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): use find_endpoint from numpy_transforms

This commit is contained in:
Enno Hermann 2023-11-14 13:37:43 +01:00
parent 5232bf9e36
commit 842a632cd5
1 changed files with 9 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
TTS/utils/audio/processor.py
View File

@ -14,6 +14,7 @@ from TTS.utils.audio.numpy_transforms import (
compute_f0, compute_f0,
db_to_amp, db_to_amp,
deemphasis, deemphasis,
find_endpoint,
griffin_lim, griffin_lim,
mel_to_spec, mel_to_spec,
millisec_to_length, millisec_to_length,
@ -527,13 +528,14 @@ class AudioProcessor(object):
Returns: Returns:
int: Last point without silence. int: Last point without silence.
""" """
window_length = int(self.sample_rate * min_silence_sec) return find_endpoint(
hop_length = int(window_length / 4) wav=wav,
threshold = db_to_amp(x=-self.trim_db, gain=self.spec_gain, base=self.base) trim_db=self.trim_db,
for x in range(hop_length, len(wav) - window_length, hop_length): sample_rate=self.sample_rate,
if np.max(wav[x : x + window_length]) < threshold: min_silence_sec=min_silence_sec,
return x + hop_length gain=self.spec_gain,
return len(wav) base=self.base,
)
def trim_silence(self, wav): def trim_silence(self, wav):
"""Trim silent parts with a threshold and 0.01 sec margin""" """Trim silent parts with a threshold and 0.01 sec margin"""