Medix-AI/coqui-tts
Medix-AI
/
coqui-tts
mirror of https://github.com/coqui-ai/TTS.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

refactor(audio.processor): remove duplicate mel_to_linear

This commit is contained in:
Enno Hermann 2023-11-14 10:44:35 +01:00
parent fd9d6d4b0f
commit 754877784b
1 changed files with 10 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
TTS/utils/audio/processor.py
View File

@ -8,7 +8,15 @@ import scipy.signal
import soundfile as sf import soundfile as sf
from TTS.tts.utils.helpers import StandardScaler from TTS.tts.utils.helpers import StandardScaler
from TTS.utils.audio.numpy_transforms import amp_to_db, compute_f0, db_to_amp, griffin_lim, spec_to_mel, stft from TTS.utils.audio.numpy_transforms import (
amp_to_db,
compute_f0,
db_to_amp,
griffin_lim,
mel_to_spec,
spec_to_mel,
stft,
)
# pylint: disable=too-many-public-methods # pylint: disable=too-many-public-methods
@ -216,7 +224,6 @@ class AudioProcessor(object):
print(" | > {}:{}".format(key, value)) print(" | > {}:{}".format(key, value))
# create spectrogram utils # create spectrogram utils
self.mel_basis = self._build_mel_basis() self.mel_basis = self._build_mel_basis()
self.inv_mel_basis = np.linalg.pinv(self._build_mel_basis())
# setup scaler # setup scaler
if stats_path and signal_norm: if stats_path and signal_norm:
mel_mean, mel_std, linear_mean, linear_std, _ = self.load_stats(stats_path) mel_mean, mel_std, linear_mean, linear_std, _ = self.load_stats(stats_path)
@ -410,10 +417,6 @@ class AudioProcessor(object):
return scipy.signal.lfilter([1], [1, -self.preemphasis], x) return scipy.signal.lfilter([1], [1, -self.preemphasis], x)
### SPECTROGRAMs ### ### SPECTROGRAMs ###
def _mel_to_linear(self, mel_spec: np.ndarray) -> np.ndarray:
"""Convert a melspectrogram to full scale spectrogram."""
return np.maximum(1e-10, np.dot(self.inv_mel_basis, mel_spec))
def spectrogram(self, y: np.ndarray) -> np.ndarray: def spectrogram(self, y: np.ndarray) -> np.ndarray:
"""Compute a spectrogram from a waveform. """Compute a spectrogram from a waveform.
@ -467,7 +470,7 @@ class AudioProcessor(object):
"""Convert a melspectrogram to a waveform using Griffi-Lim vocoder.""" """Convert a melspectrogram to a waveform using Griffi-Lim vocoder."""
D = self.denormalize(mel_spectrogram) D = self.denormalize(mel_spectrogram)
S = db_to_amp(x=D, gain=self.spec_gain, base=self.base) S = db_to_amp(x=D, gain=self.spec_gain, base=self.base)
S = self._mel_to_linear(S) # Convert back to linear S = mel_to_spec(mel=S, mel_basis=self.mel_basis) # Convert back to linear
W = self._griffin_lim(S**self.power) W = self._griffin_lim(S**self.power)
return self.apply_inv_preemphasis(W) if self.preemphasis != 0 else W return self.apply_inv_preemphasis(W) if self.preemphasis != 0 else W