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
|
import unittest
import numpy
from become_yukarin.dataset import dataset
class TestDataset(unittest.TestCase):
def setUp(self):
self.sample_rate = 24000
self.len_time = len_time = 100
self.fft_size = fft_size = 1024
self.order = order = 59
self.dummy_feature = dataset.AcousticFeature(
f0=numpy.arange(len_time).reshape((len_time, -1)),
spectrogram=numpy.arange(len_time * (fft_size // 2 + 1)).reshape((len_time, -1)),
aperiodicity=numpy.arange(len_time * (fft_size // 2 + 1)).reshape((len_time, -1)),
mfcc=numpy.arange(len_time * (order + 1)).reshape((len_time, -1)),
voiced=(numpy.arange(len_time) % 2 == 1).reshape((len_time, -1)),
)
self.feature_sizes = dataset.AcousticFeature.get_sizes(
sampling_rate=self.sample_rate,
order=self.order,
)
def test_encode_decode_feature(self):
encode_feature = dataset.EncodeFeatureProcess(['mfcc'])
decode_feature = dataset.DecodeFeatureProcess(['mfcc'], self.feature_sizes)
e = encode_feature(self.dummy_feature, test=True)
d = decode_feature(e, test=True)
self.assertTrue(numpy.all(self.dummy_feature.mfcc == d.mfcc))
def test_encode_decode_feature2(self):
encode_feature = dataset.EncodeFeatureProcess(['mfcc', 'f0'])
decode_feature = dataset.DecodeFeatureProcess(['mfcc', 'f0'], self.feature_sizes)
e = encode_feature(self.dummy_feature, test=True)
d = decode_feature(e, test=True)
self.assertTrue(numpy.all(self.dummy_feature.mfcc == d.mfcc))
self.assertTrue(numpy.all(self.dummy_feature.f0 == d.f0))
def test_encode_decode_feature3(self):
encode_feature = dataset.EncodeFeatureProcess(['mfcc', 'f0'])
decode_feature = dataset.DecodeFeatureProcess(['mfcc', 'f0'], self.feature_sizes)
e = encode_feature(self.dummy_feature, test=True)
e[0] = numpy.nan
d = decode_feature(e, test=True)
self.assertFalse(numpy.all(self.dummy_feature.mfcc == d.mfcc))
self.assertTrue(numpy.all(self.dummy_feature.f0 == d.f0))
if __name__ == '__main__':
unittest.main()
|
