
+        rp = 3
+        rs = 80
+        N, Wn = cheb2ord(wp, ws, rp, rs, False)
+        b, a = cheby2(N, rs, Wn, 'bp', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[np.logical_and(wp[0] <= w, w <= wp[1])]))
+        assert np.all(dB(h[np.logical_or(w <= ws[0], ws[1] <= w)]) < -rs + 0.1)
+
+        assert N == 9
+        xp_assert_close(Wn, [0.14876937565923479, 0.59748447842351482],
+                        rtol=1e-15)
+
+    def test_bandstop(self):
+        wp = [0.1, 0.6]
+        ws = [0.2, 0.5]
+        rp = 3
+        rs = 90
+        N, Wn = cheb2ord(wp, ws, rp, rs, False)
+        b, a = cheby2(N, rs, Wn, 'bs', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[np.logical_or(w <= wp[0], wp[1] <= w)]))
+        assert np.all(dB(h[np.logical_and(ws[0] <= w, w <= ws[1])]) < -rs + 0.1)
+
+        assert N == 10
+        xp_assert_close(Wn, [0.19926249974781743, 0.50125246585567362],
+                        rtol=1e-6)
+
+    def test_analog(self):
+        wp = [20, 50]
+        ws = [10, 60]
+        rp = 3
+        rs = 80
+        N, Wn = cheb2ord(wp, ws, rp, rs, True)
+        b, a = cheby2(N, rs, Wn, 'bp', True)
+        w, h = freqs(b, a)
+        assert np.all(-rp - 0.1 < dB(h[np.logical_and(wp[0] <= w, w <= wp[1])]))
+        assert np.all(dB(h[np.logical_or(w <= ws[0], ws[1] <= w)]) < -rs + 0.1)
+
+        assert N == 11
+        xp_assert_close(Wn, [1.673740595370124e+01, 5.974641487254268e+01],
+                        rtol=1e-15)
+
+    def test_fs_param(self):
+        wp = 150
+        ws = 100
+        rp = 3
+        rs = 70
+        fs = 1000
+        N, Wn = cheb2ord(wp, ws, rp, rs, False, fs=fs)
+        b, a = cheby2(N, rs, Wn, 'hp', False, fs=fs)
+        w, h = freqz(b, a, fs=fs)
+        assert np.all(-rp - 0.1 < dB(h[wp <= w]))
+        assert np.all(dB(h[w <= ws]) < -rs + 0.1)
+
+        assert N == 9
+        xp_assert_close(Wn, 103.4874609145164, rtol=1e-15)
+
+    def test_invalid_input(self):
+        with pytest.raises(ValueError) as exc_info:
+            cheb2ord([0.1, 0.6], [0.2, 0.5], 3, 2)
+        assert "gpass should be smaller than gstop" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            cheb2ord([0.1, 0.6], [0.2, 0.5], -1, 2)
+        assert "gpass should be larger than 0.0" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            cheb2ord([0.1, 0.6], [0.2, 0.5], 1, -2)
+        assert "gstop should be larger than 0.0" in str(exc_info.value)
+
+    def test_ellip_cheb2(self):
+        # The purpose of the test is to compare to some known output from past
+        # scipy versions. The values to compare to are generated with scipy
+        # 1.9.1 (there is nothing special about this particular version though)
+        n, wn = cheb2ord([0.1, 0.6], [0.2, 0.5], 3, 60)
+        assert n == 7
+
+        n1, w1 = cheb1ord([0.1, 0.6], [0.2, 0.5], 3, 60)
+        assert not (wn == w1).all()
+
+    def test_fs_validation(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            cheb2ord(wp, ws, rp, rs, False, fs=np.array([10, 20]))
+
+
+class TestEllipord:
+
+    def test_lowpass(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+        N, Wn = ellipord(wp, ws, rp, rs, False)
+        b, a = ellip(N, rp, rs, Wn, 'lp', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[w <= wp]))
+        assert np.all(dB(h[ws <= w]) < -rs + 0.1)
+
+        assert N == 5
+        xp_assert_close(Wn, 0.2, rtol=1e-15)
+
+    def test_lowpass_1000dB(self):
+        # failed when ellipkm1 wasn't used in ellipord and ellipap
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 1000
+        N, Wn = ellipord(wp, ws, rp, rs, False)
+        sos = ellip(N, rp, rs, Wn, 'lp', False, output='sos')
+        w, h = freqz_sos(sos)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[w <= wp]))
+        assert np.all(dB(h[ws <= w]) < -rs + 0.1)
+
+    def test_highpass(self):
+        wp = 0.3
+        ws = 0.2
+        rp = 3
+        rs = 70
+        N, Wn = ellipord(wp, ws, rp, rs, False)
+        b, a = ellip(N, rp, rs, Wn, 'hp', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[wp <= w]))
+        assert np.all(dB(h[w <= ws]) < -rs + 0.1)
+
+        assert N == 6
+        xp_assert_close(Wn, 0.3, rtol=1e-15)
+
+    def test_bandpass(self):
+        wp = [0.2, 0.5]
+        ws = [0.1, 0.6]
+        rp = 3
+        rs = 80
+        N, Wn = ellipord(wp, ws, rp, rs, False)
+        b, a = ellip(N, rp, rs, Wn, 'bp', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[np.logical_and(wp[0] <= w, w <= wp[1])]))
+        assert np.all(dB(h[np.logical_or(w <= ws[0], ws[1] <= w)]) < -rs + 0.1)
+
+        assert N == 6
+        xp_assert_close(Wn, [0.2, 0.5], rtol=1e-15)
+
+    def test_bandstop(self):
+        wp = [0.1, 0.6]
+        ws = [0.2, 0.5]
+        rp = 3
+        rs = 90
+        N, Wn = ellipord(wp, ws, rp, rs, False)
+        b, a = ellip(N, rp, rs, Wn, 'bs', False)
+        w, h = freqz(b, a)
+        w /= np.pi
+        assert np.all(-rp - 0.1 < dB(h[np.logical_or(w <= wp[0], wp[1] <= w)]))
+        assert np.all(dB(h[np.logical_and(ws[0] <= w, w <= ws[1])]) < -rs + 0.1)
+
+        assert N == 7
+        xp_assert_close(Wn, [0.14758232794342988, 0.6], rtol=1e-5)
+
+    def test_analog(self):
+        wp = [1000, 6000]
+        ws = [2000, 5000]
+        rp = 3
+        rs = 90
+        N, Wn = ellipord(wp, ws, rp, rs, True)
+        b, a = ellip(N, rp, rs, Wn, 'bs', True)
+        w, h = freqs(b, a)
+        assert np.all(-rp - 0.1 < dB(h[np.logical_or(w <= wp[0], wp[1] <= w)]))
+        assert np.all(dB(h[np.logical_and(ws[0] <= w, w <= ws[1])]) < -rs + 0.1)
+
+        assert N == 8
+        xp_assert_close(Wn, [1666.6666, 6000])
+
+        assert ellipord(1, 1.2, 1, 80, analog=True)[0] == 9
+
+    def test_fs_param(self):
+        wp = [400, 2400]
+        ws = [800, 2000]
+        rp = 3
+        rs = 90
+        fs = 8000
+        N, Wn = ellipord(wp, ws, rp, rs, False, fs=fs)
+        b, a = ellip(N, rp, rs, Wn, 'bs', False, fs=fs)
+        w, h = freqz(b, a, fs=fs)
+        assert np.all(-rp - 0.1 < dB(h[np.logical_or(w <= wp[0], wp[1] <= w)]))
+        assert np.all(dB(h[np.logical_and(ws[0] <= w, w <= ws[1])]) < -rs + 0.1)
+
+        assert N == 7
+        xp_assert_close(Wn, [590.3293117737195, 2400], rtol=1e-5)
+
+    def test_invalid_input(self):
+        with pytest.raises(ValueError) as exc_info:
+            ellipord(0.2, 0.5, 3, 2)
+        assert "gpass should be smaller than gstop" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            ellipord(0.2, 0.5, -1, 2)
+        assert "gpass should be larger than 0.0" in str(exc_info.value)
+
+        with pytest.raises(ValueError) as exc_info:
+            ellipord(0.2, 0.5, 1, -2)
+        assert "gstop should be larger than 0.0" in str(exc_info.value)
+
+    def test_ellip_butter(self):
+        # The purpose of the test is to compare to some known output from past
+        # scipy versions. The values to compare to are generated with scipy
+        # 1.9.1 (there is nothing special about this particular version though)
+        n, wn = ellipord([0.1, 0.6], [0.2, 0.5], 3, 60)
+        assert n == 5
+
+    def test_fs_validation(self):
+        wp = 0.2
+        ws = 0.3
+        rp = 3
+        rs = 60
+
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            ellipord(wp, ws, rp, rs, False, fs=np.array([10, 20]))
+
+
+class TestBessel:
+
+    def test_degenerate(self):
+        for norm in ('delay', 'phase', 'mag'):
+            # 0-order filter is just a passthrough
+            b, a = bessel(0, 1, analog=True, norm=norm)
+            xp_assert_equal(b, np.asarray([1.0]))
+            xp_assert_equal(a, np.asarray([1.0]))
+
+            # 1-order filter is same for all types
+            b, a = bessel(1, 1, analog=True, norm=norm)
+            xp_assert_close(b, np.asarray([1.0]), rtol=1e-15)
+            xp_assert_close(a, np.asarray([1.0, 1]), rtol=1e-15)
+
+            z, p, k = bessel(1, 0.3, analog=True, output='zpk', norm=norm)
+            xp_assert_equal(z, np.asarray([]))
+            xp_assert_close(p, np.asarray([-0.3+0j]), rtol=1e-14)
+            xp_assert_close(k, 0.3, rtol=1e-14)
+
+    def test_high_order(self):
+        # high even order, 'phase'
+        z, p, k = bessel(24, 100, analog=True, output='zpk')
+        z2 = []
+        p2 = [
+             -9.055312334014323e+01 + 4.844005815403969e+00j,
+             -8.983105162681878e+01 + 1.454056170018573e+01j,
+             -8.837357994162065e+01 + 2.426335240122282e+01j,
+             -8.615278316179575e+01 + 3.403202098404543e+01j,
+             -8.312326467067703e+01 + 4.386985940217900e+01j,
+             -7.921695461084202e+01 + 5.380628489700191e+01j,
+             -7.433392285433246e+01 + 6.388084216250878e+01j,
+             -6.832565803501586e+01 + 7.415032695116071e+01j,
+             -6.096221567378025e+01 + 8.470292433074425e+01j,
+             -5.185914574820616e+01 + 9.569048385258847e+01j,
+             -4.027853855197555e+01 + 1.074195196518679e+02j,
+             -2.433481337524861e+01 + 1.207298683731973e+02j,
+             ]
+        k2 = 9.999999999999989e+47
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(np.union1d(p2, np.conj(p2)), key=np.imag))
+        xp_assert_close(k, k2, rtol=1e-14)
+
+        # high odd order, 'phase'
+        z, p, k = bessel(23, 1000, analog=True, output='zpk')
+        z2 = []
+        p2 = [
+             -2.497697202208956e+02 + 1.202813187870698e+03j,
+             -4.126986617510172e+02 + 1.065328794475509e+03j,
+             -5.304922463809596e+02 + 9.439760364018479e+02j,
+             -9.027564978975828e+02 + 1.010534334242318e+02j,
+             -8.909283244406079e+02 + 2.023024699647598e+02j,
+             -8.709469394347836e+02 + 3.039581994804637e+02j,
+             -8.423805948131370e+02 + 4.062657947488952e+02j,
+             -8.045561642249877e+02 + 5.095305912401127e+02j,
+             -7.564660146766259e+02 + 6.141594859516342e+02j,
+             -6.965966033906477e+02 + 7.207341374730186e+02j,
+             -6.225903228776276e+02 + 8.301558302815096e+02j,
+             -9.066732476324988e+02]
+        k2 = 9.999999999999983e+68
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(np.union1d(p2, np.conj(p2)), key=np.imag))
+        xp_assert_close(k, k2, rtol=1e-14)
+
+        # high even order, 'delay' (Orchard 1965 "The Roots of the
+        # Maximally Flat-Delay Polynomials" Table 1)
+        z, p, k = bessel(31, 1, analog=True, output='zpk', norm='delay')
+        p2 = [-20.876706,
+              -20.826543 + 1.735732j,
+              -20.675502 + 3.473320j,
+              -20.421895 + 5.214702j,
+              -20.062802 + 6.961982j,
+              -19.593895 + 8.717546j,
+              -19.009148 + 10.484195j,
+              -18.300400 + 12.265351j,
+              -17.456663 + 14.065350j,
+              -16.463032 + 15.889910j,
+              -15.298849 + 17.746914j,
+              -13.934466 + 19.647827j,
+              -12.324914 + 21.610519j,
+              -10.395893 + 23.665701j,
+              - 8.005600 + 25.875019j,
+              - 4.792045 + 28.406037j,
+              ]
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(np.union1d(p2, np.conj(p2)), key=np.imag))
+
+        # high odd order, 'delay'
+        z, p, k = bessel(30, 1, analog=True, output='zpk', norm='delay')
+        p2 = [-20.201029 + 0.867750j,
+              -20.097257 + 2.604235j,
+              -19.888485 + 4.343721j,
+              -19.572188 + 6.088363j,
+              -19.144380 + 7.840570j,
+              -18.599342 + 9.603147j,
+              -17.929195 + 11.379494j,
+              -17.123228 + 13.173901j,
+              -16.166808 + 14.992008j,
+              -15.039580 + 16.841580j,
+              -13.712245 + 18.733902j,
+              -12.140295 + 20.686563j,
+              -10.250119 + 22.729808j,
+              - 7.901170 + 24.924391j,
+              - 4.734679 + 27.435615j,
+              ]
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(np.union1d(p2, np.conj(p2)), key=np.imag))
+
+    def test_refs(self):
+        # Compare to http://www.crbond.com/papers/bsf2.pdf
+        # "Delay Normalized Bessel Polynomial Coefficients"
+        bond_b = np.asarray([10395.0])
+        bond_a = np.asarray([1.0, 21, 210, 1260, 4725, 10395, 10395])
+        b, a = bessel(6, 1, norm='delay', analog=True)
+        xp_assert_close(b, bond_b)
+        xp_assert_close(a, bond_a)
+
+        # "Delay Normalized Bessel Pole Locations"
+        bond_poles = {
+            1: [-1.0000000000],
+            2: [-1.5000000000 + 0.8660254038j],
+            3: [-1.8389073227 + 1.7543809598j, -2.3221853546],
+            4: [-2.1037893972 + 2.6574180419j, -2.8962106028 + 0.8672341289j],
+            5: [-2.3246743032 + 3.5710229203j, -3.3519563992 + 1.7426614162j,
+                -3.6467385953],
+            6: [-2.5159322478 + 4.4926729537j, -3.7357083563 + 2.6262723114j,
+                -4.2483593959 + 0.8675096732j],
+            7: [-2.6856768789 + 5.4206941307j, -4.0701391636 + 3.5171740477j,
+                -4.7582905282 + 1.7392860611j, -4.9717868585],
+            8: [-2.8389839489 + 6.3539112986j, -4.3682892172 + 4.4144425005j,
+                -5.2048407906 + 2.6161751526j, -5.5878860433 + 0.8676144454j],
+            9: [-2.9792607982 + 7.2914636883j, -4.6384398872 + 5.3172716754j,
+                -5.6044218195 + 3.4981569179j, -6.1293679043 + 1.7378483835j,
+                -6.2970191817],
+            10: [-3.1089162336 + 8.2326994591j, -4.8862195669 + 6.2249854825j,
+                 -5.9675283286 + 4.3849471889j, -6.6152909655 + 2.6115679208j,
+                 -6.9220449054 + 0.8676651955j]
+            }
+
+        for N in range(1, 11):
+            p1 = np.sort(bond_poles[N])
+            p2 = np.sort(np.concatenate(_cplxreal(besselap(N, 'delay')[1])))
+            assert_array_almost_equal(p1, p2, decimal=10)
+
+        # "Frequency Normalized Bessel Pole Locations"
+        bond_poles = {
+            1: [-1.0000000000],
+            2: [-1.1016013306 + 0.6360098248j],
+            3: [-1.0474091610 + 0.9992644363j, -1.3226757999],
+            4: [-0.9952087644 + 1.2571057395j, -1.3700678306 + 0.4102497175j],
+            5: [-0.9576765486 + 1.4711243207j, -1.3808773259 + 0.7179095876j,
+                -1.5023162714],
+            6: [-0.9306565229 + 1.6618632689j, -1.3818580976 + 0.9714718907j,
+                -1.5714904036 + 0.3208963742j],
+            7: [-0.9098677806 + 1.8364513530j, -1.3789032168 + 1.1915667778j,
+                -1.6120387662 + 0.5892445069j, -1.6843681793],
+            8: [-0.8928697188 + 1.9983258436j, -1.3738412176 + 1.3883565759j,
+                -1.6369394181 + 0.8227956251j, -1.7574084004 + 0.2728675751j],
+            9: [-0.8783992762 + 2.1498005243j, -1.3675883098 + 1.5677337122j,
+                -1.6523964846 + 1.0313895670j, -1.8071705350 + 0.5123837306j,
+                -1.8566005012],
+            10: [-0.8657569017 + 2.2926048310j, -1.3606922784 + 1.7335057427j,
+                 -1.6618102414 + 1.2211002186j, -1.8421962445 + 0.7272575978j,
+                 -1.9276196914 + 0.2416234710j]
+            }
+
+        for N in range(1, 11):
+            p1 = np.sort(bond_poles[N])
+            p2 = np.sort(np.concatenate(_cplxreal(besselap(N, 'mag')[1])))
+            assert_array_almost_equal(p1, p2, decimal=10)
+
+        # Compare to https://www.ranecommercial.com/legacy/note147.html
+        # "Table 1 - Bessel Crossovers of Second, Third, and Fourth-Order"
+        a = np.asarray([1, 1, 1/3])
+        b2, a2 = bessel(2, 1, norm='delay', analog=True)
+        xp_assert_close(a[::-1], a2/b2)
+
+        a = np.asarray([1, 1, 2/5, 1/15])
+        b2, a2 = bessel(3, 1, norm='delay', analog=True)
+        xp_assert_close(a[::-1], a2/b2)
+
+        a = np.asarray([1, 1, 9/21, 2/21, 1/105])
+        b2, a2 = bessel(4, 1, norm='delay', analog=True)
+        xp_assert_close(a[::-1], a2/b2)
+
+        a = np.asarray([1, np.sqrt(3), 1])
+        b2, a2 = bessel(2, 1, norm='phase', analog=True)
+        xp_assert_close(a[::-1], a2/b2)
+
+        # TODO: Why so inaccurate?  Is reference flawed?
+        a = np.asarray([1, 2.481, 2.463, 1.018])
+        b2, a2 = bessel(3, 1, norm='phase', analog=True)
+        assert_array_almost_equal(a[::-1], a2/b2, decimal=1)
+
+        # TODO: Why so inaccurate?  Is reference flawed?
+        a = np.asarray([1, 3.240, 4.5, 3.240, 1.050])
+        b2, a2 = bessel(4, 1, norm='phase', analog=True)
+        assert_array_almost_equal(a[::-1], a2/b2, decimal=1)
+
+        # Table of -3 dB factors:
+        N, scale = 2, np.asarray([1.272, 1.272], dtype=np.complex128)
+        scale2 = besselap(N, 'mag')[1] / besselap(N, 'phase')[1]
+        assert_array_almost_equal(scale2, scale, decimal=3)
+
+        # TODO: Why so inaccurate?  Is reference flawed?
+        N, scale = 3, np.asarray([1.413, 1.413, 1.413], dtype=np.complex128)
+        scale2 = besselap(N, 'mag')[1] / besselap(N, 'phase')[1]
+        assert_array_almost_equal(scale2, scale, decimal=2)
+
+        # TODO: Why so inaccurate?  Is reference flawed?
+        N, scale = 4, np.asarray([1.533]*4, dtype=np.complex128)
+        scale2 = besselap(N, 'mag')[1] / besselap(N, 'phase')[1]
+        assert_array_almost_equal(scale, scale2, decimal=1)
+
+    def test_hardcoded(self):
+        # Compare to values from original hardcoded implementation
+        originals = {
+            0: [],
+            1: [-1],
+            2: [-.8660254037844386467637229 + .4999999999999999999999996j],
+            3: [-.9416000265332067855971980,
+                -.7456403858480766441810907 + .7113666249728352680992154j],
+            4: [-.6572111716718829545787788 + .8301614350048733772399715j,
+                -.9047587967882449459642624 + .2709187330038746636700926j],
+            5: [-.9264420773877602247196260,
+                -.8515536193688395541722677 + .4427174639443327209850002j,
+                -.5905759446119191779319432 + .9072067564574549539291747j],
+            6: [-.9093906830472271808050953 + .1856964396793046769246397j,
+                -.7996541858328288520243325 + .5621717346937317988594118j,
+                -.5385526816693109683073792 + .9616876881954277199245657j],
+            7: [-.9194871556490290014311619,
+                -.8800029341523374639772340 + .3216652762307739398381830j,
+                -.7527355434093214462291616 + .6504696305522550699212995j,
+                -.4966917256672316755024763 + 1.002508508454420401230220j],
+            8: [-.9096831546652910216327629 + .1412437976671422927888150j,
+                -.8473250802359334320103023 + .4259017538272934994996429j,
+                -.7111381808485399250796172 + .7186517314108401705762571j,
+                -.4621740412532122027072175 + 1.034388681126901058116589j],
+            9: [-.9154957797499037686769223,
+                -.8911217017079759323183848 + .2526580934582164192308115j,
+                -.8148021112269012975514135 + .5085815689631499483745341j,
+                -.6743622686854761980403401 + .7730546212691183706919682j,
+                -.4331415561553618854685942 + 1.060073670135929666774323j],
+            10: [-.9091347320900502436826431 + .1139583137335511169927714j,
+                 -.8688459641284764527921864 + .3430008233766309973110589j,
+                 -.7837694413101441082655890 + .5759147538499947070009852j,
+                 -.6417513866988316136190854 + .8175836167191017226233947j,
+                 -.4083220732868861566219785 + 1.081274842819124562037210j],
+            11: [-.9129067244518981934637318,
+                 -.8963656705721166099815744 + .2080480375071031919692341j,
+                 -.8453044014712962954184557 + .4178696917801248292797448j,
+                 -.7546938934722303128102142 + .6319150050721846494520941j,
+                 -.6126871554915194054182909 + .8547813893314764631518509j,
+                 -.3868149510055090879155425 + 1.099117466763120928733632j],
+            12: [-.9084478234140682638817772 + 95506365213450398415258360e-27j,
+                 -.8802534342016826507901575 + .2871779503524226723615457j,
+                 -.8217296939939077285792834 + .4810212115100676440620548j,
+                 -.7276681615395159454547013 + .6792961178764694160048987j,
+                 -.5866369321861477207528215 + .8863772751320727026622149j,
+                 -.3679640085526312839425808 + 1.114373575641546257595657j],
+            13: [-.9110914665984182781070663,
+                 -.8991314665475196220910718 + .1768342956161043620980863j,
+                 -.8625094198260548711573628 + .3547413731172988997754038j,
+                 -.7987460692470972510394686 + .5350752120696801938272504j,
+                 -.7026234675721275653944062 + .7199611890171304131266374j,
+                 -.5631559842430199266325818 + .9135900338325109684927731j,
+                 -.3512792323389821669401925 + 1.127591548317705678613239j],
+            14: [-.9077932138396487614720659 + 82196399419401501888968130e-27j,
+                 -.8869506674916445312089167 + .2470079178765333183201435j,
+                 -.8441199160909851197897667 + .4131653825102692595237260j,
+                 -.7766591387063623897344648 + .5819170677377608590492434j,
+                 -.6794256425119233117869491 + .7552857305042033418417492j,
+                 -.5418766775112297376541293 + .9373043683516919569183099j,
+                 -.3363868224902037330610040 + 1.139172297839859991370924j],
+            15: [-.9097482363849064167228581,
+                 -.9006981694176978324932918 + .1537681197278439351298882j,
+                 -.8731264620834984978337843 + .3082352470564267657715883j,
+                 -.8256631452587146506294553 + .4642348752734325631275134j,
+                 -.7556027168970728127850416 + .6229396358758267198938604j,
+                 -.6579196593110998676999362 + .7862895503722515897065645j,
+                 -.5224954069658330616875186 + .9581787261092526478889345j,
+                 -.3229963059766444287113517 + 1.149416154583629539665297j],
+            16: [-.9072099595087001356491337 + 72142113041117326028823950e-27j,
+                 -.8911723070323647674780132 + .2167089659900576449410059j,
+                 -.8584264231521330481755780 + .3621697271802065647661080j,
+                 -.8074790293236003885306146 + .5092933751171800179676218j,
+                 -.7356166304713115980927279 + .6591950877860393745845254j,
+                 -.6379502514039066715773828 + .8137453537108761895522580j,
+                 -.5047606444424766743309967 + .9767137477799090692947061j,
+                 -.3108782755645387813283867 + 1.158552841199330479412225j],
+            17: [-.9087141161336397432860029,
+                 -.9016273850787285964692844 + .1360267995173024591237303j,
+                 -.8801100704438627158492165 + .2725347156478803885651973j,
+                 -.8433414495836129204455491 + .4100759282910021624185986j,
+                 -.7897644147799708220288138 + .5493724405281088674296232j,
+                 -.7166893842372349049842743 + .6914936286393609433305754j,
+                 -.6193710717342144521602448 + .8382497252826992979368621j,
+                 -.4884629337672704194973683 + .9932971956316781632345466j,
+                 -.2998489459990082015466971 + 1.166761272925668786676672j],
+            18: [-.9067004324162775554189031 + 64279241063930693839360680e-27j,
+                 -.8939764278132455733032155 + .1930374640894758606940586j,
+                 -.8681095503628830078317207 + .3224204925163257604931634j,
+                 -.8281885016242836608829018 + .4529385697815916950149364j,
+                 -.7726285030739558780127746 + .5852778162086640620016316j,
+                 -.6987821445005273020051878 + .7204696509726630531663123j,
+                 -.6020482668090644386627299 + .8602708961893664447167418j,
+                 -.4734268069916151511140032 + 1.008234300314801077034158j,
+                 -.2897592029880489845789953 + 1.174183010600059128532230j],
+            19: [-.9078934217899404528985092,
+                 -.9021937639390660668922536 + .1219568381872026517578164j,
+                 -.8849290585034385274001112 + .2442590757549818229026280j,
+                 -.8555768765618421591093993 + .3672925896399872304734923j,
+                 -.8131725551578197705476160 + .4915365035562459055630005j,
+                 -.7561260971541629355231897 + .6176483917970178919174173j,
+                 -.6818424412912442033411634 + .7466272357947761283262338j,
+                 -.5858613321217832644813602 + .8801817131014566284786759j,
+                 -.4595043449730988600785456 + 1.021768776912671221830298j,
+                 -.2804866851439370027628724 + 1.180931628453291873626003j],
+            20: [-.9062570115576771146523497 + 57961780277849516990208850e-27j,
+                 -.8959150941925768608568248 + .1740317175918705058595844j,
+                 -.8749560316673332850673214 + .2905559296567908031706902j,
+                 -.8427907479956670633544106 + .4078917326291934082132821j,
+                 -.7984251191290606875799876 + .5264942388817132427317659j,
+                 -.7402780309646768991232610 + .6469975237605228320268752j,
+                 -.6658120544829934193890626 + .7703721701100763015154510j,
+                 -.5707026806915714094398061 + .8982829066468255593407161j,
+                 -.4465700698205149555701841 + 1.034097702560842962315411j,
+                 -.2719299580251652601727704 + 1.187099379810885886139638j],
+            21: [-.9072262653142957028884077,
+                 -.9025428073192696303995083 + .1105252572789856480992275j,
+                 -.8883808106664449854431605 + .2213069215084350419975358j,
+                 -.8643915813643204553970169 + .3326258512522187083009453j,
+                 -.8299435470674444100273463 + .4448177739407956609694059j,
+                 -.7840287980408341576100581 + .5583186348022854707564856j,
+                 -.7250839687106612822281339 + .6737426063024382240549898j,
+                 -.6506315378609463397807996 + .7920349342629491368548074j,
+                 -.5564766488918562465935297 + .9148198405846724121600860j,
+                 -.4345168906815271799687308 + 1.045382255856986531461592j,
+                 -.2640041595834031147954813 + 1.192762031948052470183960j],
+            22: [-.9058702269930872551848625 + 52774908289999045189007100e-27j,
+                 -.8972983138153530955952835 + .1584351912289865608659759j,
+                 -.8799661455640176154025352 + .2644363039201535049656450j,
+                 -.8534754036851687233084587 + .3710389319482319823405321j,
+                 -.8171682088462720394344996 + .4785619492202780899653575j,
+                 -.7700332930556816872932937 + .5874255426351153211965601j,
+                 -.7105305456418785989070935 + .6982266265924524000098548j,
+                 -.6362427683267827226840153 + .8118875040246347267248508j,
+                 -.5430983056306302779658129 + .9299947824439872998916657j,
+                 -.4232528745642628461715044 + 1.055755605227545931204656j,
+                 -.2566376987939318038016012 + 1.197982433555213008346532j],
+            23: [-.9066732476324988168207439,
+                 -.9027564979912504609412993 + .1010534335314045013252480j,
+                 -.8909283242471251458653994 + .2023024699381223418195228j,
+                 -.8709469395587416239596874 + .3039581993950041588888925j,
+                 -.8423805948021127057054288 + .4062657948237602726779246j,
+                 -.8045561642053176205623187 + .5095305912227258268309528j,
+                 -.7564660146829880581478138 + .6141594859476032127216463j,
+                 -.6965966033912705387505040 + .7207341374753046970247055j,
+                 -.6225903228771341778273152 + .8301558302812980678845563j,
+                 -.5304922463810191698502226 + .9439760364018300083750242j,
+                 -.4126986617510148836149955 + 1.065328794475513585531053j,
+                 -.2497697202208956030229911 + 1.202813187870697831365338j],
+            24: [-.9055312363372773709269407 + 48440066540478700874836350e-27j,
+                 -.8983105104397872954053307 + .1454056133873610120105857j,
+                 -.8837358034555706623131950 + .2426335234401383076544239j,
+                 -.8615278304016353651120610 + .3403202112618624773397257j,
+                 -.8312326466813240652679563 + .4386985933597305434577492j,
+                 -.7921695462343492518845446 + .5380628490968016700338001j,
+                 -.7433392285088529449175873 + .6388084216222567930378296j,
+                 -.6832565803536521302816011 + .7415032695091650806797753j,
+                 -.6096221567378335562589532 + .8470292433077202380020454j,
+                 -.5185914574820317343536707 + .9569048385259054576937721j,
+                 -.4027853855197518014786978 + 1.074195196518674765143729j,
+                 -.2433481337524869675825448 + 1.207298683731972524975429j],
+            25: [-.9062073871811708652496104,
+                 -.9028833390228020537142561 + 93077131185102967450643820e-27j,
+                 -.8928551459883548836774529 + .1863068969804300712287138j,
+                 -.8759497989677857803656239 + .2798521321771408719327250j,
+                 -.8518616886554019782346493 + .3738977875907595009446142j,
+                 -.8201226043936880253962552 + .4686668574656966589020580j,
+                 -.7800496278186497225905443 + .5644441210349710332887354j,
+                 -.7306549271849967721596735 + .6616149647357748681460822j,
+                 -.6704827128029559528610523 + .7607348858167839877987008j,
+                 -.5972898661335557242320528 + .8626676330388028512598538j,
+                 -.5073362861078468845461362 + .9689006305344868494672405j,
+                 -.3934529878191079606023847 + 1.082433927173831581956863j,
+                 -.2373280669322028974199184 + 1.211476658382565356579418j],
+            }
+        for N in originals:
+            p1 = sorted(np.union1d(originals[N],
+                                   np.conj(originals[N])), key=np.imag)
+            p2 = sorted(besselap(N)[1], key=np.imag)
+            xp_assert_close(p1,
+                            p2, rtol=1e-14, check_dtype=False)
+
+    def test_norm_phase(self):
+        # Test some orders and frequencies and see that they have the right
+        # phase at w0
+        for N in (1, 2, 3, 4, 5, 51, 72):
+            for w0 in (1, 100):
+                b, a = bessel(N, w0, analog=True, norm='phase')
+                w = np.linspace(0, w0, 100)
+                w, h = freqs(b, a, w)
+                phase = np.unwrap(np.angle(h))
+                xp_assert_close(phase[[0, -1]], (0, -N*pi/4), rtol=1e-1)
+
+    def test_norm_mag(self):
+        # Test some orders and frequencies and see that they have the right
+        # mag at w0
+        for N in (1, 2, 3, 4, 5, 51, 72):
+            for w0 in (1, 100):
+                b, a = bessel(N, w0, analog=True, norm='mag')
+                w = (0, w0)
+                w, h = freqs(b, a, w)
+                mag = abs(h)
+                xp_assert_close(mag, (1, 1/np.sqrt(2)))
+
+    def test_norm_delay(self):
+        # Test some orders and frequencies and see that they have the right
+        # delay at DC
+        for N in (1, 2, 3, 4, 5, 51, 72):
+            for w0 in (1, 100):
+                b, a = bessel(N, w0, analog=True, norm='delay')
+                w = np.linspace(0, 10*w0, 1000)
+                w, h = freqs(b, a, w)
+                delay = -np.diff(np.unwrap(np.angle(h)))/np.diff(w)
+                xp_assert_close(delay[0], 1/w0, rtol=1e-4)
+
+    def test_norm_factor(self):
+        mpmath_values = {
+            1: 1.0, 2: 1.361654128716130520, 3: 1.755672368681210649,
+            4: 2.113917674904215843, 5: 2.427410702152628137,
+            6: 2.703395061202921876, 7: 2.951722147038722771,
+            8: 3.179617237510651330, 9: 3.391693138911660101,
+            10: 3.590980594569163482, 11: 3.779607416439620092,
+            12: 3.959150821144285315, 13: 4.130825499383535980,
+            14: 4.295593409533637564, 15: 4.454233021624377494,
+            16: 4.607385465472647917, 17: 4.755586548961147727,
+            18: 4.899289677284488007, 19: 5.038882681488207605,
+            20: 5.174700441742707423, 21: 5.307034531360917274,
+            22: 5.436140703250035999, 23: 5.562244783787878196,
+            24: 5.685547371295963521, 25: 5.806227623775418541,
+            50: 8.268963160013226298, 51: 8.352374541546012058,
+            }
+        for N in mpmath_values:
+            z, p, k = besselap(N, 'delay')
+            xp_assert_close(mpmath_values[N], _norm_factor(p, k), rtol=1e-13)
+
+    def test_bessel_poly(self):
+        xp_assert_equal(_bessel_poly(5), [945, 945, 420, 105, 15, 1])
+        xp_assert_equal(_bessel_poly(4, True), [1, 10, 45, 105, 105])
+
+    def test_bessel_zeros(self):
+        xp_assert_equal(_bessel_zeros(0), [])
+
+    def test_invalid(self):
+        assert_raises(ValueError, besselap, 5, 'nonsense')
+        assert_raises(ValueError, besselap, -5)
+        assert_raises(ValueError, besselap, 3.2)
+        assert_raises(ValueError, _bessel_poly, -3)
+        assert_raises(ValueError, _bessel_poly, 3.3)
+
+    @pytest.mark.fail_slow(10)
+    def test_fs_param(self):
+        for norm in ('phase', 'mag', 'delay'):
+            for fs in (900, 900.1, 1234.567):
+                for N in (0, 1, 2, 3, 10):
+                    for fc in (100, 100.1, 432.12345):
+                        for btype in ('lp', 'hp'):
+                            ba1 = bessel(N, fc, btype, norm=norm, fs=fs)
+                            ba2 = bessel(N, fc/(fs/2), btype, norm=norm)
+                            for ba1_, ba2_ in zip(ba1, ba2):
+                                xp_assert_close(ba1_, ba2_)
+                    for fc in ((100, 200), (100.1, 200.2), (321.123, 432.123)):
+                        for btype in ('bp', 'bs'):
+                            ba1 = bessel(N, fc, btype, norm=norm, fs=fs)
+                            for seq in (list, tuple, array):
+                                fcnorm = seq([f/(fs/2) for f in fc])
+                                ba2 = bessel(N, fcnorm, btype, norm=norm)
+                                for ba1_, ba2_ in zip(ba1, ba2):
+                                    xp_assert_close(ba1_, ba2_)
+
+
+class TestButter:
+
+    def test_degenerate(self):
+        # 0-order filter is just a passthrough
+        b, a = butter(0, 1, analog=True)
+        xp_assert_equal(b, np.asarray([1.0]))
+        xp_assert_equal(a, np.asarray([1.0]))
+
+        # 1-order filter is same for all types
+        b, a = butter(1, 1, analog=True)
+        assert_array_almost_equal(b, [1])
+        assert_array_almost_equal(a, [1, 1])
+
+        z, p, k = butter(1, 0.3, output='zpk')
+        xp_assert_equal(z, np.asarray([-1.0]))
+        xp_assert_close(p, [3.249196962329063e-01 + 0j], rtol=1e-14)
+        xp_assert_close(k, 3.375401518835469e-01, rtol=1e-14)
+
+    def test_basic(self):
+        # analog s-plane
+        for N in range(25):
+            wn = 0.01
+            z, p, k = butter(N, wn, 'low', analog=True, output='zpk')
+            assert_array_almost_equal([], z)
+            assert len(p) == N
+            # All poles should be at distance wn from origin
+            assert_array_almost_equal(abs(p), np.asarray(wn))
+            assert all(np.real(p) <= 0)  # No poles in right half of S-plane
+            assert_array_almost_equal(wn**N, k)
+
+        # digital z-plane
+        for N in range(25):
+            wn = 0.01
+            z, p, k = butter(N, wn, 'high', analog=False, output='zpk')
+            xp_assert_equal(np.ones(N), z)  # All zeros exactly at DC
+            assert all(np.abs(p) <= 1)  # No poles outside unit circle
+
+        b1, a1 = butter(2, 1, analog=True)
+        assert_array_almost_equal(b1, [1])
+        assert_array_almost_equal(a1, [1, np.sqrt(2), 1])
+
+        b2, a2 = butter(5, 1, analog=True)
+        assert_array_almost_equal(b2, [1])
+        assert_array_almost_equal(a2, [1, 3.2361, 5.2361,
+                                       5.2361, 3.2361, 1], decimal=4)
+
+        b3, a3 = butter(10, 1, analog=True)
+        assert_array_almost_equal(b3, [1])
+        assert_array_almost_equal(a3, [1, 6.3925, 20.4317, 42.8021, 64.8824,
+                                       74.2334, 64.8824, 42.8021, 20.4317,
+                                       6.3925, 1], decimal=4)
+
+        b2, a2 = butter(19, 1.0441379169150726, analog=True)
+        assert_array_almost_equal(b2, [2.2720], decimal=4)
+        assert_array_almost_equal(a2, 1.0e+004 * np.array([
+                        0.0001, 0.0013, 0.0080, 0.0335, 0.1045, 0.2570,
+                        0.5164, 0.8669, 1.2338, 1.5010, 1.5672, 1.4044,
+                        1.0759, 0.6986, 0.3791, 0.1681, 0.0588, 0.0153,
+                        0.0026, 0.0002]), decimal=0)
+
+        b, a = butter(5, 0.4)
+        assert_array_almost_equal(b, [0.0219, 0.1097, 0.2194,
+                                      0.2194, 0.1097, 0.0219], decimal=4)
+        assert_array_almost_equal(a, [1.0000, -0.9853, 0.9738,
+                                      -0.3864, 0.1112, -0.0113], decimal=4)
+
+    def test_highpass(self):
+        # highpass, high even order
+        z, p, k = butter(28, 0.43, 'high', output='zpk')
+        z2 = np.ones(28)
+        p2 = [
+            2.068257195514592e-01 + 9.238294351481734e-01j,
+            2.068257195514592e-01 - 9.238294351481734e-01j,
+            1.874933103892023e-01 + 8.269455076775277e-01j,
+            1.874933103892023e-01 - 8.269455076775277e-01j,
+            1.717435567330153e-01 + 7.383078571194629e-01j,
+            1.717435567330153e-01 - 7.383078571194629e-01j,
+            1.588266870755982e-01 + 6.564623730651094e-01j,
+            1.588266870755982e-01 - 6.564623730651094e-01j,
+            1.481881532502603e-01 + 5.802343458081779e-01j,
+            1.481881532502603e-01 - 5.802343458081779e-01j,
+            1.394122576319697e-01 + 5.086609000582009e-01j,
+            1.394122576319697e-01 - 5.086609000582009e-01j,
+            1.321840881809715e-01 + 4.409411734716436e-01j,
+            1.321840881809715e-01 - 4.409411734716436e-01j,
+            1.262633413354405e-01 + 3.763990035551881e-01j,
+            1.262633413354405e-01 - 3.763990035551881e-01j,
+            1.214660449478046e-01 + 3.144545234797277e-01j,
+            1.214660449478046e-01 - 3.144545234797277e-01j,
+            1.104868766650320e-01 + 2.771505404367791e-02j,
+            1.104868766650320e-01 - 2.771505404367791e-02j,
+            1.111768629525075e-01 + 8.331369153155753e-02j,
+            1.111768629525075e-01 - 8.331369153155753e-02j,
+            1.125740630842972e-01 + 1.394219509611784e-01j,
+            1.125740630842972e-01 - 1.394219509611784e-01j,
+            1.147138487992747e-01 + 1.963932363793666e-01j,
+            1.147138487992747e-01 - 1.963932363793666e-01j,
+            1.176516491045901e-01 + 2.546021573417188e-01j,
+            1.176516491045901e-01 - 2.546021573417188e-01j,
+            ]
+        k2 = 1.446671081817286e-06
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-7)
+        xp_assert_close(k, k2, rtol=1e-10)
+
+        # highpass, high odd order
+        z, p, k = butter(27, 0.56, 'high', output='zpk')
+        z2 = np.ones(27)
+        p2 = [
+            -1.772572785680147e-01 + 9.276431102995948e-01j,
+            -1.772572785680147e-01 - 9.276431102995948e-01j,
+            -1.600766565322114e-01 + 8.264026279893268e-01j,
+            -1.600766565322114e-01 - 8.264026279893268e-01j,
+            -1.461948419016121e-01 + 7.341841939120078e-01j,
+            -1.461948419016121e-01 - 7.341841939120078e-01j,
+            -1.348975284762046e-01 + 6.493235066053785e-01j,
+            -1.348975284762046e-01 - 6.493235066053785e-01j,
+            -1.256628210712206e-01 + 5.704921366889227e-01j,
+            -1.256628210712206e-01 - 5.704921366889227e-01j,
+            -1.181038235962314e-01 + 4.966120551231630e-01j,
+            -1.181038235962314e-01 - 4.966120551231630e-01j,
+            -1.119304913239356e-01 + 4.267938916403775e-01j,
+            -1.119304913239356e-01 - 4.267938916403775e-01j,
+            -1.069237739782691e-01 + 3.602914879527338e-01j,
+            -1.069237739782691e-01 - 3.602914879527338e-01j,
+            -1.029178030691416e-01 + 2.964677964142126e-01j,
+            -1.029178030691416e-01 - 2.964677964142126e-01j,
+            -9.978747500816100e-02 + 2.347687643085738e-01j,
+            -9.978747500816100e-02 - 2.347687643085738e-01j,
+            -9.743974496324025e-02 + 1.747028739092479e-01j,
+            -9.743974496324025e-02 - 1.747028739092479e-01j,
+            -9.580754551625957e-02 + 1.158246860771989e-01j,
+            -9.580754551625957e-02 - 1.158246860771989e-01j,
+            -9.484562207782568e-02 + 5.772118357151691e-02j,
+            -9.484562207782568e-02 - 5.772118357151691e-02j,
+            -9.452783117928215e-02
+            ]
+        k2 = 9.585686688851069e-09
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-8)
+        xp_assert_close(k, k2)
+
+    def test_bandpass(self):
+        z, p, k = butter(8, [0.25, 0.33], 'band', output='zpk')
+        z2 = [1, 1, 1, 1, 1, 1, 1, 1,
+              -1, -1, -1, -1, -1, -1, -1, -1]
+        p2 = [
+            4.979909925436156e-01 + 8.367609424799387e-01j,
+            4.979909925436156e-01 - 8.367609424799387e-01j,
+            4.913338722555539e-01 + 7.866774509868817e-01j,
+            4.913338722555539e-01 - 7.866774509868817e-01j,
+            5.035229361778706e-01 + 7.401147376726750e-01j,
+            5.035229361778706e-01 - 7.401147376726750e-01j,
+            5.307617160406101e-01 + 7.029184459442954e-01j,
+            5.307617160406101e-01 - 7.029184459442954e-01j,
+            5.680556159453138e-01 + 6.788228792952775e-01j,
+            5.680556159453138e-01 - 6.788228792952775e-01j,
+            6.100962560818854e-01 + 6.693849403338664e-01j,
+            6.100962560818854e-01 - 6.693849403338664e-01j,
+            6.904694312740631e-01 + 6.930501690145245e-01j,
+            6.904694312740631e-01 - 6.930501690145245e-01j,
+            6.521767004237027e-01 + 6.744414640183752e-01j,
+            6.521767004237027e-01 - 6.744414640183752e-01j,
+            ]
+        k2 = 3.398854055800844e-08
+        xp_assert_equal(z, z2, check_dtype=False)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-13)
+        xp_assert_close(k, k2, rtol=1e-13)
+
+        # bandpass analog
+        z, p, k = butter(4, [90.5, 110.5], 'bp', analog=True, output='zpk')
+        z2 = np.zeros(4, dtype=z.dtype)
+        p2 = [
+            -4.179137760733086e+00 + 1.095935899082837e+02j,
+            -4.179137760733086e+00 - 1.095935899082837e+02j,
+            -9.593598668443835e+00 + 1.034745398029734e+02j,
+            -9.593598668443835e+00 - 1.034745398029734e+02j,
+            -8.883991981781929e+00 + 9.582087115567160e+01j,
+            -8.883991981781929e+00 - 9.582087115567160e+01j,
+            -3.474530886568715e+00 + 9.111599925805801e+01j,
+            -3.474530886568715e+00 - 9.111599925805801e+01j,
+            ]
+        k2 = 1.600000000000001e+05
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag))
+        xp_assert_close(k, k2, rtol=1e-15)
+
+    def test_bandstop(self):
+        z, p, k = butter(7, [0.45, 0.56], 'stop', output='zpk')
+        z2 = [-1.594474531383421e-02 + 9.998728744679880e-01j,
+              -1.594474531383421e-02 - 9.998728744679880e-01j,
+              -1.594474531383421e-02 + 9.998728744679880e-01j,
+              -1.594474531383421e-02 - 9.998728744679880e-01j,
+              -1.594474531383421e-02 + 9.998728744679880e-01j,
+              -1.594474531383421e-02 - 9.998728744679880e-01j,
+              -1.594474531383421e-02 + 9.998728744679880e-01j,
+              -1.594474531383421e-02 - 9.998728744679880e-01j,
+              -1.594474531383421e-02 + 9.998728744679880e-01j,
+              -1.594474531383421e-02 - 9.998728744679880e-01j,
+              -1.594474531383421e-02 + 9.998728744679880e-01j,
+              -1.594474531383421e-02 - 9.998728744679880e-01j,
+              -1.594474531383421e-02 + 9.998728744679880e-01j,
+              -1.594474531383421e-02 - 9.998728744679880e-01j]
+        p2 = [-1.766850742887729e-01 + 9.466951258673900e-01j,
+              -1.766850742887729e-01 - 9.466951258673900e-01j,
+               1.467897662432886e-01 + 9.515917126462422e-01j,
+               1.467897662432886e-01 - 9.515917126462422e-01j,
+              -1.370083529426906e-01 + 8.880376681273993e-01j,
+              -1.370083529426906e-01 - 8.880376681273993e-01j,
+               1.086774544701390e-01 + 8.915240810704319e-01j,
+               1.086774544701390e-01 - 8.915240810704319e-01j,
+              -7.982704457700891e-02 + 8.506056315273435e-01j,
+              -7.982704457700891e-02 - 8.506056315273435e-01j,
+               5.238812787110331e-02 + 8.524011102699969e-01j,
+               5.238812787110331e-02 - 8.524011102699969e-01j,
+              -1.357545000491310e-02 + 8.382287744986582e-01j,
+              -1.357545000491310e-02 - 8.382287744986582e-01j]
+        k2 = 4.577122512960063e-01
+        xp_assert_close(sorted(z, key=np.imag),
+                        sorted(z2, key=np.imag))
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag))
+        xp_assert_close(k, k2, rtol=1e-14)
+
+    def test_ba_output(self):
+        b, a = butter(4, [100, 300], 'bandpass', analog=True)
+        b2 = [1.6e+09, 0, 0, 0, 0]
+        a2 = [1.000000000000000e+00, 5.226251859505511e+02,
+              2.565685424949238e+05, 6.794127417357160e+07,
+              1.519411254969542e+10, 2.038238225207147e+12,
+              2.309116882454312e+14, 1.411088002066486e+16,
+              8.099999999999991e+17]
+        xp_assert_close(b, b2, rtol=1e-14)
+        xp_assert_close(a, a2, rtol=1e-14)
+
+    def test_fs_param(self):
+        for fs in (900, 900.1, 1234.567):
+            for N in (0, 1, 2, 3, 10):
+                for fc in (100, 100.1, 432.12345):
+                    for btype in ('lp', 'hp'):
+                        ba1 = butter(N, fc, btype, fs=fs)
+                        ba2 = butter(N, fc/(fs/2), btype)
+                        for ba1_, ba2_ in zip(ba1, ba2):
+                            xp_assert_close(ba1_, ba2_)
+                for fc in ((100, 200), (100.1, 200.2), (321.123, 432.123)):
+                    for btype in ('bp', 'bs'):
+                        ba1 = butter(N, fc, btype, fs=fs)
+                        for seq in (list, tuple, array):
+                            fcnorm = seq([f/(fs/2) for f in fc])
+                            ba2 = butter(N, fcnorm, btype)
+                            for ba1_, ba2_ in zip(ba1, ba2):
+                                xp_assert_close(ba1_, ba2_)
+
+
+class TestCheby1:
+
+    def test_degenerate(self):
+        # 0-order filter is just a passthrough
+        # Even-order filters have DC gain of -rp dB
+        b, a = cheby1(0, 10*np.log10(2), 1, analog=True)
+        assert_array_almost_equal(b, [1/np.sqrt(2)])
+        xp_assert_equal(a, np.asarray([1.0]))
+
+        # 1-order filter is same for all types
+        b, a = cheby1(1, 10*np.log10(2), 1, analog=True)
+        assert_array_almost_equal(b, [1])
+        assert_array_almost_equal(a, [1, 1])
+
+        z, p, k = cheby1(1, 0.1, 0.3, output='zpk')
+        xp_assert_equal(z, np.asarray([-1.0]))
+        xp_assert_close(p, [-5.390126972799615e-01 + 0j], rtol=1e-14)
+        xp_assert_close(k, 7.695063486399808e-01, rtol=1e-14)
+
+    def test_basic(self):
+        for N in range(25):
+            wn = 0.01
+            z, p, k = cheby1(N, 1, wn, 'low', analog=True, output='zpk')
+            assert_array_almost_equal([], z)
+            assert len(p) == N
+            assert all(np.real(p) <= 0)  # No poles in right half of S-plane
+
+        for N in range(25):
+            wn = 0.01
+            z, p, k = cheby1(N, 1, wn, 'high', analog=False, output='zpk')
+            xp_assert_equal(np.ones(N), z)  # All zeros exactly at DC
+            assert all(np.abs(p) <= 1)  # No poles outside unit circle
+
+        # Same test as TestNormalize
+        b, a = cheby1(8, 0.5, 0.048)
+        assert_array_almost_equal(b, [
+                             2.150733144728282e-11, 1.720586515782626e-10,
+                             6.022052805239190e-10, 1.204410561047838e-09,
+                             1.505513201309798e-09, 1.204410561047838e-09,
+                             6.022052805239190e-10, 1.720586515782626e-10,
+                             2.150733144728282e-11], decimal=14)
+        assert_array_almost_equal(a, [
+                             1.000000000000000e+00, -7.782402035027959e+00,
+                             2.654354569747454e+01, -5.182182531666387e+01,
+                             6.334127355102684e+01, -4.963358186631157e+01,
+                             2.434862182949389e+01, -6.836925348604676e+00,
+                             8.412934944449140e-01], decimal=14)
+
+        b, a = cheby1(4, 1, [0.4, 0.7], btype='band')
+        assert_array_almost_equal(b, [0.0084, 0, -0.0335, 0, 0.0502, 0,
+                                      -0.0335, 0, 0.0084], decimal=4)
+        assert_array_almost_equal(a, [1.0, 1.1191, 2.862, 2.2986, 3.4137,
+                                      1.8653, 1.8982, 0.5676, 0.4103],
+                                  decimal=4)
+
+        b2, a2 = cheby1(5, 3, 1, analog=True)
+        assert_array_almost_equal(b2, [0.0626], decimal=4)
+        assert_array_almost_equal(a2, [1, 0.5745, 1.4150, 0.5489, 0.4080,
+                                       0.0626], decimal=4)
+
+        b, a = cheby1(8, 0.5, 0.1)
+        assert_array_almost_equal(b, 1.0e-006 * np.array([
+            0.00703924326028, 0.05631394608227, 0.19709881128793,
+            0.39419762257586, 0.49274702821983, 0.39419762257586,
+            0.19709881128793, 0.05631394608227, 0.00703924326028]),
+            decimal=13)
+        assert_array_almost_equal(a, [
+              1.00000000000000, -7.44912258934158, 24.46749067762108,
+              -46.27560200466141, 55.11160187999928, -42.31640010161038,
+              20.45543300484147, -5.69110270561444, 0.69770374759022],
+            decimal=13)
+
+        b, a = cheby1(8, 0.5, 0.25)
+        assert_array_almost_equal(b, 1.0e-003 * np.array([
+            0.00895261138923, 0.07162089111382, 0.25067311889837,
+            0.50134623779673, 0.62668279724591, 0.50134623779673,
+            0.25067311889837, 0.07162089111382, 0.00895261138923]),
+            decimal=13)
+        assert_array_almost_equal(a, [1.00000000000000, -5.97529229188545,
+                                      16.58122329202101, -27.71423273542923,
+                                      30.39509758355313, -22.34729670426879,
+                                      10.74509800434910, -3.08924633697497,
+                                      0.40707685889802], decimal=13)
+
+    def test_highpass(self):
+        # high even order
+        z, p, k = cheby1(24, 0.7, 0.2, 'high', output='zpk')
+        z2 = np.ones(24)
+        p2 = [-6.136558509657073e-01 + 2.700091504942893e-01j,
+              -6.136558509657073e-01 - 2.700091504942893e-01j,
+              -3.303348340927516e-01 + 6.659400861114254e-01j,
+              -3.303348340927516e-01 - 6.659400861114254e-01j,
+              8.779713780557169e-03 + 8.223108447483040e-01j,
+              8.779713780557169e-03 - 8.223108447483040e-01j,
+              2.742361123006911e-01 + 8.356666951611864e-01j,
+              2.742361123006911e-01 - 8.356666951611864e-01j,
+              4.562984557158206e-01 + 7.954276912303594e-01j,
+              4.562984557158206e-01 - 7.954276912303594e-01j,
+              5.777335494123628e-01 + 7.435821817961783e-01j,
+              5.777335494123628e-01 - 7.435821817961783e-01j,
+              6.593260977749194e-01 + 6.955390907990932e-01j,
+              6.593260977749194e-01 - 6.955390907990932e-01j,
+              7.149590948466562e-01 + 6.559437858502012e-01j,
+              7.149590948466562e-01 - 6.559437858502012e-01j,
+              7.532432388188739e-01 + 6.256158042292060e-01j,
+              7.532432388188739e-01 - 6.256158042292060e-01j,
+              7.794365244268271e-01 + 6.042099234813333e-01j,
+              7.794365244268271e-01 - 6.042099234813333e-01j,
+              7.967253874772997e-01 + 5.911966597313203e-01j,
+              7.967253874772997e-01 - 5.911966597313203e-01j,
+              8.069756417293870e-01 + 5.862214589217275e-01j,
+              8.069756417293870e-01 - 5.862214589217275e-01j]
+        k2 = 6.190427617192018e-04
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-10)
+        xp_assert_close(k, k2, rtol=1e-10)
+
+        # high odd order
+        z, p, k = cheby1(23, 0.8, 0.3, 'high', output='zpk')
+        z2 = np.ones(23)
+        p2 = [-7.676400532011010e-01,
+              -6.754621070166477e-01 + 3.970502605619561e-01j,
+              -6.754621070166477e-01 - 3.970502605619561e-01j,
+              -4.528880018446727e-01 + 6.844061483786332e-01j,
+              -4.528880018446727e-01 - 6.844061483786332e-01j,
+              -1.986009130216447e-01 + 8.382285942941594e-01j,
+              -1.986009130216447e-01 - 8.382285942941594e-01j,
+              2.504673931532608e-02 + 8.958137635794080e-01j,
+              2.504673931532608e-02 - 8.958137635794080e-01j,
+              2.001089429976469e-01 + 9.010678290791480e-01j,
+              2.001089429976469e-01 - 9.010678290791480e-01j,
+              3.302410157191755e-01 + 8.835444665962544e-01j,
+              3.302410157191755e-01 - 8.835444665962544e-01j,
+              4.246662537333661e-01 + 8.594054226449009e-01j,
+              4.246662537333661e-01 - 8.594054226449009e-01j,
+              4.919620928120296e-01 + 8.366772762965786e-01j,
+              4.919620928120296e-01 - 8.366772762965786e-01j,
+              5.385746917494749e-01 + 8.191616180796720e-01j,
+              5.385746917494749e-01 - 8.191616180796720e-01j,
+              5.855636993537203e-01 + 8.060680937701062e-01j,
+              5.855636993537203e-01 - 8.060680937701062e-01j,
+              5.688812849391721e-01 + 8.086497795114683e-01j,
+              5.688812849391721e-01 - 8.086497795114683e-01j]
+        k2 = 1.941697029206324e-05
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-10)
+        xp_assert_close(k, k2, rtol=1e-10)
+
+        z, p, k = cheby1(10, 1, 1000, 'high', analog=True, output='zpk')
+        z2 = np.zeros(10)
+        p2 = [-3.144743169501551e+03 + 3.511680029092744e+03j,
+              -3.144743169501551e+03 - 3.511680029092744e+03j,
+              -5.633065604514602e+02 + 2.023615191183945e+03j,
+              -5.633065604514602e+02 - 2.023615191183945e+03j,
+              -1.946412183352025e+02 + 1.372309454274755e+03j,
+              -1.946412183352025e+02 - 1.372309454274755e+03j,
+              -7.987162953085479e+01 + 1.105207708045358e+03j,
+              -7.987162953085479e+01 - 1.105207708045358e+03j,
+              -2.250315039031946e+01 + 1.001723931471477e+03j,
+              -2.250315039031946e+01 - 1.001723931471477e+03j]
+        k2 = 8.912509381337453e-01
+        xp_assert_equal(z, z2)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-13)
+        xp_assert_close(k, k2, rtol=1e-15)
+
+    def test_bandpass(self):
+        z, p, k = cheby1(8, 1, [0.3, 0.4], 'bp', output='zpk')
+        z2 = [1, 1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1]
+        p2 = [3.077784854851463e-01 + 9.453307017592942e-01j,
+              3.077784854851463e-01 - 9.453307017592942e-01j,
+              3.280567400654425e-01 + 9.272377218689016e-01j,
+              3.280567400654425e-01 - 9.272377218689016e-01j,
+              3.677912763284301e-01 + 9.038008865279966e-01j,
+              3.677912763284301e-01 - 9.038008865279966e-01j,
+              4.194425632520948e-01 + 8.769407159656157e-01j,
+              4.194425632520948e-01 - 8.769407159656157e-01j,
+              4.740921994669189e-01 + 8.496508528630974e-01j,
+              4.740921994669189e-01 - 8.496508528630974e-01j,
+              5.234866481897429e-01 + 8.259608422808477e-01j,
+              5.234866481897429e-01 - 8.259608422808477e-01j,
+              5.844717632289875e-01 + 8.052901363500210e-01j,
+              5.844717632289875e-01 - 8.052901363500210e-01j,
+              5.615189063336070e-01 + 8.100667803850766e-01j,
+              5.615189063336070e-01 - 8.100667803850766e-01j]
+        k2 = 5.007028718074307e-09
+        xp_assert_equal(z, z2, check_dtype=False)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-13)
+        xp_assert_close(k, k2, rtol=1e-13)
+
+    def test_bandstop(self):
+        z, p, k = cheby1(7, 1, [0.5, 0.6], 'stop', output='zpk')
+        z2 = [-1.583844403245361e-01 + 9.873775210440450e-01j,
+              -1.583844403245361e-01 - 9.873775210440450e-01j,
+              -1.583844403245361e-01 + 9.873775210440450e-01j,
+              -1.583844403245361e-01 - 9.873775210440450e-01j,
+              -1.583844403245361e-01 + 9.873775210440450e-01j,
+              -1.583844403245361e-01 - 9.873775210440450e-01j,
+              -1.583844403245361e-01 + 9.873775210440450e-01j,
+              -1.583844403245361e-01 - 9.873775210440450e-01j,
+              -1.583844403245361e-01 + 9.873775210440450e-01j,
+              -1.583844403245361e-01 - 9.873775210440450e-01j,
+              -1.583844403245361e-01 + 9.873775210440450e-01j,
+              -1.583844403245361e-01 - 9.873775210440450e-01j,
+              -1.583844403245361e-01 + 9.873775210440450e-01j,
+              -1.583844403245361e-01 - 9.873775210440450e-01j]
+        p2 = [-8.942974551472813e-02 + 3.482480481185926e-01j,
+              -8.942974551472813e-02 - 3.482480481185926e-01j,
+               1.293775154041798e-01 + 8.753499858081858e-01j,
+               1.293775154041798e-01 - 8.753499858081858e-01j,
+               3.399741945062013e-02 + 9.690316022705607e-01j,
+               3.399741945062013e-02 - 9.690316022705607e-01j,
+               4.167225522796539e-04 + 9.927338161087488e-01j,
+               4.167225522796539e-04 - 9.927338161087488e-01j,
+              -3.912966549550960e-01 + 8.046122859255742e-01j,
+              -3.912966549550960e-01 - 8.046122859255742e-01j,
+              -3.307805547127368e-01 + 9.133455018206508e-01j,
+              -3.307805547127368e-01 - 9.133455018206508e-01j,
+              -3.072658345097743e-01 + 9.443589759799366e-01j,
+              -3.072658345097743e-01 - 9.443589759799366e-01j]
+        k2 = 3.619438310405028e-01
+        xp_assert_close(sorted(z, key=np.imag),
+                        sorted(z2, key=np.imag), rtol=1e-13)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-13)
+        xp_assert_close(k, k2, rtol=0, atol=5e-16)
+
+    def test_ba_output(self):
+        # with transfer function conversion,  without digital conversion
+        b, a = cheby1(5, 0.9, [210, 310], 'stop', analog=True)
+        b2 = [1.000000000000006e+00, 0,
+              3.255000000000020e+05, 0,
+              4.238010000000026e+10, 0,
+              2.758944510000017e+15, 0,
+              8.980364380050052e+19, 0,
+              1.169243442282517e+24
+              ]
+        a2 = [1.000000000000000e+00, 4.630555945694342e+02,
+              4.039266454794788e+05, 1.338060988610237e+08,
+              5.844333551294591e+10, 1.357346371637638e+13,
+              3.804661141892782e+15, 5.670715850340080e+17,
+              1.114411200988328e+20, 8.316815934908471e+21,
+              1.169243442282517e+24
+              ]
+        xp_assert_close(b, b2, rtol=1e-14)
+        xp_assert_close(a, a2, rtol=1e-14)
+
+    def test_fs_param(self):
+        for fs in (900, 900.1, 1234.567):
+            for N in (0, 1, 2, 3, 10):
+                for fc in (100, 100.1, 432.12345):
+                    for btype in ('lp', 'hp'):
+                        ba1 = cheby1(N, 1, fc, btype, fs=fs)
+                        ba2 = cheby1(N, 1, fc/(fs/2), btype)
+                        for ba1_, ba2_ in zip(ba1, ba2):
+                            xp_assert_close(ba1_, ba2_)
+                for fc in ((100, 200), (100.1, 200.2), (321.123, 432.123)):
+                    for btype in ('bp', 'bs'):
+                        ba1 = cheby1(N, 1, fc, btype, fs=fs)
+                        for seq in (list, tuple, array):
+                            fcnorm = seq([f/(fs/2) for f in fc])
+                            ba2 = cheby1(N, 1, fcnorm, btype)
+                            for ba1_, ba2_ in zip(ba1, ba2):
+                                xp_assert_close(ba1_, ba2_)
+
+class TestCheby2:
+
+    def test_degenerate(self):
+        # 0-order filter is just a passthrough
+        # Stopband ripple factor doesn't matter
+        b, a = cheby2(0, 123.456, 1, analog=True)
+        xp_assert_equal(b, np.asarray([1.0]))
+        xp_assert_equal(a, np.asarray([1.0]))
+
+        # 1-order filter is same for all types
+        b, a = cheby2(1, 10*np.log10(2), 1, analog=True)
+        assert_array_almost_equal(b, [1])
+        assert_array_almost_equal(a, [1, 1])
+
+        z, p, k = cheby2(1, 50, 0.3, output='zpk')
+        xp_assert_equal(z, np.asarray([-1], dtype=np.complex128))
+        xp_assert_close(p, [9.967826460175649e-01 + 0j], rtol=1e-14)
+        xp_assert_close(k, 1.608676991217512e-03, rtol=1e-14)
+
+    def test_basic(self):
+        for N in range(25):
+            wn = 0.01
+            z, p, k = cheby2(N, 40, wn, 'low', analog=True, output='zpk')
+            assert len(p) == N
+            assert all(np.real(p) <= 0)  # No poles in right half of S-plane
+
+        for N in range(25):
+            wn = 0.01
+            z, p, k = cheby2(N, 40, wn, 'high', analog=False, output='zpk')
+            assert all(np.abs(p) <= 1)  # No poles outside unit circle
+
+        B, A = cheby2(18, 100, 0.5)
+        assert_array_almost_equal(B, [
+            0.00167583914216, 0.01249479541868, 0.05282702120282,
+            0.15939804265706, 0.37690207631117, 0.73227013789108,
+            1.20191856962356, 1.69522872823393, 2.07598674519837,
+            2.21972389625291, 2.07598674519838, 1.69522872823395,
+            1.20191856962359, 0.73227013789110, 0.37690207631118,
+            0.15939804265707, 0.05282702120282, 0.01249479541868,
+            0.00167583914216], decimal=13)
+        assert_array_almost_equal(A, [
+            1.00000000000000, -0.27631970006174, 3.19751214254060,
+            -0.15685969461355, 4.13926117356269, 0.60689917820044,
+            2.95082770636540, 0.89016501910416, 1.32135245849798,
+            0.51502467236824, 0.38906643866660, 0.15367372690642,
+            0.07255803834919, 0.02422454070134, 0.00756108751837,
+            0.00179848550988, 0.00033713574499, 0.00004258794833,
+            0.00000281030149], decimal=13)
+
+    def test_highpass(self):
+        # high even order
+        z, p, k = cheby2(26, 60, 0.3, 'high', output='zpk')
+        z2 = [9.981088955489852e-01 + 6.147058341984388e-02j,
+              9.981088955489852e-01 - 6.147058341984388e-02j,
+              9.832702870387426e-01 + 1.821525257215483e-01j,
+              9.832702870387426e-01 - 1.821525257215483e-01j,
+              9.550760158089112e-01 + 2.963609353922882e-01j,
+              9.550760158089112e-01 - 2.963609353922882e-01j,
+              9.162054748821922e-01 + 4.007087817803773e-01j,
+              9.162054748821922e-01 - 4.007087817803773e-01j,
+              8.700619897368064e-01 + 4.929423232136168e-01j,
+              8.700619897368064e-01 - 4.929423232136168e-01j,
+              5.889791753434985e-01 + 8.081482110427953e-01j,
+              5.889791753434985e-01 - 8.081482110427953e-01j,
+              5.984900456570295e-01 + 8.011302423760501e-01j,
+              5.984900456570295e-01 - 8.011302423760501e-01j,
+              6.172880888914629e-01 + 7.867371958365343e-01j,
+              6.172880888914629e-01 - 7.867371958365343e-01j,
+              6.448899971038180e-01 + 7.642754030030161e-01j,
+              6.448899971038180e-01 - 7.642754030030161e-01j,
+              6.804845629637927e-01 + 7.327624168637228e-01j,
+              6.804845629637927e-01 - 7.327624168637228e-01j,
+              8.202619107108660e-01 + 5.719881098737678e-01j,
+              8.202619107108660e-01 - 5.719881098737678e-01j,
+              7.228410452536148e-01 + 6.910143437705678e-01j,
+              7.228410452536148e-01 - 6.910143437705678e-01j,
+              7.702121399578629e-01 + 6.377877856007792e-01j,
+              7.702121399578629e-01 - 6.377877856007792e-01j]
+        p2 = [7.365546198286450e-01 + 4.842085129329526e-02j,
+              7.365546198286450e-01 - 4.842085129329526e-02j,
+              7.292038510962885e-01 + 1.442201672097581e-01j,
+              7.292038510962885e-01 - 1.442201672097581e-01j,
+              7.151293788040354e-01 + 2.369925800458584e-01j,
+              7.151293788040354e-01 - 2.369925800458584e-01j,
+              6.955051820787286e-01 + 3.250341363856910e-01j,
+              6.955051820787286e-01 - 3.250341363856910e-01j,
+              6.719122956045220e-01 + 4.070475750638047e-01j,
+              6.719122956045220e-01 - 4.070475750638047e-01j,
+              6.461722130611300e-01 + 4.821965916689270e-01j,
+              6.461722130611300e-01 - 4.821965916689270e-01j,
+              5.528045062872224e-01 + 8.162920513838372e-01j,
+              5.528045062872224e-01 - 8.162920513838372e-01j,
+              5.464847782492791e-01 + 7.869899955967304e-01j,
+              5.464847782492791e-01 - 7.869899955967304e-01j,
+              5.488033111260949e-01 + 7.520442354055579e-01j,
+              5.488033111260949e-01 - 7.520442354055579e-01j,
+              6.201874719022955e-01 + 5.500894392527353e-01j,
+              6.201874719022955e-01 - 5.500894392527353e-01j,
+              5.586478152536709e-01 + 7.112676877332921e-01j,
+              5.586478152536709e-01 - 7.112676877332921e-01j,
+              5.958145844148228e-01 + 6.107074340842115e-01j,
+              5.958145844148228e-01 - 6.107074340842115e-01j,
+              5.747812938519067e-01 + 6.643001536914696e-01j,
+              5.747812938519067e-01 - 6.643001536914696e-01j]
+        k2 = 9.932997786497189e-02
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-13)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-12)
+        xp_assert_close(k, k2, rtol=1e-11)
+
+        # high odd order
+        z, p, k = cheby2(25, 80, 0.5, 'high', output='zpk')
+        z2 = [9.690690376586687e-01 + 2.467897896011971e-01j,
+              9.690690376586687e-01 - 2.467897896011971e-01j,
+              9.999999999999492e-01,
+              8.835111277191199e-01 + 4.684101698261429e-01j,
+              8.835111277191199e-01 - 4.684101698261429e-01j,
+              7.613142857900539e-01 + 6.483830335935022e-01j,
+              7.613142857900539e-01 - 6.483830335935022e-01j,
+              6.232625173626231e-01 + 7.820126817709752e-01j,
+              6.232625173626231e-01 - 7.820126817709752e-01j,
+              4.864456563413621e-01 + 8.737108351316745e-01j,
+              4.864456563413621e-01 - 8.737108351316745e-01j,
+              3.618368136816749e-01 + 9.322414495530347e-01j,
+              3.618368136816749e-01 - 9.322414495530347e-01j,
+              2.549486883466794e-01 + 9.669545833752675e-01j,
+              2.549486883466794e-01 - 9.669545833752675e-01j,
+              1.676175432109457e-01 + 9.858520980390212e-01j,
+              1.676175432109457e-01 - 9.858520980390212e-01j,
+              1.975218468277521e-03 + 9.999980492540941e-01j,
+              1.975218468277521e-03 - 9.999980492540941e-01j,
+              1.786959496651858e-02 + 9.998403260399917e-01j,
+              1.786959496651858e-02 - 9.998403260399917e-01j,
+              9.967933660557139e-02 + 9.950196127985684e-01j,
+              9.967933660557139e-02 - 9.950196127985684e-01j,
+              5.013970951219547e-02 + 9.987422137518890e-01j,
+              5.013970951219547e-02 - 9.987422137518890e-01j]
+        p2 = [4.218866331906864e-01,
+              4.120110200127552e-01 + 1.361290593621978e-01j,
+              4.120110200127552e-01 - 1.361290593621978e-01j,
+              3.835890113632530e-01 + 2.664910809911026e-01j,
+              3.835890113632530e-01 - 2.664910809911026e-01j,
+              3.399195570456499e-01 + 3.863983538639875e-01j,
+              3.399195570456499e-01 - 3.863983538639875e-01j,
+              2.855977834508353e-01 + 4.929444399540688e-01j,
+              2.855977834508353e-01 - 4.929444399540688e-01j,
+              2.255765441339322e-01 + 5.851631870205766e-01j,
+              2.255765441339322e-01 - 5.851631870205766e-01j,
+              1.644087535815792e-01 + 6.637356937277153e-01j,
+              1.644087535815792e-01 - 6.637356937277153e-01j,
+              -7.293633845273095e-02 + 9.739218252516307e-01j,
+              -7.293633845273095e-02 - 9.739218252516307e-01j,
+              1.058259206358626e-01 + 7.304739464862978e-01j,
+              1.058259206358626e-01 - 7.304739464862978e-01j,
+              -5.703971947785402e-02 + 9.291057542169088e-01j,
+              -5.703971947785402e-02 - 9.291057542169088e-01j,
+              5.263875132656864e-02 + 7.877974334424453e-01j,
+              5.263875132656864e-02 - 7.877974334424453e-01j,
+              -3.007943405982616e-02 + 8.846331716180016e-01j,
+              -3.007943405982616e-02 - 8.846331716180016e-01j,
+              6.857277464483946e-03 + 8.383275456264492e-01j,
+              6.857277464483946e-03 - 8.383275456264492e-01j]
+        k2 = 6.507068761705037e-03
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-13)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-12)
+        xp_assert_close(k, k2, rtol=1e-11)
+
+    def test_bandpass(self):
+        z, p, k = cheby2(9, 40, [0.07, 0.2], 'pass', output='zpk')
+        z2 = [-9.999999999999999e-01,
+               3.676588029658514e-01 + 9.299607543341383e-01j,
+               3.676588029658514e-01 - 9.299607543341383e-01j,
+               7.009689684982283e-01 + 7.131917730894889e-01j,
+               7.009689684982283e-01 - 7.131917730894889e-01j,
+               7.815697973765858e-01 + 6.238178033919218e-01j,
+               7.815697973765858e-01 - 6.238178033919218e-01j,
+               8.063793628819866e-01 + 5.913986160941200e-01j,
+               8.063793628819866e-01 - 5.913986160941200e-01j,
+               1.000000000000001e+00,
+               9.944493019920448e-01 + 1.052168511576739e-01j,
+               9.944493019920448e-01 - 1.052168511576739e-01j,
+               9.854674703367308e-01 + 1.698642543566085e-01j,
+               9.854674703367308e-01 - 1.698642543566085e-01j,
+               9.762751735919308e-01 + 2.165335665157851e-01j,
+               9.762751735919308e-01 - 2.165335665157851e-01j,
+               9.792277171575134e-01 + 2.027636011479496e-01j,
+               9.792277171575134e-01 - 2.027636011479496e-01j]
+        p2 = [8.143803410489621e-01 + 5.411056063397541e-01j,
+              8.143803410489621e-01 - 5.411056063397541e-01j,
+              7.650769827887418e-01 + 5.195412242095543e-01j,
+              7.650769827887418e-01 - 5.195412242095543e-01j,
+              6.096241204063443e-01 + 3.568440484659796e-01j,
+              6.096241204063443e-01 - 3.568440484659796e-01j,
+              6.918192770246239e-01 + 4.770463577106911e-01j,
+              6.918192770246239e-01 - 4.770463577106911e-01j,
+              6.986241085779207e-01 + 1.146512226180060e-01j,
+              6.986241085779207e-01 - 1.146512226180060e-01j,
+              8.654645923909734e-01 + 1.604208797063147e-01j,
+              8.654645923909734e-01 - 1.604208797063147e-01j,
+              9.164831670444591e-01 + 1.969181049384918e-01j,
+              9.164831670444591e-01 - 1.969181049384918e-01j,
+              9.630425777594550e-01 + 2.317513360702271e-01j,
+              9.630425777594550e-01 - 2.317513360702271e-01j,
+              9.438104703725529e-01 + 2.193509900269860e-01j,
+              9.438104703725529e-01 - 2.193509900269860e-01j]
+        k2 = 9.345352824659604e-03
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-13)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-13)
+        xp_assert_close(k, k2, rtol=1e-11)
+
+    def test_bandstop(self):
+        z, p, k = cheby2(6, 55, [0.1, 0.9], 'stop', output='zpk')
+        z2 = [6.230544895101009e-01 + 7.821784343111114e-01j,
+              6.230544895101009e-01 - 7.821784343111114e-01j,
+              9.086608545660115e-01 + 4.175349702471991e-01j,
+              9.086608545660115e-01 - 4.175349702471991e-01j,
+              9.478129721465802e-01 + 3.188268649763867e-01j,
+              9.478129721465802e-01 - 3.188268649763867e-01j,
+              -6.230544895100982e-01 + 7.821784343111109e-01j,
+              -6.230544895100982e-01 - 7.821784343111109e-01j,
+              -9.086608545660116e-01 + 4.175349702472088e-01j,
+              -9.086608545660116e-01 - 4.175349702472088e-01j,
+              -9.478129721465784e-01 + 3.188268649763897e-01j,
+              -9.478129721465784e-01 - 3.188268649763897e-01j]
+        p2 = [-9.464094036167638e-01 + 1.720048695084344e-01j,
+              -9.464094036167638e-01 - 1.720048695084344e-01j,
+              -8.715844103386737e-01 + 1.370665039509297e-01j,
+              -8.715844103386737e-01 - 1.370665039509297e-01j,
+              -8.078751204586425e-01 + 5.729329866682983e-02j,
+              -8.078751204586425e-01 - 5.729329866682983e-02j,
+               9.464094036167665e-01 + 1.720048695084332e-01j,
+               9.464094036167665e-01 - 1.720048695084332e-01j,
+               8.078751204586447e-01 + 5.729329866683007e-02j,
+               8.078751204586447e-01 - 5.729329866683007e-02j,
+               8.715844103386721e-01 + 1.370665039509331e-01j,
+               8.715844103386721e-01 - 1.370665039509331e-01j]
+        k2 = 2.917823332763358e-03
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-13)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-13)
+        xp_assert_close(k, k2, rtol=1e-11)
+
+    def test_ba_output(self):
+        # with transfer function conversion, without digital conversion
+        b, a = cheby2(5, 20, [2010, 2100], 'stop', True)
+        b2 = [1.000000000000000e+00, 0,  # Matlab: 6.683253076978249e-12,
+              2.111512500000000e+07, 0,  # Matlab: 1.134325604589552e-04,
+              1.782966433781250e+14, 0,  # Matlab: 7.216787944356781e+02,
+              7.525901316990656e+20, 0,  # Matlab: 2.039829265789886e+09,
+              1.587960565565748e+27, 0,  # Matlab: 2.161236218626134e+15,
+              1.339913493808585e+33]
+        a2 = [1.000000000000000e+00, 1.849550755473371e+02,
+              2.113222918998538e+07, 3.125114149732283e+09,
+              1.785133457155609e+14, 1.979158697776348e+16,
+              7.535048322653831e+20, 5.567966191263037e+22,
+              1.589246884221346e+27, 5.871210648525566e+28,
+              1.339913493808590e+33]
+        xp_assert_close(b, b2, rtol=1e-14)
+        xp_assert_close(a, a2, rtol=1e-14)
+
+    def test_fs_param(self):
+        for fs in (900, 900.1, 1234.567):
+            for N in (0, 1, 2, 3, 10):
+                for fc in (100, 100.1, 432.12345):
+                    for btype in ('lp', 'hp'):
+                        ba1 = cheby2(N, 20, fc, btype, fs=fs)
+                        ba2 = cheby2(N, 20, fc/(fs/2), btype)
+                        for ba1_, ba2_ in zip(ba1, ba2):
+                            xp_assert_close(ba1_, ba2_)
+                for fc in ((100, 200), (100.1, 200.2), (321.123, 432.123)):
+                    for btype in ('bp', 'bs'):
+                        ba1 = cheby2(N, 20, fc, btype, fs=fs)
+                        for seq in (list, tuple, array):
+                            fcnorm = seq([f/(fs/2) for f in fc])
+                            ba2 = cheby2(N, 20, fcnorm, btype)
+                            for ba1_, ba2_ in zip(ba1, ba2):
+                                xp_assert_close(ba1_, ba2_)
+
+class TestEllip:
+
+    def test_degenerate(self):
+        # 0-order filter is just a passthrough
+        # Even-order filters have DC gain of -rp dB
+        # Stopband ripple factor doesn't matter
+        b, a = ellip(0, 10*np.log10(2), 123.456, 1, analog=True)
+        assert_array_almost_equal(b, [1/np.sqrt(2)])
+        xp_assert_equal(a, np.asarray([1.0]))
+
+        # 1-order filter is same for all types
+        b, a = ellip(1, 10*np.log10(2), 1, 1, analog=True)
+        assert_array_almost_equal(b, [1])
+        assert_array_almost_equal(a, [1, 1])
+
+        z, p, k = ellip(1, 1, 55, 0.3, output='zpk')
+        xp_assert_close(z, [-9.999999999999998e-01], rtol=1e-14)
+        xp_assert_close(p, [-6.660721153525525e-04], rtol=1e-10)
+        xp_assert_close(k, 5.003330360576763e-01, rtol=1e-14)
+
+    def test_basic(self):
+        for N in range(25):
+            wn = 0.01
+            z, p, k = ellip(N, 1, 40, wn, 'low', analog=True, output='zpk')
+            assert len(p) == N
+            assert all(np.real(p) <= 0)  # No poles in right half of S-plane
+
+        for N in range(25):
+            wn = 0.01
+            z, p, k = ellip(N, 1, 40, wn, 'high', analog=False, output='zpk')
+            assert all(np.abs(p) <= 1)  # No poles outside unit circle
+
+        b3, a3 = ellip(5, 3, 26, 1, analog=True)
+        assert_array_almost_equal(b3, [0.1420, 0, 0.3764, 0,
+                                       0.2409], decimal=4)
+        assert_array_almost_equal(a3, [1, 0.5686, 1.8061, 0.8017, 0.8012,
+                                       0.2409], decimal=4)
+
+        b, a = ellip(3, 1, 60, [0.4, 0.7], 'stop')
+        assert_array_almost_equal(b, [0.3310, 0.3469, 1.1042, 0.7044, 1.1042,
+                                      0.3469, 0.3310], decimal=4)
+        assert_array_almost_equal(a, [1.0000, 0.6973, 1.1441, 0.5878, 0.7323,
+                                      0.1131, -0.0060], decimal=4)
+
+    def test_highpass(self):
+        # high even order
+        z, p, k = ellip(24, 1, 80, 0.3, 'high', output='zpk')
+        z2 = [9.761875332501075e-01 + 2.169283290099910e-01j,
+              9.761875332501075e-01 - 2.169283290099910e-01j,
+              8.413503353963494e-01 + 5.404901600661900e-01j,
+              8.413503353963494e-01 - 5.404901600661900e-01j,
+              7.160082576305009e-01 + 6.980918098681732e-01j,
+              7.160082576305009e-01 - 6.980918098681732e-01j,
+              6.456533638965329e-01 + 7.636306264739803e-01j,
+              6.456533638965329e-01 - 7.636306264739803e-01j,
+              6.127321820971366e-01 + 7.902906256703928e-01j,
+              6.127321820971366e-01 - 7.902906256703928e-01j,
+              5.983607817490196e-01 + 8.012267936512676e-01j,
+              5.983607817490196e-01 - 8.012267936512676e-01j,
+              5.922577552594799e-01 + 8.057485658286990e-01j,
+              5.922577552594799e-01 - 8.057485658286990e-01j,
+              5.896952092563588e-01 + 8.076258788449631e-01j,
+              5.896952092563588e-01 - 8.076258788449631e-01j,
+              5.886248765538837e-01 + 8.084063054565607e-01j,
+              5.886248765538837e-01 - 8.084063054565607e-01j,
+              5.881802711123132e-01 + 8.087298490066037e-01j,
+              5.881802711123132e-01 - 8.087298490066037e-01j,
+              5.879995719101164e-01 + 8.088612386766461e-01j,
+              5.879995719101164e-01 - 8.088612386766461e-01j,
+              5.879354086709576e-01 + 8.089078780868164e-01j,
+              5.879354086709576e-01 - 8.089078780868164e-01j]
+        p2 = [-3.184805259081650e-01 + 4.206951906775851e-01j,
+              -3.184805259081650e-01 - 4.206951906775851e-01j,
+               1.417279173459985e-01 + 7.903955262836452e-01j,
+               1.417279173459985e-01 - 7.903955262836452e-01j,
+               4.042881216964651e-01 + 8.309042239116594e-01j,
+               4.042881216964651e-01 - 8.309042239116594e-01j,
+               5.128964442789670e-01 + 8.229563236799665e-01j,
+               5.128964442789670e-01 - 8.229563236799665e-01j,
+               5.569614712822724e-01 + 8.155957702908510e-01j,
+               5.569614712822724e-01 - 8.155957702908510e-01j,
+               5.750478870161392e-01 + 8.118633973883931e-01j,
+               5.750478870161392e-01 - 8.118633973883931e-01j,
+               5.825314018170804e-01 + 8.101960910679270e-01j,
+               5.825314018170804e-01 - 8.101960910679270e-01j,
+               5.856397379751872e-01 + 8.094825218722543e-01j,
+               5.856397379751872e-01 - 8.094825218722543e-01j,
+               5.869326035251949e-01 + 8.091827531557583e-01j,
+               5.869326035251949e-01 - 8.091827531557583e-01j,
+               5.874697218855733e-01 + 8.090593298213502e-01j,
+               5.874697218855733e-01 - 8.090593298213502e-01j,
+               5.876904783532237e-01 + 8.090127161018823e-01j,
+               5.876904783532237e-01 - 8.090127161018823e-01j,
+               5.877753105317594e-01 + 8.090050577978136e-01j,
+               5.877753105317594e-01 - 8.090050577978136e-01j]
+        k2 = 4.918081266957108e-02
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-4)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-4)
+        xp_assert_close(k, k2, rtol=1e-3)
+
+        # high odd order
+        z, p, k = ellip(23, 1, 70, 0.5, 'high', output='zpk')
+        z2 = [9.999999999998661e-01,
+              6.603717261750994e-01 + 7.509388678638675e-01j,
+              6.603717261750994e-01 - 7.509388678638675e-01j,
+              2.788635267510325e-01 + 9.603307416968041e-01j,
+              2.788635267510325e-01 - 9.603307416968041e-01j,
+              1.070215532544218e-01 + 9.942567008268131e-01j,
+              1.070215532544218e-01 - 9.942567008268131e-01j,
+              4.049427369978163e-02 + 9.991797705105507e-01j,
+              4.049427369978163e-02 - 9.991797705105507e-01j,
+              1.531059368627931e-02 + 9.998827859909265e-01j,
+              1.531059368627931e-02 - 9.998827859909265e-01j,
+              5.808061438534933e-03 + 9.999831330689181e-01j,
+              5.808061438534933e-03 - 9.999831330689181e-01j,
+              2.224277847754599e-03 + 9.999975262909676e-01j,
+              2.224277847754599e-03 - 9.999975262909676e-01j,
+              8.731857107534554e-04 + 9.999996187732845e-01j,
+              8.731857107534554e-04 - 9.999996187732845e-01j,
+              3.649057346914968e-04 + 9.999999334218996e-01j,
+              3.649057346914968e-04 - 9.999999334218996e-01j,
+              1.765538109802615e-04 + 9.999999844143768e-01j,
+              1.765538109802615e-04 - 9.999999844143768e-01j,
+              1.143655290967426e-04 + 9.999999934602630e-01j,
+              1.143655290967426e-04 - 9.999999934602630e-01j]
+        p2 = [-6.322017026545028e-01,
+              -4.648423756662754e-01 + 5.852407464440732e-01j,
+              -4.648423756662754e-01 - 5.852407464440732e-01j,
+              -2.249233374627773e-01 + 8.577853017985717e-01j,
+              -2.249233374627773e-01 - 8.577853017985717e-01j,
+              -9.234137570557621e-02 + 9.506548198678851e-01j,
+              -9.234137570557621e-02 - 9.506548198678851e-01j,
+              -3.585663561241373e-02 + 9.821494736043981e-01j,
+              -3.585663561241373e-02 - 9.821494736043981e-01j,
+              -1.363917242312723e-02 + 9.933844128330656e-01j,
+              -1.363917242312723e-02 - 9.933844128330656e-01j,
+              -5.131505238923029e-03 + 9.975221173308673e-01j,
+              -5.131505238923029e-03 - 9.975221173308673e-01j,
+              -1.904937999259502e-03 + 9.990680819857982e-01j,
+              -1.904937999259502e-03 - 9.990680819857982e-01j,
+              -6.859439885466834e-04 + 9.996492201426826e-01j,
+              -6.859439885466834e-04 - 9.996492201426826e-01j,
+              -2.269936267937089e-04 + 9.998686250679161e-01j,
+              -2.269936267937089e-04 - 9.998686250679161e-01j,
+              -5.687071588789117e-05 + 9.999527573294513e-01j,
+              -5.687071588789117e-05 - 9.999527573294513e-01j,
+              -6.948417068525226e-07 + 9.999882737700173e-01j,
+              -6.948417068525226e-07 - 9.999882737700173e-01j]
+        k2 = 1.220910020289434e-02
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-4)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-4)
+        xp_assert_close(k, k2, rtol=1e-3)
+
+    def test_bandpass(self):
+        z, p, k = ellip(7, 1, 40, [0.07, 0.2], 'pass', output='zpk')
+        z2 = [-9.999999999999991e-01,
+               6.856610961780020e-01 + 7.279209168501619e-01j,
+               6.856610961780020e-01 - 7.279209168501619e-01j,
+               7.850346167691289e-01 + 6.194518952058737e-01j,
+               7.850346167691289e-01 - 6.194518952058737e-01j,
+               7.999038743173071e-01 + 6.001281461922627e-01j,
+               7.999038743173071e-01 - 6.001281461922627e-01j,
+               9.999999999999999e-01,
+               9.862938983554124e-01 + 1.649980183725925e-01j,
+               9.862938983554124e-01 - 1.649980183725925e-01j,
+               9.788558330548762e-01 + 2.045513580850601e-01j,
+               9.788558330548762e-01 - 2.045513580850601e-01j,
+               9.771155231720003e-01 + 2.127093189691258e-01j,
+               9.771155231720003e-01 - 2.127093189691258e-01j]
+        p2 = [8.063992755498643e-01 + 5.858071374778874e-01j,
+              8.063992755498643e-01 - 5.858071374778874e-01j,
+              8.050395347071724e-01 + 5.639097428109795e-01j,
+              8.050395347071724e-01 - 5.639097428109795e-01j,
+              8.113124936559144e-01 + 4.855241143973142e-01j,
+              8.113124936559144e-01 - 4.855241143973142e-01j,
+              8.665595314082394e-01 + 3.334049560919331e-01j,
+              8.665595314082394e-01 - 3.334049560919331e-01j,
+              9.412369011968871e-01 + 2.457616651325908e-01j,
+              9.412369011968871e-01 - 2.457616651325908e-01j,
+              9.679465190411238e-01 + 2.228772501848216e-01j,
+              9.679465190411238e-01 - 2.228772501848216e-01j,
+              9.747235066273385e-01 + 2.178937926146544e-01j,
+              9.747235066273385e-01 - 2.178937926146544e-01j]
+        k2 = 8.354782670263239e-03
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-4)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-4)
+        xp_assert_close(k, k2, rtol=1e-3)
+
+        z, p, k = ellip(5, 1, 75, [90.5, 110.5], 'pass', True, 'zpk')
+        z2 = [-5.583607317695175e-14 + 1.433755965989225e+02j,
+              -5.583607317695175e-14 - 1.433755965989225e+02j,
+               5.740106416459296e-14 + 1.261678754570291e+02j,
+               5.740106416459296e-14 - 1.261678754570291e+02j,
+              -2.199676239638652e-14 + 6.974861996895196e+01j,
+              -2.199676239638652e-14 - 6.974861996895196e+01j,
+              -3.372595657044283e-14 + 7.926145989044531e+01j,
+              -3.372595657044283e-14 - 7.926145989044531e+01j,
+              0]
+        p2 = [-8.814960004852743e-01 + 1.104124501436066e+02j,
+              -8.814960004852743e-01 - 1.104124501436066e+02j,
+              -2.477372459140184e+00 + 1.065638954516534e+02j,
+              -2.477372459140184e+00 - 1.065638954516534e+02j,
+              -3.072156842945799e+00 + 9.995404870405324e+01j,
+              -3.072156842945799e+00 - 9.995404870405324e+01j,
+              -2.180456023925693e+00 + 9.379206865455268e+01j,
+              -2.180456023925693e+00 - 9.379206865455268e+01j,
+              -7.230484977485752e-01 + 9.056598800801140e+01j,
+              -7.230484977485752e-01 - 9.056598800801140e+01j]
+        k2 = 3.774571622827070e-02
+        xp_assert_close(sorted(z, key=np.imag),
+                        sorted(z2, key=np.imag), rtol=1e-4)
+        xp_assert_close(sorted(p, key=np.imag),
+                        sorted(p2, key=np.imag), rtol=1e-6)
+        xp_assert_close(k, k2, rtol=1e-3)
+
+    def test_bandstop(self):
+        z, p, k = ellip(8, 1, 65, [0.2, 0.4], 'stop', output='zpk')
+        z2 = [3.528578094286510e-01 + 9.356769561794296e-01j,
+              3.528578094286510e-01 - 9.356769561794296e-01j,
+              3.769716042264783e-01 + 9.262248159096587e-01j,
+              3.769716042264783e-01 - 9.262248159096587e-01j,
+              4.406101783111199e-01 + 8.976985411420985e-01j,
+              4.406101783111199e-01 - 8.976985411420985e-01j,
+              5.539386470258847e-01 + 8.325574907062760e-01j,
+              5.539386470258847e-01 - 8.325574907062760e-01j,
+              6.748464963023645e-01 + 7.379581332490555e-01j,
+              6.748464963023645e-01 - 7.379581332490555e-01j,
+              7.489887970285254e-01 + 6.625826604475596e-01j,
+              7.489887970285254e-01 - 6.625826604475596e-01j,
+              7.913118471618432e-01 + 6.114127579150699e-01j,
+              7.913118471618432e-01 - 6.114127579150699e-01j,
+              7.806804740916381e-01 + 6.249303940216475e-01j,
+              7.806804740916381e-01 - 6.249303940216475e-01j]
+
+        p2 = [-1.025299146693730e-01 + 5.662682444754943e-01j,
+              -1.025299146693730e-01 - 5.662682444754943e-01j,
+               1.698463595163031e-01 + 8.926678667070186e-01j,
+               1.698463595163031e-01 - 8.926678667070186e-01j,
+               2.750532687820631e-01 + 9.351020170094005e-01j,
+               2.750532687820631e-01 - 9.351020170094005e-01j,
+               3.070095178909486e-01 + 9.457373499553291e-01j,
+               3.070095178909486e-01 - 9.457373499553291e-01j,
+               7.695332312152288e-01 + 2.792567212705257e-01j,
+               7.695332312152288e-01 - 2.792567212705257e-01j,
+               8.083818999225620e-01 + 4.990723496863960e-01j,
+               8.083818999225620e-01 - 4.990723496863960e-01j,
+               8.066158014414928e-01 + 5.649811440393374e-01j,
+               8.066158014414928e-01 - 5.649811440393374e-01j,
+               8.062787978834571e-01 + 5.855780880424964e-01j,
+               8.062787978834571e-01 - 5.855780880424964e-01j]
+        k2 = 2.068622545291259e-01
+        xp_assert_close(sorted(z, key=np.angle),
+                        sorted(z2, key=np.angle), rtol=1e-6)
+        xp_assert_close(sorted(p, key=np.angle),
+                        sorted(p2, key=np.angle), rtol=1e-5)
+        xp_assert_close(k, k2, rtol=1e-5)
+
+    def test_ba_output(self):
+        # with transfer function conversion,  without digital conversion
+        b, a = ellip(5, 1, 40, [201, 240], 'stop', True)
+        b2 = [
+             1.000000000000000e+00, 0,  # Matlab: 1.743506051190569e-13,
+             2.426561778314366e+05, 0,  # Matlab: 3.459426536825722e-08,
+             2.348218683400168e+10, 0,  # Matlab: 2.559179747299313e-03,
+             1.132780692872241e+15, 0,  # Matlab: 8.363229375535731e+01,
+             2.724038554089566e+19, 0,  # Matlab: 1.018700994113120e+06,
+             2.612380874940186e+23
+             ]
+        a2 = [
+             1.000000000000000e+00, 1.337266601804649e+02,
+             2.486725353510667e+05, 2.628059713728125e+07,
+             2.436169536928770e+10, 1.913554568577315e+12,
+             1.175208184614438e+15, 6.115751452473410e+16,
+             2.791577695211466e+19, 7.241811142725384e+20,
+             2.612380874940182e+23
+             ]
+        xp_assert_close(b, b2, rtol=1e-6)
+        xp_assert_close(a, a2, rtol=1e-4)
+
+    def test_fs_param(self):
+        for fs in (900, 900.1, 1234.567):
+            for N in (0, 1, 2, 3, 10):
+                for fc in (100, 100.1, 432.12345):
+                    for btype in ('lp', 'hp'):
+                        ba1 = ellip(N, 1, 20, fc, btype, fs=fs)
+                        ba2 = ellip(N, 1, 20, fc/(fs/2), btype)
+                        for ba1_, ba2_ in zip(ba1, ba2):
+                            xp_assert_close(ba1_, ba2_)
+                for fc in ((100, 200), (100.1, 200.2), (321.123, 432.123)):
+                    for btype in ('bp', 'bs'):
+                        ba1 = ellip(N, 1, 20, fc, btype, fs=fs)
+                        for seq in (list, tuple, array):
+                            fcnorm = seq([f/(fs/2) for f in fc])
+                            ba2 = ellip(N, 1, 20, fcnorm, btype)
+                            for ba1_, ba2_ in zip(ba1, ba2):
+                                xp_assert_close(ba1_, ba2_)
+
+    def test_fs_validation(self):
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            iirnotch(0.06, 30, fs=np.array([10, 20]))
+
+        with pytest.raises(ValueError, match="Sampling.*be none"):
+            iirnotch(0.06, 30, fs=None)
+
+
+def test_sos_consistency():
+    # Consistency checks of output='sos' for the specialized IIR filter
+    # design functions.
+    design_funcs = [(bessel, (0.1,)),
+                    (butter, (0.1,)),
+                    (cheby1, (45.0, 0.1)),
+                    (cheby2, (0.087, 0.1)),
+                    (ellip, (0.087, 45, 0.1))]
+    for func, args in design_funcs:
+        name = func.__name__
+
+        b, a = func(2, *args, output='ba')
+        sos = func(2, *args, output='sos')
+        xp_assert_close(sos, [np.hstack((b, a))], err_msg=f"{name}(2,...)")
+
+        zpk = func(3, *args, output='zpk')
+        sos = func(3, *args, output='sos')
+        xp_assert_close(sos, zpk2sos(*zpk), err_msg=f"{name}(3,...)")
+
+        zpk = func(4, *args, output='zpk')
+        sos = func(4, *args, output='sos')
+        xp_assert_close(sos, zpk2sos(*zpk), err_msg=f"{name}(4,...)")
+
+
+class TestIIRNotch:
+
+    def test_ba_output(self):
+        # Compare coefficients with Matlab ones
+        # for the equivalent input:
+        b, a = iirnotch(0.06, 30)
+        b2 = [
+             9.9686824e-01, -1.9584219e+00,
+             9.9686824e-01
+             ]
+        a2 = [
+             1.0000000e+00, -1.9584219e+00,
+             9.9373647e-01
+             ]
+
+        xp_assert_close(b, b2, rtol=1e-8)
+        xp_assert_close(a, a2, rtol=1e-8)
+
+    def test_frequency_response(self):
+        # Get filter coefficients
+        b, a = iirnotch(0.3, 30)
+
+        # Get frequency response
+        w, h = freqz(b, a, 1000)
+
+        # Pick 5 point
+        p = [200,  # w0 = 0.200
+             295,  # w0 = 0.295
+             300,  # w0 = 0.300
+             305,  # w0 = 0.305
+             400]  # w0 = 0.400
+
+        # Get frequency response correspondent to each of those points
+        hp = h[p]
+
+        # Check if the frequency response fulfill the specifications:
+        # hp[0] and hp[4]  correspond to frequencies distant from
+        # w0 = 0.3 and should be close to 1
+        xp_assert_close(abs(hp[0]), np.asarray(1.), rtol=1e-2, check_0d=False)
+        xp_assert_close(abs(hp[4]), np.asarray(1.), rtol=1e-2, check_0d=False)
+
+        # hp[1] and hp[3] correspond to frequencies approximately
+        # on the edges of the passband and should be close to -3dB
+        xp_assert_close(abs(hp[1]), 1/np.sqrt(2), rtol=1e-2)
+        xp_assert_close(abs(hp[3]), 1/np.sqrt(2), rtol=1e-2)
+
+        # hp[2] correspond to the frequency that should be removed
+        # the frequency response should be very close to 0
+        xp_assert_close(abs(hp[2]), np.asarray(0.0), atol=1e-10, check_0d=False)
+
+    def test_errors(self):
+        # Exception should be raised if w0 > 1 or w0 <0
+        assert_raises(ValueError, iirnotch, w0=2, Q=30)
+        assert_raises(ValueError, iirnotch, w0=-1, Q=30)
+
+        # Exception should be raised if any of the parameters
+        # are not float (or cannot be converted to one)
+        assert_raises(ValueError, iirnotch, w0="blabla", Q=30)
+        assert_raises(TypeError, iirnotch, w0=-1, Q=[1, 2, 3])
+
+    def test_fs_param(self):
+        # Get filter coefficients
+        b, a = iirnotch(1500, 30, fs=10000)
+
+        # Get frequency response
+        w, h = freqz(b, a, 1000, fs=10000)
+
+        # Pick 5 point
+        p = [200,  # w0 = 1000
+             295,  # w0 = 1475
+             300,  # w0 = 1500
+             305,  # w0 = 1525
+             400]  # w0 = 2000
+
+        # Get frequency response correspondent to each of those points
+        hp = h[p]
+
+        # Check if the frequency response fulfill the specifications:
+        # hp[0] and hp[4]  correspond to frequencies distant from
+        # w0 = 1500 and should be close to 1
+        xp_assert_close(abs(hp[0]), np.ones_like(abs(hp[0])), rtol=1e-2,
+                        check_0d=False)
+        xp_assert_close(abs(hp[4]), np.ones_like(abs(hp[4])), rtol=1e-2,
+                        check_0d=False)
+
+        # hp[1] and hp[3] correspond to frequencies approximately
+        # on the edges of the passband and should be close to -3dB
+        xp_assert_close(abs(hp[1]), 1/np.sqrt(2), rtol=1e-2)
+        xp_assert_close(abs(hp[3]), 1/np.sqrt(2), rtol=1e-2)
+
+        # hp[2] correspond to the frequency that should be removed
+        # the frequency response should be very close to 0
+        xp_assert_close(abs(hp[2]), np.asarray(0.0), atol=1e-10, check_0d=False)
+
+
+class TestIIRPeak:
+
+    def test_ba_output(self):
+        # Compare coefficients with Matlab ones
+        # for the equivalent input:
+        b, a = iirpeak(0.06, 30)
+        b2 = [
+             3.131764229e-03, 0,
+             -3.131764229e-03
+             ]
+        a2 = [
+             1.0000000e+00, -1.958421917e+00,
+             9.9373647e-01
+             ]
+        xp_assert_close(b, b2, rtol=1e-8)
+        xp_assert_close(a, a2, rtol=1e-8)
+
+    def test_frequency_response(self):
+        # Get filter coefficients
+        b, a = iirpeak(0.3, 30)
+
+        # Get frequency response
+        w, h = freqz(b, a, 1000)
+
+        # Pick 5 point
+        p = [30,  # w0 = 0.030
+             295,  # w0 = 0.295
+             300,  # w0 = 0.300
+             305,  # w0 = 0.305
+             800]  # w0 = 0.800
+
+        # Get frequency response correspondent to each of those points
+        hp = h[p]
+
+        # Check if the frequency response fulfill the specifications:
+        # hp[0] and hp[4]  correspond to frequencies distant from
+        # w0 = 0.3 and should be close to 0
+        xp_assert_close(abs(hp[0]),
+                        np.zeros_like(abs(hp[0])), atol=1e-2, check_0d=False)
+        xp_assert_close(abs(hp[4]),
+                        np.zeros_like(abs(hp[4])), atol=1e-2, check_0d=False)
+
+        # hp[1] and hp[3] correspond to frequencies approximately
+        # on the edges of the passband and should be close to 10**(-3/20)
+        xp_assert_close(abs(hp[1]), 1/np.sqrt(2), rtol=1e-2)
+        xp_assert_close(abs(hp[3]), 1/np.sqrt(2), rtol=1e-2)
+
+        # hp[2] correspond to the frequency that should be retained and
+        # the frequency response should be very close to 1
+        xp_assert_close(abs(hp[2]), np.asarray(1.0), rtol=1e-10, check_0d=False)
+
+    def test_errors(self):
+        # Exception should be raised if w0 > 1 or w0 <0
+        assert_raises(ValueError, iirpeak, w0=2, Q=30)
+        assert_raises(ValueError, iirpeak, w0=-1, Q=30)
+
+        # Exception should be raised if any of the parameters
+        # are not float (or cannot be converted to one)
+        assert_raises(ValueError, iirpeak, w0="blabla", Q=30)
+        assert_raises(TypeError, iirpeak, w0=-1, Q=[1, 2, 3])
+
+    def test_fs_param(self):
+        # Get filter coefficients
+        b, a = iirpeak(1200, 30, fs=8000)
+
+        # Get frequency response
+        w, h = freqz(b, a, 1000, fs=8000)
+
+        # Pick 5 point
+        p = [30,  # w0 = 120
+             295,  # w0 = 1180
+             300,  # w0 = 1200
+             305,  # w0 = 1220
+             800]  # w0 = 3200
+
+        # Get frequency response correspondent to each of those points
+        hp = h[p]
+
+        # Check if the frequency response fulfill the specifications:
+        # hp[0] and hp[4]  correspond to frequencies distant from
+        # w0 = 1200 and should be close to 0
+        xp_assert_close(abs(hp[0]),
+                        np.zeros_like(abs(hp[0])), atol=1e-2, check_0d=False)
+        xp_assert_close(abs(hp[4]),
+                        np.zeros_like(abs(hp[4])), atol=1e-2, check_0d=False)
+
+        # hp[1] and hp[3] correspond to frequencies approximately
+        # on the edges of the passband and should be close to 10**(-3/20)
+        xp_assert_close(abs(hp[1]), 1/np.sqrt(2), rtol=1e-2)
+        xp_assert_close(abs(hp[3]), 1/np.sqrt(2), rtol=1e-2)
+
+        # hp[2] correspond to the frequency that should be retained and
+        # the frequency response should be very close to 1
+        xp_assert_close(abs(hp[2]),
+                        np.ones_like(abs(hp[2])), rtol=1e-10, check_0d=False)
+
+
+class TestIIRComb:
+    # Test erroneous input cases
+    def test_invalid_input(self):
+        # w0 is <= 0 or >= fs / 2
+        fs = 1000
+        for args in [(-fs, 30), (0, 35), (fs / 2, 40), (fs, 35)]:
+            with pytest.raises(ValueError, match='w0 must be between '):
+                iircomb(*args, fs=fs)
+
+        # fs is not divisible by w0
+        for args in [(120, 30), (157, 35)]:
+            with pytest.raises(ValueError, match='fs must be divisible '):
+                iircomb(*args, fs=fs)
+
+        # https://github.com/scipy/scipy/issues/14043#issuecomment-1107349140
+        # Previously, fs=44100, w0=49.999 was rejected, but fs=2,
+        # w0=49.999/int(44100/2) was accepted. Now it is rejected, too.
+        with pytest.raises(ValueError, match='fs must be divisible '):
+            iircomb(w0=49.999/int(44100/2), Q=30)
+
+        with pytest.raises(ValueError, match='fs must be divisible '):
+            iircomb(w0=49.999, Q=30, fs=44100)
+
+        # Filter type is not notch or peak
+        for args in [(0.2, 30, 'natch'), (0.5, 35, 'comb')]:
+            with pytest.raises(ValueError, match='ftype must be '):
+                iircomb(*args)
+
+    # Verify that the filter's frequency response contains a
+    # notch at the cutoff frequency
+    @pytest.mark.parametrize('ftype', ('notch', 'peak'))
+    def test_frequency_response(self, ftype):
+        # Create a notching or peaking comb filter at 1000 Hz
+        b, a = iircomb(1000, 30, ftype=ftype, fs=10000)
+
+        # Compute the frequency response
+        freqs, response = freqz(b, a, 1000, fs=10000)
+
+        # Find the notch using argrelextrema
+        comb_points = argrelextrema(abs(response), np.less)[0]
+
+        # Verify that the first notch sits at 1000 Hz
+        comb1 = comb_points[0]
+        xp_assert_close(freqs[comb1], np.asarray(1000.), check_0d=False)
+
+    # Verify pass_zero parameter
+    @pytest.mark.parametrize('ftype,pass_zero,peak,notch',
+                             [('peak', True, 123.45, 61.725),
+                              ('peak', False, 61.725, 123.45),
+                              ('peak', None, 61.725, 123.45),
+                              ('notch', None, 61.725, 123.45),
+                              ('notch', True, 123.45, 61.725),
+                              ('notch', False, 61.725, 123.45)])
+    def test_pass_zero(self, ftype, pass_zero, peak, notch):
+        # Create a notching or peaking comb filter
+        b, a = iircomb(123.45, 30, ftype=ftype, fs=1234.5, pass_zero=pass_zero)
+
+        # Compute the frequency response
+        freqs, response = freqz(b, a, [peak, notch], fs=1234.5)
+
+        # Verify that expected notches are notches and peaks are peaks
+        assert abs(response[0]) > 0.99
+        assert abs(response[1]) < 1e-10
+
+    # All built-in IIR filters are real, so should have perfectly
+    # symmetrical poles and zeros. Then ba representation (using
+    # numpy.poly) will be purely real instead of having negligible
+    # imaginary parts.
+    def test_iir_symmetry(self):
+        b, a = iircomb(400, 30, fs=24000)
+        z, p, k = tf2zpk(b, a)
+        xp_assert_equal(sorted(z), sorted(z.conj()))
+        xp_assert_equal(sorted(p), sorted(p.conj()))
+        xp_assert_equal(k, np.real(k))
+
+        assert issubclass(b.dtype.type, np.floating)
+        assert issubclass(a.dtype.type, np.floating)
+
+    # Verify filter coefficients with MATLAB's iircomb function
+    def test_ba_output(self):
+        b_notch, a_notch = iircomb(60, 35, ftype='notch', fs=600)
+        b_notch2 = [0.957020174408697, 0.0, 0.0, 0.0, 0.0, 0.0,
+                    0.0, 0.0, 0.0, 0.0, -0.957020174408697]
+        a_notch2 = [1.0, 0.0, 0.0, 0.0, 0.0, 0.0,
+                    0.0, 0.0, 0.0, 0.0, -0.914040348817395]
+        xp_assert_close(b_notch, b_notch2)
+        xp_assert_close(a_notch, a_notch2)
+
+        b_peak, a_peak = iircomb(60, 35, ftype='peak', fs=600)
+        b_peak2 = [0.0429798255913026, 0.0, 0.0, 0.0, 0.0, 0.0,
+                   0.0, 0.0, 0.0, 0.0, -0.0429798255913026]
+        a_peak2 = [1.0, 0.0, 0.0, 0.0, 0.0, 0.0,
+                   0.0, 0.0, 0.0, 0.0, 0.914040348817395]
+        xp_assert_close(b_peak, b_peak2)
+        xp_assert_close(a_peak, a_peak2)
+
+    # Verify that https://github.com/scipy/scipy/issues/14043 is fixed
+    def test_nearest_divisor(self):
+        # Create a notching comb filter
+        b, a = iircomb(50/int(44100/2), 50.0, ftype='notch')
+
+        # Compute the frequency response at an upper harmonic of 50
+        freqs, response = freqz(b, a, [22000], fs=44100)
+
+        # Before bug fix, this would produce N = 881, so that 22 kHz was ~0 dB.
+        # Now N = 882 correctly and 22 kHz should be a notch <-220 dB
+        assert abs(response[0]) < 1e-10
+
+    def test_fs_validation(self):
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            iircomb(1000, 30, fs=np.array([10, 20]))
+
+        with pytest.raises(ValueError, match="Sampling.*be none"):
+            iircomb(1000, 30, fs=None)
+
+
+class TestIIRDesign:
+
+    def test_exceptions(self):
+        with pytest.raises(ValueError, match="the same shape"):
+            iirdesign(0.2, [0.1, 0.3], 1, 40)
+        with pytest.raises(ValueError, match="the same shape"):
+            iirdesign(np.array([[0.3, 0.6], [0.3, 0.6]]),
+                      np.array([[0.4, 0.5], [0.4, 0.5]]), 1, 40)
+
+        # discrete filter with non-positive frequency
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign(0, 0.5, 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign(-0.1, 0.5, 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign(0.1, 0, 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign(0.1, -0.5, 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0, 0.3], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([-0.1, 0.3], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, 0], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, -0.3], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, 0.3], [0, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, 0.3], [-0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, 0.3], [0.1, 0], 1, 40)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, 0.3], [0.1, -0.5], 1, 40)
+
+        # analog filter with negative frequency
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign(-0.1, 0.5, 1, 40, analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign(0.1, -0.5, 1, 40, analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([-0.1, 0.3], [0.1, 0.5], 1, 40, analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, -0.3], [0.1, 0.5], 1, 40, analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, 0.3], [-0.1, 0.5], 1, 40, analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirdesign([0.1, 0.3], [0.1, -0.5], 1, 40, analog=True)
+
+        # discrete filter with fs=None, freq > 1
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign(1, 0.5, 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign(1.1, 0.5, 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign(0.1, 1, 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign(0.1, 1.5, 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([1, 0.3], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([1.1, 0.3], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([0.1, 1], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([0.1, 1.1], [0.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([0.1, 0.3], [1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([0.1, 0.3], [1.1, 0.5], 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([0.1, 0.3], [0.1, 1], 1, 40)
+        with pytest.raises(ValueError, match="must be less than 1"):
+            iirdesign([0.1, 0.3], [0.1, 1.5], 1, 40)
+
+        # discrete filter with fs>2, wp, ws < fs/2 must pass
+        iirdesign(100, 500, 1, 40, fs=2000)
+        iirdesign(500, 100, 1, 40, fs=2000)
+        iirdesign([200, 400], [100, 500], 1, 40, fs=2000)
+        iirdesign([100, 500], [200, 400], 1, 40, fs=2000)
+
+        # discrete filter with fs>2, freq > fs/2: this must raise
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign(1000, 400, 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign(1100, 500, 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign(100, 1000, 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign(100, 1100, 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([1000, 400], [100, 500], 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([1100, 400], [100, 500], 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([200, 1000], [100, 500], 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([200, 1100], [100, 500], 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([200, 400], [1000, 500], 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([200, 400], [1100, 500], 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([200, 400], [100, 1000], 1, 40, fs=2000)
+        with pytest.raises(ValueError, match="must be less than fs/2"):
+            iirdesign([200, 400], [100, 1100], 1, 40, fs=2000)
+
+        with pytest.raises(ValueError, match="strictly inside stopband"):
+            iirdesign([0.1, 0.4], [0.5, 0.6], 1, 40)
+        with pytest.raises(ValueError, match="strictly inside stopband"):
+            iirdesign([0.5, 0.6], [0.1, 0.4], 1, 40)
+        with pytest.raises(ValueError, match="strictly inside stopband"):
+            iirdesign([0.3, 0.6], [0.4, 0.7], 1, 40)
+        with pytest.raises(ValueError, match="strictly inside stopband"):
+            iirdesign([0.4, 0.7], [0.3, 0.6], 1, 40)
+
+    def test_fs_validation(self):
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            iirfilter(1, 1, btype="low", fs=np.array([10, 20]))
+
+
+class TestIIRFilter:
+
+    def test_symmetry(self):
+        # All built-in IIR filters are real, so should have perfectly
+        # symmetrical poles and zeros. Then ba representation (using
+        # numpy.poly) will be purely real instead of having negligible
+        # imaginary parts.
+        for N in np.arange(1, 26):
+            for ftype in ('butter', 'bessel', 'cheby1', 'cheby2', 'ellip'):
+                z, p, k = iirfilter(N, 1.1, 1, 20, 'low', analog=True,
+                                    ftype=ftype, output='zpk')
+                xp_assert_equal(sorted(z),
+                                sorted(z.conj()))
+                xp_assert_equal(sorted(p),
+                                sorted(p.conj()))
+                xp_assert_equal(k, np.real(k))
+
+                b, a = iirfilter(N, 1.1, 1, 20, 'low', analog=True,
+                                 ftype=ftype, output='ba')
+                assert issubclass(b.dtype.type, np.floating)
+                assert issubclass(a.dtype.type, np.floating)
+
+    def test_int_inputs(self):
+        # Using integer frequency arguments and large N should not produce
+        # numpy integers that wraparound to negative numbers
+        k = iirfilter(24, 100, btype='low', analog=True, ftype='bessel',
+                      output='zpk')[2]
+        k2 = 9.999999999999989e+47
+        xp_assert_close(np.asarray(k),  np.asarray(k2))
+        # if fs is specified then the normalization of Wn to have
+        # 0 <= Wn <= 1 should not cause an integer overflow
+        # the following line should not raise an exception
+        iirfilter(20, [1000000000, 1100000000], btype='bp',
+                      analog=False, fs=6250000000)
+
+    def test_invalid_wn_size(self):
+        # low and high have 1 Wn, band and stop have 2 Wn
+        assert_raises(ValueError, iirfilter, 1, [0.1, 0.9], btype='low')
+        assert_raises(ValueError, iirfilter, 1, [0.2, 0.5], btype='high')
+        assert_raises(ValueError, iirfilter, 1, 0.2, btype='bp')
+        assert_raises(ValueError, iirfilter, 1, 400, btype='bs', analog=True)
+
+    def test_invalid_wn_range(self):
+        # For digital filters, 0 <= Wn <= 1
+        assert_raises(ValueError, iirfilter, 1, 2, btype='low')
+        assert_raises(ValueError, iirfilter, 1, [0.5, 1], btype='band')
+        assert_raises(ValueError, iirfilter, 1, [0., 0.5], btype='band')
+        assert_raises(ValueError, iirfilter, 1, -1, btype='high')
+        assert_raises(ValueError, iirfilter, 1, [1, 2], btype='band')
+        assert_raises(ValueError, iirfilter, 1, [10, 20], btype='stop')
+
+        # analog=True with non-positive critical frequencies
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirfilter(2, 0, btype='low', analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirfilter(2, -1, btype='low', analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirfilter(2, [0, 100], analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirfilter(2, [-1, 100], analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirfilter(2, [10, 0], analog=True)
+        with pytest.raises(ValueError, match="must be greater than 0"):
+            iirfilter(2, [10, -1], analog=True)
+
+    def test_analog_sos(self):
+        # first order Butterworth filter with Wn = 1 has tf 1/(s+1)
+        sos = [[0., 0., 1., 0., 1., 1.]]
+        sos2 = iirfilter(N=1, Wn=1, btype='low', analog=True, output='sos')
+        assert_array_almost_equal(sos, sos2)
+
+    def test_wn1_ge_wn0(self):
+        # gh-15773: should raise error if Wn[0] >= Wn[1]
+        with pytest.raises(ValueError,
+                           match=r"Wn\[0\] must be less than Wn\[1\]"):
+            iirfilter(2, [0.5, 0.5])
+        with pytest.raises(ValueError,
+                           match=r"Wn\[0\] must be less than Wn\[1\]"):
+            iirfilter(2, [0.6, 0.5])
+
+
+class TestGroupDelay:
+    def test_identity_filter(self):
+        w, gd = group_delay((1, 1))
+        assert_array_almost_equal(w, pi * np.arange(512) / 512)
+        assert_array_almost_equal(gd, np.zeros(512))
+        w, gd = group_delay((1, 1), whole=True)
+        assert_array_almost_equal(w, 2 * pi * np.arange(512) / 512)
+        assert_array_almost_equal(gd, np.zeros(512))
+
+    def test_fir(self):
+        # Let's design linear phase FIR and check that the group delay
+        # is constant.
+        N = 100
+        b = firwin(N + 1, 0.1)
+        w, gd = group_delay((b, 1))
+        xp_assert_close(gd, np.ones_like(gd)*(0.5 * N))
+
+    def test_iir(self):
+        # Let's design Butterworth filter and test the group delay at
+        # some points against MATLAB answer.
+        b, a = butter(4, 0.1)
+        w = np.linspace(0, pi, num=10, endpoint=False)
+        w, gd = group_delay((b, a), w=w)
+        matlab_gd = np.array([8.249313898506037, 11.958947880907104,
+                              2.452325615326005, 1.048918665702008,
+                              0.611382575635897, 0.418293269460578,
+                              0.317932917836572, 0.261371844762525,
+                              0.229038045801298, 0.212185774208521])
+        assert_array_almost_equal(gd, matlab_gd)
+
+    @pytest.mark.thread_unsafe
+    def test_singular(self):
+        # Let's create a filter with zeros and poles on the unit circle and
+        # check if warnings are raised at those frequencies.
+        z1 = np.exp(1j * 0.1 * pi)
+        z2 = np.exp(1j * 0.25 * pi)
+        p1 = np.exp(1j * 0.5 * pi)
+        p2 = np.exp(1j * 0.8 * pi)
+        b = np.convolve([1, -z1], [1, -z2])
+        a = np.convolve([1, -p1], [1, -p2])
+        w = np.array([0.1 * pi, 0.25 * pi, -0.5 * pi, -0.8 * pi])
+
+        w, gd = assert_warns(UserWarning, group_delay, (b, a), w=w)
+
+    def test_backward_compat(self):
+        # For backward compatibility, test if None act as a wrapper for default
+        w1, gd1 = group_delay((1, 1))
+        w2, gd2 = group_delay((1, 1), None)
+        assert_array_almost_equal(w1, w2)
+        assert_array_almost_equal(gd1, gd2)
+
+    def test_fs_param(self):
+        # Let's design Butterworth filter and test the group delay at
+        # some points against the normalized frequency answer.
+        b, a = butter(4, 4800, fs=96000)
+        w = np.linspace(0, 96000/2, num=10, endpoint=False)
+        w, gd = group_delay((b, a), w=w, fs=96000)
+        norm_gd = np.array([8.249313898506037, 11.958947880907104,
+                            2.452325615326005, 1.048918665702008,
+                            0.611382575635897, 0.418293269460578,
+                            0.317932917836572, 0.261371844762525,
+                            0.229038045801298, 0.212185774208521])
+        assert_array_almost_equal(gd, norm_gd)
+
+    def test_w_or_N_types(self):
+        # Measure at 8 equally-spaced points
+        for N in (8, np.int8(8), np.int16(8), np.int32(8), np.int64(8),
+                  np.array(8)):
+            w, gd = group_delay((1, 1), N)
+            assert_array_almost_equal(w, pi * np.arange(8) / 8)
+            assert_array_almost_equal(gd, np.zeros(8))
+
+        # Measure at frequency 8 rad/sec
+        for w in (8.0, 8.0+0j):
+            w_out, gd = group_delay((1, 1), w)
+            assert_array_almost_equal(w_out, [8])
+            assert_array_almost_equal(gd, [0])
+
+    def test_complex_coef(self):
+        # gh-19586: handle complex coef TFs
+        #
+        # for g(z) = (alpha*z+1)/(1+conjugate(alpha)), group delay is
+        # given by function below.
+        #
+        # def gd_expr(w, alpha):
+        #     num = 1j*(abs(alpha)**2-1)*np.exp(1j*w)
+        #     den = (alpha*np.exp(1j*w)+1)*(np.exp(1j*w)+np.conj(alpha))
+        #     return -np.imag(num/den)
+
+        # arbitrary non-real alpha
+        alpha = -0.6143077933232609+0.3355978770229421j
+        # 8 points from from -pi to pi
+        wref = np.array([-3.141592653589793 ,
+                         -2.356194490192345 ,
+                         -1.5707963267948966,
+                         -0.7853981633974483,
+                         0.                ,
+                         0.7853981633974483,
+                         1.5707963267948966,
+                         2.356194490192345 ])
+        gdref =  array([0.18759548150354619,
+                        0.17999770352712252,
+                        0.23598047471879877,
+                        0.46539443069907194,
+                        1.9511492420564165 ,
+                        3.478129975138865  ,
+                        0.6228594960517333 ,
+                        0.27067831839471224])
+        b = [alpha,1]
+        a = [1, np.conjugate(alpha)]
+        gdtest = group_delay((b,a), wref)[1]
+        # need nulp=14 for macOS arm64 wheel builds; added 2 for some
+        # robustness on other platforms.
+        assert_array_almost_equal_nulp(gdtest, gdref, nulp=16)
+
+    def test_fs_validation(self):
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            group_delay((1, 1), fs=np.array([10, 20]))
+
+        with pytest.raises(ValueError, match="Sampling.*be none"):
+            group_delay((1, 1), fs=None)
+
+
+class TestGammatone:
+    # Test erroneous input cases.
+    def test_invalid_input(self):
+        # Cutoff frequency is <= 0 or >= fs / 2.
+        fs = 16000
+        for args in [(-fs, 'iir'), (0, 'fir'), (fs / 2, 'iir'), (fs, 'fir')]:
+            with pytest.raises(ValueError, match='The frequency must be '
+                               'between '):
+                gammatone(*args, fs=fs)
+
+        # Filter type is not fir or iir
+        for args in [(440, 'fie'), (220, 'it')]:
+            with pytest.raises(ValueError, match='ftype must be '):
+                gammatone(*args, fs=fs)
+
+        # Order is <= 0 or > 24 for FIR filter.
+        for args in [(440, 'fir', -50), (220, 'fir', 0), (110, 'fir', 25),
+                     (55, 'fir', 50)]:
+            with pytest.raises(ValueError, match='Invalid order: '):
+                gammatone(*args, numtaps=None, fs=fs)
+
+    # Verify that the filter's frequency response is approximately
+    # 1 at the cutoff frequency.
+    def test_frequency_response(self):
+        fs = 16000
+        ftypes = ['fir', 'iir']
+        for ftype in ftypes:
+            # Create a gammatone filter centered at 1000 Hz.
+            b, a = gammatone(1000, ftype, fs=fs)
+
+            # Calculate the frequency response.
+            freqs, response = freqz(b, a)
+
+            # Determine peak magnitude of the response
+            # and corresponding frequency.
+            response_max = np.max(np.abs(response))
+            freq_hz = freqs[np.argmax(np.abs(response))] / ((2 * np.pi) / fs)
+
+            # Check that the peak magnitude is 1 and the frequency is 1000 Hz.
+            xp_assert_close(response_max,
+                            np.ones_like(response_max), rtol=1e-2, check_0d=False)
+            xp_assert_close(freq_hz,
+                            1000*np.ones_like(freq_hz), rtol=1e-2, check_0d=False)
+
+    # All built-in IIR filters are real, so should have perfectly
+    # symmetrical poles and zeros. Then ba representation (using
+    # numpy.poly) will be purely real instead of having negligible
+    # imaginary parts.
+    def test_iir_symmetry(self):
+        b, a = gammatone(440, 'iir', fs=24000)
+        z, p, k = tf2zpk(b, a)
+        xp_assert_equal(sorted(z), sorted(z.conj()))
+        xp_assert_equal(sorted(p), sorted(p.conj()))
+        xp_assert_equal(k, np.real(k))
+
+        assert issubclass(b.dtype.type, np.floating)
+        assert issubclass(a.dtype.type, np.floating)
+
+    # Verify FIR filter coefficients with the paper's
+    # Mathematica implementation
+    def test_fir_ba_output(self):
+        b, _ = gammatone(15, 'fir', fs=1000)
+        b2 = [0.0, 2.2608075649884e-04,
+              1.5077903981357e-03, 4.2033687753998e-03,
+              8.1508962726503e-03, 1.2890059089154e-02,
+              1.7833890391666e-02, 2.2392613558564e-02,
+              2.6055195863104e-02, 2.8435872863284e-02,
+              2.9293319149544e-02, 2.852976858014e-02,
+              2.6176557156294e-02, 2.2371510270395e-02,
+              1.7332485267759e-02]
+        xp_assert_close(b, b2)
+
+    # Verify IIR filter coefficients with the paper's MATLAB implementation
+    def test_iir_ba_output(self):
+        b, a = gammatone(440, 'iir', fs=16000)
+        b2 = [1.31494461367464e-06, -5.03391196645395e-06,
+              7.00649426000897e-06, -4.18951968419854e-06,
+              9.02614910412011e-07]
+        a2 = [1.0, -7.65646235454218,
+              25.7584699322366, -49.7319214483238,
+              60.2667361289181, -46.9399590980486,
+              22.9474798808461, -6.43799381299034,
+              0.793651554625368]
+        xp_assert_close(b, b2)
+        xp_assert_close(a, a2)
+
+    def test_fs_validation(self):
+        with pytest.raises(ValueError, match="Sampling.*single scalar"):
+            gammatone(440, 'iir', fs=np.array([10, 20]))