def split_pair_to_vars(sample_batch_pair):
"""
Splits the generated batch data and creates combination of pairs.
Input argument sample_batch_pair consists of a batch_size number of
[clean_signal, noisy_signal] pairs.
This function creates three pytorch Variables - a clean_signal, noisy_signal pair,
clean signal only, and noisy signal only.
It goes through preemphasis preprocessing before converted into variable.
Args:
sample_batch_pair(torch.Tensor): batch of [clean_signal, noisy_signal] pairs
Returns:
batch_pairs_var(Variable): batch of pairs containing clean signal and noisy signal
clean_batch_var(Variable): clean signal batch
noisy_batch_var(Varialbe): noisy signal batch
"""
# preemphasis
sample_batch_pair = emph.pre_emphasis(sample_batch_pair.numpy(),
emph_coeff=0.95)
batch_pairs_var = Variable(
torch.from_numpy(sample_batch_pair).type(
torch.FloatTensor)).cuda() # [40 x 2 x 16384]
clean_batch = np.stack(
[pair[0].reshape(1, -1) for pair in sample_batch_pair])
clean_batch_var = Variable(
torch.from_numpy(clean_batch).type(torch.FloatTensor)).cuda()
noisy_batch = np.stack(
[pair[1].reshape(1, -1) for pair in sample_batch_pair])
noisy_batch_var = Variable(
torch.from_numpy(noisy_batch).type(torch.FloatTensor)).cuda()
return batch_pairs_var, clean_batch_var, noisy_batch_var
def test_pre_emphasis(self):
"""
Tests equality after de-emphasizing pre-emphasized signal.
"""
rand_signal_batch = np.random.randint(low=1,
high=10,
size=(10, 1, 400))
reconst_batch = emph.de_emphasis(emph.pre_emphasis(rand_signal_batch))
# after de-emphasis, the signal must have been restored
self.assertEqual(rand_signal_batch.shape, reconst_batch.shape)
self.assertTrue(np.allclose(rand_signal_batch, reconst_batch))