def test_stft_raises_for_invalid_dimensions(self):
sr = SR22050()
samples = SilenceSynthesizer(sr).synthesize(Milliseconds(6666))
wscheme = sr.windowing_scheme(512, 256)
tf = stft(samples, wscheme, HanningWindowingFunc())
self.assertRaises(ValueError,
lambda: stft(tf, wscheme, HanningWindowingFunc()))
def test_has_correct_shape(self):
sr = SR22050()
samples = SilenceSynthesizer(sr).synthesize(Milliseconds(9999))
wscheme = sr.windowing_scheme(256, 128)
scale = GeometricScale(50, sr.nyquist, 0.4, 32)
scale.ensure_overlap_ratio()
tf = stft(samples, wscheme, HanningWindowingFunc())
geom = apply_scale(tf, scale, window=HanningWindowingFunc())
self.assertEqual(tf.shape[:-1] + (len(scale), ), geom.shape)
def test_apply_scale_to_self_is_identity_function(self):
samplerate = SR22050()
samples = SineSynthesizer(samplerate).synthesize(Milliseconds(8888))
wscheme = samplerate.windowing_scheme(256, 128)
tf = stft(samples, wscheme, HanningWindowingFunc())
scale = tf.dimensions[-1].scale
transformed = apply_scale(tf, scale, HanningWindowingFunc())
self.assertEqual(tf.shape, transformed.shape)
self.assertEqual(tf.dimensions[0], transformed.dimensions[0])
self.assertEqual(tf.dimensions[1], transformed.dimensions[1])
def test_has_correct_dimensions(self):
sr = SR22050()
samples = SilenceSynthesizer(sr).synthesize(Milliseconds(9999))
wscheme = sr.windowing_scheme(256, 128)
scale = GeometricScale(50, sr.nyquist, 0.4, 32)
scale.ensure_overlap_ratio()
tf = stft(samples, wscheme, HanningWindowingFunc())
geom = apply_scale(tf, scale, window=HanningWindowingFunc())
self.assertIsInstance(geom, ArrayWithUnits)
self.assertEqual(2, len(geom.dimensions))
self.assertIsInstance(geom.dimensions[0], TimeDimension)
self.assertEqual(geom.dimensions[0].samplerate, wscheme)
self.assertIsInstance(geom.dimensions[1], FrequencyDimension)
self.assertEqual(scale, geom.dimensions[1].scale)
def _stft(self, samples):
samples = samples / np.abs(samples.max())
wscheme = SampleRate(
frequency=samples.samplerate.frequency * 128,
duration=samples.samplerate.frequency * 256)
coeffs = stft(samples, wscheme, HanningWindowingFunc())
return rainbowgram(coeffs)
def test_preserves_time_dimension(self):
sr = SR22050()
samples = TickSynthesizer(sr).synthesize(Milliseconds(10000),
Milliseconds(1000))
wscheme = sr.windowing_scheme(256, 128)
scale = GeometricScale(50, sr.nyquist, 0.4, 32)
scale.ensure_overlap_ratio()
tf = stft(samples, wscheme, HanningWindowingFunc())
geom = apply_scale(tf, scale, window=HanningWindowingFunc())
# get the loudness envelope of each
tf_envelope = np.abs(tf.real).sum(axis=1)
geom_envelope = geom.sum(axis=1)
tf_zeros = np.where(tf_envelope == 0)
geom_zeros = np.where(geom_envelope == 0)
np.testing.assert_allclose(tf_zeros, geom_zeros)
def test_has_correct_dimensions(self):
sr = SR22050()
samples = SilenceSynthesizer(sr).synthesize(Milliseconds(6666))
wscheme = sr.windowing_scheme(512, 256)
tf = stft(samples, wscheme, HanningWindowingFunc())
self.assertIsInstance(tf, ArrayWithUnits)
self.assertEqual(2, len(tf.dimensions))
self.assertIsInstance(tf.dimensions[0], TimeDimension)
self.assertEqual(tf.dimensions[0].samplerate, wscheme)
self.assertIsInstance(tf.dimensions[1], FrequencyDimension)
self.assertIsInstance(tf.dimensions[1].scale, LinearScale)
def test_should_have_correct_shape_and_dimensions(self):
samplerate = SR22050()
samples = SineSynthesizer(samplerate).synthesize(Milliseconds(8888))
wscheme = samplerate.windowing_scheme(256, 128)
tf = stft(samples, wscheme, HanningWindowingFunc())
rg = rainbowgram(tf, cm.rainbow)
self.assertEqual(3, rg.ndim)
self.assertEqual(tf.shape + (3, ), rg.shape)
self.assertEqual(3, len(rg.dimensions))
self.assertEqual(tf.dimensions[0], rg.dimensions[0])
self.assertEqual(tf.dimensions[1], rg.dimensions[1])
self.assertEqual(rg.dimensions[2], IdentityDimension())
self.assertEqual(3, rg.shape[-1])
def test_can_take_stft_of_batch(self):
sr = SR22050()
samples = SilenceSynthesizer(sr).synthesize(Milliseconds(6666))
stacked = ArrayWithUnits(
np.zeros((10, ) + samples.shape, dtype=samples.dtype),
(IdentityDimension(), ) + samples.dimensions)
stacked[:] = samples
wscheme = sr.windowing_scheme(512, 256)
tf = stft(stacked, wscheme, HanningWindowingFunc())
self.assertEqual(10, len(tf))
self.assertIsInstance(tf, ArrayWithUnits)
self.assertEqual(3, len(tf.dimensions))
self.assertIsInstance(tf.dimensions[0], IdentityDimension)
self.assertIsInstance(tf.dimensions[1], TimeDimension)
self.assertEqual(tf.dimensions[1].samplerate, wscheme)
self.assertIsInstance(tf.dimensions[2], FrequencyDimension)
self.assertIsInstance(tf.dimensions[2].scale, LinearScale)
def test_can_do_short_time_dct_transform(self):
t = torch.FloatTensor(3, 1, 512)
v = Variable(t)
dct_trasform = DctTransform()
stdct = dct_trasform.short_time_dct(v, 128, 64, HanningWindowingFunc())
self.assertEqual((3, 128, 7), tuple(stdct.shape))
def test_has_correct_number_of_bins(self):
sr = SR22050()
samples = SilenceSynthesizer(sr).synthesize(Milliseconds(6666))
wscheme = sr.windowing_scheme(512, 256)
tf = stft(samples, wscheme, HanningWindowingFunc())
self.assertEqual(tf.shape[1], 257)