def __init__(self, config: dict):
super().__init__(config)
config1 = copy.deepcopy(config)
config1['is_fbank'] = False
config1['output_type'] = 2
self.spect = Spectrum(config1)
self.fbank = Fbank(config)
def compute_fbank():
parser = get_parser()
args = parser.parse_args()
config = {}
config['sample_rate'] = float(args.sample_rate)
config['upper_frequency_limit'] = float(args.upper_frequency_limit)
config['lower_frequency_limit'] = float(args.lower_frequency_limit)
config['filterbank_channel_count'] = float(args.filterbank_channel_count)
config['window_length'] = args.window_length
config['frame_length'] = args.frame_length
config['output_type'] = args.output_type
fbank = Fbank.params(config).instantiate()
with kaldiio.ReadHelper(args.rspecifier,
segments=args.segments) as reader, \
KaldiWriter(args.wspecifier, write_num_frames=args.write_num_frames,
compress=args.compress, compression_method=args.compression_method) as writer:
for utt_id, (sample_rate, array) in reader:
if sample_rate != args.sample_rate:
args.sample_rate = sample_rate
array = array.astype(np.float32)
audio_data = tf.constant(array, dtype=tf.float32)
fbank_test = tf.squeeze(fbank(audio_data, args.sample_rate))
sess = tf.compat.v1.Session()
fbank_feats = fbank_test.eval(session=sess)
writer[utt_id] = fbank_feats
def test_fbank(self):
wav_path = str(Path(PACKAGE_OPS_DIR).joinpath('data/sm1_cln.wav'))
with self.cached_session(use_gpu=False, force_gpu=False):
read_wav = ReadWav.params().instantiate()
input_data, sample_rate = read_wav(wav_path)
config = {
'window_length': 0.025,
'output_type': 1,
'frame_length': 0.010,
'snip_edges': True
}
fbank = Fbank.params(config).instantiate()
fbank_test = fbank(input_data, sample_rate)
self.assertEqual(tf.rank(fbank_test).eval(), 3)
real_fank_feats = np.array(
[[3.768338, 4.946218, 6.289874, 6.330853, 6.761764, 6.884573],
[3.803553, 5.450971, 6.547878, 5.796172, 6.397846, 7.242926]])
self.assertAllClose(
np.squeeze(fbank_test.eval()[0:2, 0:6, 0]),
real_fank_feats,
rtol=1e-05,
atol=1e-05)
def test_fbank(self):
wav_path = str(
Path(PACKAGE_ROOT_DIR).joinpath(
'layers/ops/data/sm1_cln.wav'))
with self.cached_session(use_gpu=False, force_gpu=False):
read_wav = ReadWav.params().instantiate()
input_data, sample_rate = read_wav(wav_path)
config = {'window_length': 0.025, 'output_type': 1, 'frame_length': 0.010}
fbank = Fbank.params(config).instantiate()
fbank_test = fbank(input_data, sample_rate)
self.assertEqual(tf.rank(fbank_test).eval(), 3)
def test_fbank(self):
wav_path = str(
Path(os.environ['MAIN_ROOT']).joinpath(
'delta/layers/ops/data/sm1_cln.wav'))
with self.session():
read_wav = ReadWav.params().instantiate()
input_data, sample_rate = read_wav(wav_path)
config = {
'window_length': 0.025,
'output_type': 1,
'frame_length': 0.010
}
fbank = Fbank.params(config).instantiate()
fbank_test = fbank(input_data, sample_rate)
self.assertEqual(tf.rank(fbank_test).eval(), 3)
def test_fbank(self):
wav_path = str(Path(PACKAGE_OPS_DIR).joinpath('data/sm1_cln.wav'))
with self.cached_session(use_gpu=False, force_gpu=False):
read_wav = ReadWav.params().instantiate()
input_data, sample_rate = read_wav(wav_path)
config = {
'window_length': 0.025,
'output_type': 1,
'frame_length': 0.010,
'snip_edges': True
}
fbank = Fbank.params(config).instantiate()
fbank_test = fbank(input_data, sample_rate)
self.assertEqual(tf.rank(fbank_test).eval(), 3)
if tf.executing_eagerly():
print(fbank_test.numpy()[0:2, 0:6, 0])
else:
print(fbank_test.eval()[0:2, 0:6, 0])
def __init__(self, config: dict):
super().__init__(config)
self.fbank = Fbank(config)
self.pitch = Pitch(config)
def __init__(self, config: dict):
super().__init__(config)
self.framepow = Framepow(config)
self.fbank = Fbank(config)