def read_QI(info,ist,file,line=None): """ QI[it][il][ib*ia*im,ie] """ QI = dict() for it in info.Nt[ist]: QI[it] = ND_list(info.Nl[it]) for il in range(info.Nl[it]): QI[it][il] = torch_complex.ComplexTensor( np.empty((info.Nb[ist],info.Na[ist][it],info.Nm(il),info.Ne[it]),dtype=np.float32), np.empty((info.Nb[ist],info.Na[ist][it],info.Nm(il),info.Ne[it]),dtype=np.float32) ) for ib in range(info.Nb[ist]): for it in info.Nt[ist]: for ia in range(info.Na[ist][it]): for il in range(info.Nl[it]): for im in range(info.Nm(il)): for ie in range(info.Ne[it]): if not line: line = file.readline().split() QI[it][il].real[ib,ia,im,ie] = float(line.pop(0)) if not line: line = file.readline().split() QI[it][il].imag[ib,ia,im,ie] = float(line.pop(0)) for it in info.Nt[ist]: for il in range(info.Nl[it]): QI[it][il] = torch_complex.ComplexTensor( torch.from_numpy(QI[it][il].real).view(-1,info.Ne[it]), torch.from_numpy(QI[it][il].imag).view(-1,info.Ne[it])) return QI,line
def read_SI(info,ist,file,line=None): """ SI[it1,it2][il1][il2][ie1,ia1,im1,ia2,im2,ie2] """ SI = dict() for it1,it2 in itertools.product( info.Nt[ist], info.Nt[ist] ): SI[it1,it2] = ND_list(info.Nl[it1],info.Nl[it2]) for il1,il2 in itertools.product( range(info.Nl[it1]), range(info.Nl[it2]) ): SI[it1,it2][il1][il2] = torch_complex.ComplexTensor( np.empty((info.Na[ist][it1],info.Nm(il1),info.Ne[it1],info.Na[ist][it2],info.Nm(il2),info.Ne[it2]),dtype=np.float32), np.empty((info.Na[ist][it1],info.Nm(il1),info.Ne[it1],info.Na[ist][it2],info.Nm(il2),info.Ne[it2]),dtype=np.float32) ) for it1 in info.Nt[ist]: for ia1 in range(info.Na[ist][it1]): for il1 in range(info.Nl[it1]): for im1 in range(info.Nm(il1)): for it2 in info.Nt[ist]: for ia2 in range(info.Na[ist][it2]): for il2 in range(info.Nl[it2]): for im2 in range(info.Nm(il2)): for ie1 in range(info.Ne[it1]): for ie2 in range(info.Ne[it2]): if not line: line = file.readline().split() SI[it1,it2][il1][il2].real[ia1,im1,ie1,ia2,im2,ie2] = float(line.pop(0)) if not line: line = file.readline().split() SI[it1,it2][il1][il2].imag[ia1,im1,ie1,ia2,im2,ie2] = float(line.pop(0)) for it1,it2 in itertools.product( info.Nt[ist], info.Nt[ist] ): for il1,il2 in itertools.product( range(info.Nl[it1]), range(info.Nl[it2]) ): SI[it1,it2][il1][il2] = torch_complex.ComplexTensor( torch.from_numpy(SI[it1,it2][il1][il2].real), torch.from_numpy(SI[it1,it2][il1][il2].imag)) return SI,line
def stft(
time_signal,
size: int = 1024,
shift: int = 256,
*,
# axis=-1, # I never use this and it complicated the code
window: [str, typing.Callable] = 'blackman',
window_length: int = None,
fading: typing.Optional[typing.Union[bool, str]] = 'full',
pad: bool = True,
symmetric_window: bool = False,
):
"""
>>> import numpy as np
>>> import random
>>> from paderbox.transform.module_stft import stft as np_stft, istft as np_istft
>>> kwargs = dict(
... size=np.random.randint(100, 200),
... shift=np.random.randint(40, 100),
... window=random.choice(['blackman', 'hann', 'hamming']),
... fading=random.choice(['full', 'half', False]),
... )
>>> num_samples = np.random.randint(200, 500)
>>> a = np.random.rand(num_samples)
>>> A_np = np_stft(a, **kwargs)
>>> A_pt = stft(torch.tensor(a), **kwargs)
>>> np.testing.assert_allclose(
... A_np, A_pt.numpy(), err_msg=str(kwargs), atol=1e-10)
"""
assert isinstance(time_signal, torch.Tensor)
if window_length is None:
window_length = size
else:
if window_length != size:
raise NotImplementedError(
'Torch does not support window_length != size\n'
'window_length = {window_length} != {size} = size')
# Pad with zeros to have enough samples for the window function to fade.
assert fading in [None, True, False, 'full',
'half'], (fading, type(fading))
if fading not in [False, None]:
if fading == 'half':
pad_width = [
(window_length - shift) // 2,
math.ceil((window_length - shift) / 2),
]
else:
pad_width = [
window_length - shift,
window_length - shift,
]
time_signal = torch.nn.functional.pad(time_signal,
pad_width,
mode='constant')
window = _get_window(
window=window,
symmetric_window=symmetric_window,
window_length=window_length,
)
time_signal_seg = segment_axis(time_signal,
window_length,
shift=shift,
axis=-1,
end='pad' if pad else 'cut')
out = torch.rfft(
time_signal_seg * window,
1,
# size,
)
assert out.shape[-1] == 2, out.shape
return torch_complex.ComplexTensor(out[..., 0], out[..., 1])
def complex_numpy(self, example, device):
import torch_complex
return torch_complex.ComplexTensor(example, device=device)