File size: 6,205 Bytes
62bee25
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
import torch
import numpy as np
import torch.nn.functional as F
from torch.autograd import Variable
from scipy.signal import get_window
import librosa.util as librosa_util
from librosa.util import pad_center, tiny
# from audio_processing import window_sumsquare

def window_sumsquare(window, n_frames, hop_length=512, win_length=1024,
                     n_fft=1024, dtype=np.float32, norm=None):
    """
    # from librosa 0.6
    Compute the sum-square envelope of a window function at a given hop length.
    This is used to estimate modulation effects induced by windowing
    observations in short-time fourier transforms.
    Parameters
    ----------
    window : string, tuple, number, callable, or list-like
        Window specification, as in `get_window`
    n_frames : int > 0
        The number of analysis frames
    hop_length : int > 0
        The number of samples to advance between frames
    win_length : [optional]
        The length of the window function.  By default, this matches `n_fft`.
    n_fft : int > 0
        The length of each analysis frame.
    dtype : np.dtype
        The data type of the output
    Returns
    -------
    wss : np.ndarray, shape=`(n_fft + hop_length * (n_frames - 1))`
        The sum-squared envelope of the window function
    """
    if win_length is None:
        win_length = n_fft

    n = n_fft + hop_length * (n_frames - 1)
    x = np.zeros(n, dtype=dtype)

    # Compute the squared window at the desired length
    win_sq = get_window(window, win_length, fftbins=True)
    win_sq = librosa_util.normalize(win_sq, norm=norm)**2
    win_sq = librosa_util.pad_center(win_sq, n_fft)

    # Fill the envelope
    for i in range(n_frames):
        sample = i * hop_length
        x[sample:min(n, sample + n_fft)] += win_sq[:max(0, min(n_fft, n - sample))]
    return x

class STFT(torch.nn.Module):
    """adapted from Prem Seetharaman's https://github.com/pseeth/pytorch-stft"""
    def __init__(self, filter_length=1024, hop_length=512, win_length=1024,
                 window='hann'):
        super(STFT, self).__init__()
        self.filter_length = filter_length
        self.hop_length = hop_length
        self.win_length = win_length
        self.window = window
        self.forward_transform = None
        scale = self.filter_length / self.hop_length
        fourier_basis = np.fft.fft(np.eye(self.filter_length))

        cutoff = int((self.filter_length / 2 + 1))
        fourier_basis = np.vstack([np.real(fourier_basis[:cutoff, :]),
                                   np.imag(fourier_basis[:cutoff, :])])

        forward_basis = torch.FloatTensor(fourier_basis[:, None, :])
        inverse_basis = torch.FloatTensor(
            np.linalg.pinv(scale * fourier_basis).T[:, None, :])

        if window is not None:
            assert(filter_length >= win_length)
            # get window and zero center pad it to filter_length
            fft_window = get_window(window, win_length, fftbins=True)
            fft_window = pad_center(fft_window, filter_length)
            fft_window = torch.from_numpy(fft_window).float()

            # window the bases
            forward_basis *= fft_window
            inverse_basis *= fft_window

        self.register_buffer('forward_basis', forward_basis.float())
        self.register_buffer('inverse_basis', inverse_basis.float())

    def transform(self, input_data):
        num_batches = input_data.size(0)
        num_samples = input_data.size(1)

        self.num_samples = num_samples

        # similar to librosa, reflect-pad the input
        input_data = input_data.view(num_batches, 1, num_samples)
        input_data = F.pad(
            input_data.unsqueeze(1),
            (int(self.filter_length / 2), int(self.filter_length / 2), 0, 0),
            mode='reflect')
        input_data = input_data.squeeze(1)

        forward_transform = F.conv1d(
            input_data,
            Variable(self.forward_basis, requires_grad=False),
            stride=self.hop_length,
            padding=0)

        cutoff = int((self.filter_length / 2) + 1)
        real_part = forward_transform[:, :cutoff, :]
        imag_part = forward_transform[:, cutoff:, :]

        magnitude = torch.sqrt(real_part**2 + imag_part**2)
        phase = torch.autograd.Variable(
            torch.atan2(imag_part.data, real_part.data))

        return magnitude, phase # [batch_size, F(513), T(1251)]

    def inverse(self, magnitude, phase):
        recombine_magnitude_phase = torch.cat(
            [magnitude*torch.cos(phase), magnitude*torch.sin(phase)], dim=1)

        inverse_transform = F.conv_transpose1d(
            recombine_magnitude_phase,
            Variable(self.inverse_basis, requires_grad=False),
            stride=self.hop_length,
            padding=0)

        if self.window is not None:
            window_sum = window_sumsquare(
                self.window, magnitude.size(-1), hop_length=self.hop_length,
                win_length=self.win_length, n_fft=self.filter_length,
                dtype=np.float32)
            # remove modulation effects
            approx_nonzero_indices = torch.from_numpy(
                np.where(window_sum > tiny(window_sum))[0])
            window_sum = torch.autograd.Variable(
                torch.from_numpy(window_sum), requires_grad=False)
            window_sum = window_sum.cuda() if magnitude.is_cuda else window_sum
            inverse_transform[:, :, approx_nonzero_indices] /= window_sum[approx_nonzero_indices]

            # scale by hop ratio
            inverse_transform *= float(self.filter_length) / self.hop_length

        inverse_transform = inverse_transform[:, :, int(self.filter_length/2):]
        inverse_transform = inverse_transform[:, :, :-int(self.filter_length/2):]

        return inverse_transform #[batch_size, 1, sample_num]

    def forward(self, input_data):
        self.magnitude, self.phase = self.transform(input_data)
        reconstruction = self.inverse(self.magnitude, self.phase)
        return reconstruction 

if __name__ == '__main__':
    a = torch.randn(4, 320000)
    stft = STFT()
    mag, phase = stft.transform(a)
    # rec_a = stft.inverse(mag, phase)
    print(mag.shape)