File size: 6,786 Bytes
be9690e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
""" from https://github.com/jik876/hifi-gan """

import math
import os
import random

import numpy as np
import torch
import torch.utils.data
from librosa.filters import mel as librosa_mel_fn
from librosa.util import normalize
from scipy.io.wavfile import read

MAX_WAV_VALUE = 32768.0


def load_wav(full_path):
    sampling_rate, data = read(full_path)
    return data, sampling_rate


def dynamic_range_compression(x, C=1, clip_val=1e-5):
    return np.log(np.clip(x, a_min=clip_val, a_max=None) * C)


def dynamic_range_decompression(x, C=1):
    return np.exp(x) / C


def dynamic_range_compression_torch(x, C=1, clip_val=1e-5):
    return torch.log(torch.clamp(x, min=clip_val) * C)


def dynamic_range_decompression_torch(x, C=1):
    return torch.exp(x) / C


def spectral_normalize_torch(magnitudes):
    output = dynamic_range_compression_torch(magnitudes)
    return output


def spectral_de_normalize_torch(magnitudes):
    output = dynamic_range_decompression_torch(magnitudes)
    return output


mel_basis = {}
hann_window = {}


def mel_spectrogram(y, n_fft, num_mels, sampling_rate, hop_size, win_size, fmin, fmax, center=False):
    if torch.min(y) < -1.0:
        print("min value is ", torch.min(y))
    if torch.max(y) > 1.0:
        print("max value is ", torch.max(y))

    global mel_basis, hann_window  # pylint: disable=global-statement
    if fmax not in mel_basis:
        mel = librosa_mel_fn(sampling_rate, n_fft, num_mels, fmin, fmax)
        mel_basis[str(fmax) + "_" + str(y.device)] = torch.from_numpy(mel).float().to(y.device)
        hann_window[str(y.device)] = torch.hann_window(win_size).to(y.device)

    y = torch.nn.functional.pad(
        y.unsqueeze(1), (int((n_fft - hop_size) / 2), int((n_fft - hop_size) / 2)), mode="reflect"
    )
    y = y.squeeze(1)

    spec = torch.view_as_real(
        torch.stft(
            y,
            n_fft,
            hop_length=hop_size,
            win_length=win_size,
            window=hann_window[str(y.device)],
            center=center,
            pad_mode="reflect",
            normalized=False,
            onesided=True,
            return_complex=True,
        )
    )

    spec = torch.sqrt(spec.pow(2).sum(-1) + (1e-9))

    spec = torch.matmul(mel_basis[str(fmax) + "_" + str(y.device)], spec)
    spec = spectral_normalize_torch(spec)

    return spec


def get_dataset_filelist(a):
    with open(a.input_training_file, encoding="utf-8") as fi:
        training_files = [
            os.path.join(a.input_wavs_dir, x.split("|")[0] + ".wav") for x in fi.read().split("\n") if len(x) > 0
        ]

    with open(a.input_validation_file, encoding="utf-8") as fi:
        validation_files = [
            os.path.join(a.input_wavs_dir, x.split("|")[0] + ".wav") for x in fi.read().split("\n") if len(x) > 0
        ]
    return training_files, validation_files


class MelDataset(torch.utils.data.Dataset):
    def __init__(
        self,
        training_files,
        segment_size,
        n_fft,
        num_mels,
        hop_size,
        win_size,
        sampling_rate,
        fmin,
        fmax,
        split=True,
        shuffle=True,
        n_cache_reuse=1,
        device=None,
        fmax_loss=None,
        fine_tuning=False,
        base_mels_path=None,
    ):
        self.audio_files = training_files
        random.seed(1234)
        if shuffle:
            random.shuffle(self.audio_files)
        self.segment_size = segment_size
        self.sampling_rate = sampling_rate
        self.split = split
        self.n_fft = n_fft
        self.num_mels = num_mels
        self.hop_size = hop_size
        self.win_size = win_size
        self.fmin = fmin
        self.fmax = fmax
        self.fmax_loss = fmax_loss
        self.cached_wav = None
        self.n_cache_reuse = n_cache_reuse
        self._cache_ref_count = 0
        self.device = device
        self.fine_tuning = fine_tuning
        self.base_mels_path = base_mels_path

    def __getitem__(self, index):
        filename = self.audio_files[index]
        if self._cache_ref_count == 0:
            audio, sampling_rate = load_wav(filename)
            audio = audio / MAX_WAV_VALUE
            if not self.fine_tuning:
                audio = normalize(audio) * 0.95
            self.cached_wav = audio
            if sampling_rate != self.sampling_rate:
                raise ValueError(f"{sampling_rate} SR doesn't match target {self.sampling_rate} SR")
            self._cache_ref_count = self.n_cache_reuse
        else:
            audio = self.cached_wav
            self._cache_ref_count -= 1

        audio = torch.FloatTensor(audio)
        audio = audio.unsqueeze(0)

        if not self.fine_tuning:
            if self.split:
                if audio.size(1) >= self.segment_size:
                    max_audio_start = audio.size(1) - self.segment_size
                    audio_start = random.randint(0, max_audio_start)
                    audio = audio[:, audio_start : audio_start + self.segment_size]
                else:
                    audio = torch.nn.functional.pad(audio, (0, self.segment_size - audio.size(1)), "constant")

            mel = mel_spectrogram(
                audio,
                self.n_fft,
                self.num_mels,
                self.sampling_rate,
                self.hop_size,
                self.win_size,
                self.fmin,
                self.fmax,
                center=False,
            )
        else:
            mel = np.load(os.path.join(self.base_mels_path, os.path.splitext(os.path.split(filename)[-1])[0] + ".npy"))
            mel = torch.from_numpy(mel)

            if len(mel.shape) < 3:
                mel = mel.unsqueeze(0)

            if self.split:
                frames_per_seg = math.ceil(self.segment_size / self.hop_size)

                if audio.size(1) >= self.segment_size:
                    mel_start = random.randint(0, mel.size(2) - frames_per_seg - 1)
                    mel = mel[:, :, mel_start : mel_start + frames_per_seg]
                    audio = audio[:, mel_start * self.hop_size : (mel_start + frames_per_seg) * self.hop_size]
                else:
                    mel = torch.nn.functional.pad(mel, (0, frames_per_seg - mel.size(2)), "constant")
                    audio = torch.nn.functional.pad(audio, (0, self.segment_size - audio.size(1)), "constant")

        mel_loss = mel_spectrogram(
            audio,
            self.n_fft,
            self.num_mels,
            self.sampling_rate,
            self.hop_size,
            self.win_size,
            self.fmin,
            self.fmax_loss,
            center=False,
        )

        return (mel.squeeze(), audio.squeeze(0), filename, mel_loss.squeeze())

    def __len__(self):
        return len(self.audio_files)