Skip to content

Commit

Permalink
Update tts_conversion.py
Browse files Browse the repository at this point in the history
  • Loading branch information
@Bebra777228
Bebra777228 authored Aug 5, 2024
1 parent 08017fb commit 41acdfe
Showing 1 changed file with 31 additions and 50 deletions.
81 changes: 31 additions & 50 deletions src/scripts/conversion/tts_conversion.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -7,87 +7,68 @@
import numpy as np
import gradio as gr
import edge_tts
import logging
import asyncio

logging.basicConfig(level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')

now_dir = os.getcwd()

from src.rvc import Config, load_hubert, get_vc, rvc_infer

RVC_MODELS_DIR = os.path.join(now_dir, 'models', 'rvc_models')
HUBERT_MODEL_PATH = os.path.join(now_dir, 'models', 'assets', 'hubert_base.pt')
OUTPUT_DIR = os.path.join(now_dir, 'song_output')
OUTPUT_DIR = os.path.join(now_dir, 'output')

def get_rvc_model(voice_model):
model_dir = os.path.join(RVC_MODELS_DIR, voice_model)
rvc_model_path = next((os.path.join(model_dir, f) for f in os.listdir(model_dir) if f.endswith('.pth')), None)
rvc_index_path = next((os.path.join(model_dir, f) for f in os.listdir(model_dir) if f.endswith('.index')), None)
if not rvc_model_path:
logging.error(f'В каталоге {model_dir} отсутствует файл модели.')
print(f'В каталоге {model_dir} отсутствует файл модели.')
raise FileNotFoundError(f'В каталоге {model_dir} отсутствует файл модели.')
return rvc_model_path, rvc_index_path

def convert_to_stereo(audio_path):
try:
wave, sr = librosa.load(audio_path, mono=False, sr=44100)
if wave.ndim == 1:
stereo_path = os.path.join(OUTPUT_DIR, 'Voice_stereo.wav')
subprocess.run(shlex.split(f'ffmpeg -y -loglevel error -i "{audio_path}" -ac 2 -f wav "{stereo_path}"'))
return stereo_path
return audio_path
except Exception as e:
logging.error(f"Ошибка при конвертации в стерео: {e}")
raise
wave, sr = librosa.load(audio_path, mono=False, sr=44100)
if wave.ndim == 1:
stereo_path = os.path.join(OUTPUT_DIR, 'Voice_stereo.wav')
subprocess.run(shlex.split(f'ffmpeg -y -loglevel error -i "{audio_path}" -ac 2 -f wav "{stereo_path}"'))
return stereo_path
return audio_path

def display_progress(percent, message, progress=gr.Progress()):
progress(percent, desc=message)

def voice_change(voice_model, vocals_path, output_path, pitch_change, f0_method, index_rate, filter_radius, volume_envelope, protect, hop_length, f0autotune, f0_min, f0_max):
try:
rvc_model_path, rvc_index_path = get_rvc_model(voice_model)
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
config = Config(device, True)
hubert_model = load_hubert(device, config.is_half, HUBERT_MODEL_PATH)
cpt, version, net_g, tgt_sr, vc = get_vc(device, config.is_half, config, rvc_model_path)
rvc_model_path, rvc_index_path = get_rvc_model(voice_model)
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
config = Config(device, True)
hubert_model = load_hubert(device, config.is_half, HUBERT_MODEL_PATH)
cpt, version, net_g, tgt_sr, vc = get_vc(device, config.is_half, config, rvc_model_path)

rvc_infer(rvc_index_path, index_rate, vocals_path, output_path, pitch_change, f0_method, cpt, version, net_g,
filter_radius, tgt_sr, volume_envelope, protect, hop_length, vc, hubert_model, f0autotune, f0_min, f0_max)
rvc_infer(rvc_index_path, index_rate, vocals_path, output_path, pitch_change, f0_method, cpt, version, net_g,
filter_radius, tgt_sr, volume_envelope, protect, hop_length, vc, hubert_model, f0autotune, f0_min, f0_max)

del hubert_model, cpt, net_g, vc
gc.collect()
torch.cuda.empty_cache()
except Exception as e:
logging.error(f"Ошибка при преобразовании голоса: {e}")
raise
del hubert_model, cpt, net_g, vc
gc.collect()
torch.cuda.empty_cache()

async def text_to_speech(text, voice, output_path):
try:
communicate = edge_tts.Communicate(text=text, voice=voice)
await communicate.save(output_path)
except Exception as e:
logging.error(f"Ошибка при преобразовании текста в речь: {e}")
raise
communicate = edge_tts.Communicate(text=text, voice=voice)
await communicate.save(output_path)

def tts_conversion(text, voice_model, voice, pitch_change, index_rate=0.5, filter_radius=3, volume_envelope=0.25, f0_method='rmvpe',
hop_length=128, protect=0.33, output_format='mp3', progress=gr.Progress(), f0autotune=False, f0_min=50, f0_max=1100):
try:
if not text or not voice_model or not voice:
logging.error('Убедитесь, что все поля заполнены.')
raise ValueError('Убедитесь, что все поля заполнены.')
if not text or not voice_model or not voice:
print('Убедитесь, что все поля заполнены.')
raise ValueError('Убедитесь, что все поля заполнены.')

display_progress(0, '[~] Запуск конвейера генерации TTS...', progress)
display_progress(0, '[~] Запуск конвейера генерации TTS...', progress)

tts_voice_path = os.path.join(OUTPUT_DIR, f'TTS_Output.wav')
asyncio.run(text_to_speech(text, voice, tts_voice_path))
tts_voice_path = os.path.join(OUTPUT_DIR, f'TTS_Output.wav')
asyncio.run(text_to_speech(text, voice, tts_voice_path))

display_progress(0.5, '[~] Преобразование голоса...', progress)
final_output_path = os.path.join(OUTPUT_DIR, f'Converted_TTS_Voice.{output_format}')
voice_change(voice_model, tts_voice_path, final_output_path, pitch_change, f0_method, index_rate,
filter_radius, volume_envelope, protect, hop_length, f0autotune, f0_min, f0_max)
display_progress(0.5, '[~] Преобразование голоса...', progress)
final_output_path = os.path.join(OUTPUT_DIR, f'Converted_TTS_Voice.{output_format}')
voice_change(voice_model, tts_voice_path, final_output_path, pitch_change, f0_method, index_rate,
filter_radius, volume_envelope, protect, hop_length, f0autotune, f0_min, f0_max)

return final_output_path, tts_voice_path
except Exception as e:
logging.error(f"Ошибка в процессе конвертации: {e}")
raise
return final_output_path, tts_voice_path

0 comments on commit 41acdfe

Please sign in to comment.