def call(self, audio_data, sample_rate):
"""
Caculate fbank features of audio data.
:param audio_data: the audio signal from which to compute spectrum. Should be an (1, N) tensor.
:param sample_rate: the samplerate of the signal we working with.
:return: A float tensor of size (num_channels, num_frames, num_frequencies) containing
fbank features of every frame in speech.
"""
p = self.config
with tf.name_scope('fbank'):
spectrum = self.spect(audio_data, sample_rate)
spectrum = tf.expand_dims(spectrum, 0)
sample_rate = tf.cast(sample_rate, dtype=tf.int32)
fbank = py_x_ops.fbank(
spectrum,
sample_rate,
upper_frequency_limit=p.upper_frequency_limit,
lower_frequency_limit=p.lower_frequency_limit,
filterbank_channel_count=p.filterbank_channel_count)
fbank = tf.squeeze(fbank, axis=0)
shape = tf.shape(fbank)
nframe = shape[0]
nfbank = shape[1]
if p.delta_delta:
fbank = py_x_ops.delta_delta(fbank, p.order, p.window)
if p.type == 'Fbank':
fbank = self.cmvn(fbank)
fbank = tf.reshape(fbank, (nframe, nfbank, p.channel))
return fbank
def call(self, audio_data, sample_rate):
"""Caculate fbank features of audio data.
Args:
audio_data: the audio signal from which to compute spectrum.
sample_rate: the sample rate of the signal we working with.
Shape:
- audio_data: :math:`(1, N)`
- sample_rate: float
"""
p = self.config
with tf.name_scope('fbank'):
spectrum = self.spect(audio_data, sample_rate)
spectrum = tf.expand_dims(spectrum, 0)
sample_rate = tf.cast(sample_rate, dtype=tf.int32)
fbank = py_x_ops.fbank(
spectrum,
sample_rate,
upper_frequency_limit=p.upper_frequency_limit,
lower_frequency_limit=p.lower_frequency_limit,
filterbank_channel_count=p.filterbank_channel_count,
is_log10=p.is_log10)
fbank = tf.squeeze(fbank, axis=0)
shape = tf.shape(fbank)
nframe = shape[0]
nfbank = shape[1]
if p.delta_delta:
fbank = py_x_ops.delta_delta(fbank, p.order, p.window)
if p.type == 'Fbank':
fbank = self.cmvn(fbank)
fbank = tf.reshape(fbank, (nframe, nfbank, p.channel))
return fbank