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

Fix style tests

This commit is contained in:
Edresson Casanova 2022-03-23 15:05:32 -03:00
parent 0ae1e0248c
commit 3435bc8fca
2 changed files with 22 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
TTS/bin/remove_silence_using_vad.py
View File

@ -4,6 +4,7 @@ import os
import pathlib import pathlib
from tqdm import tqdm from tqdm import tqdm
from TTS.utils.vad import get_vad_model_and_utils, remove_silence from TTS.utils.vad import get_vad_model_and_utils, remove_silence
@ -16,7 +17,13 @@ def adjust_path_and_remove_silence(audio_path):
# create all directory structure # create all directory structure
pathlib.Path(output_path).parent.mkdir(parents=True, exist_ok=True) pathlib.Path(output_path).parent.mkdir(parents=True, exist_ok=True)
# remove the silence and save the audio # remove the silence and save the audio
output_path = remove_silence(model_and_utils, audio_path, output_path, trim_just_beginning_and_end=args.trim_just_beginning_and_end, use_cuda=args.use_cuda) output_path = remove_silence(
model_and_utils,
audio_path,
output_path,
trim_just_beginning_and_end=args.trim_just_beginning_and_end,
use_cuda=args.use_cuda,
)
return output_path return output_path

22
TTS/utils/vad.py
View File

@ -1,6 +1,7 @@
import torch import torch
import torchaudio import torchaudio
def read_audio(path): def read_audio(path):
wav, sr = torchaudio.load(path) wav, sr = torchaudio.load(path)
@ -9,39 +10,42 @@ def read_audio(path):
return wav.squeeze(0), sr return wav.squeeze(0), sr
def resample_wav(wav, sr, new_sr): def resample_wav(wav, sr, new_sr):
wav = wav.unsqueeze(0) wav = wav.unsqueeze(0)
transform = torchaudio.transforms.Resample(orig_freq=sr, new_freq=new_sr) transform = torchaudio.transforms.Resample(orig_freq=sr, new_freq=new_sr)
wav = transform(wav) wav = transform(wav)
return wav.squeeze(0) return wav.squeeze(0)
def map_timestamps_to_new_sr(vad_sr, new_sr, timestamps, just_begging_end=False): def map_timestamps_to_new_sr(vad_sr, new_sr, timestamps, just_begging_end=False):
factor = new_sr / vad_sr factor = new_sr / vad_sr
new_timestamps = [] new_timestamps = []
if just_begging_end and timestamps: if just_begging_end and timestamps:
# get just the start and end timestamps # get just the start and end timestamps
new_dict = {'start': int(timestamps[0]['start']*factor), 'end': int(timestamps[-1]['end']*factor)} new_dict = {"start": int(timestamps[0]["start"] * factor), "end": int(timestamps[-1]["end"] * factor)}
new_timestamps.append(new_dict) new_timestamps.append(new_dict)
else: else:
for ts in timestamps: for ts in timestamps:
# map to the new SR # map to the new SR
new_dict = {'start': int(ts['start']*factor), 'end': int(ts['end']*factor)} new_dict = {"start": int(ts["start"] * factor), "end": int(ts["end"] * factor)}
new_timestamps.append(new_dict) new_timestamps.append(new_dict)
return new_timestamps return new_timestamps
def get_vad_model_and_utils(use_cuda=False): def get_vad_model_and_utils(use_cuda=False):
model, utils = torch.hub.load(repo_or_dir='snakers4/silero-vad', model, utils = torch.hub.load(repo_or_dir="snakers4/silero-vad", model="silero_vad", force_reload=True, onnx=False)
model='silero_vad',
force_reload=True,
onnx=False)
if use_cuda: if use_cuda:
model = model.cuda() model = model.cuda()
get_speech_timestamps, save_audio, _, _, collect_chunks = utils get_speech_timestamps, save_audio, _, _, collect_chunks = utils
return model, get_speech_timestamps, save_audio, collect_chunks return model, get_speech_timestamps, save_audio, collect_chunks
def remove_silence(model_and_utils, audio_path, out_path, vad_sample_rate=8000, trim_just_beginning_and_end=True, use_cuda=False):
def remove_silence(
model_and_utils, audio_path, out_path, vad_sample_rate=8000, trim_just_beginning_and_end=True, use_cuda=False
):
# get the VAD model and utils functions # get the VAD model and utils functions
model, get_speech_timestamps, save_audio, collect_chunks = model_and_utils model, get_speech_timestamps, save_audio, collect_chunks = model_and_utils
@ -62,7 +66,9 @@ def remove_silence(model_and_utils, audio_path, out_path, vad_sample_rate=8000,
speech_timestamps = get_speech_timestamps(wav_vad, model, sampling_rate=vad_sample_rate, window_size_samples=768) speech_timestamps = get_speech_timestamps(wav_vad, model, sampling_rate=vad_sample_rate, window_size_samples=768)
# map the current speech_timestamps to the sample rate of the ground truth audio # map the current speech_timestamps to the sample rate of the ground truth audio
new_speech_timestamps = map_timestamps_to_new_sr(vad_sample_rate, gt_sample_rate, speech_timestamps, trim_just_beginning_and_end) new_speech_timestamps = map_timestamps_to_new_sr(
vad_sample_rate, gt_sample_rate, speech_timestamps, trim_just_beginning_and_end
)
# if have speech timestamps else save the wav # if have speech timestamps else save the wav
if new_speech_timestamps: if new_speech_timestamps: