2023-04-01 15:02:53 +08:00
|
|
|
import ffmpeg
|
|
|
|
|
import numpy as np
|
2023-07-23 03:47:53 +00:00
|
|
|
|
|
|
|
|
# import praatio
|
|
|
|
|
# import praatio.praat_scripts
|
2023-07-21 20:36:54 +07:00
|
|
|
import os
|
2023-07-26 06:24:23 +07:00
|
|
|
import random
|
2023-04-15 20:44:24 +09:00
|
|
|
|
2023-07-25 05:53:47 +07:00
|
|
|
import sqlite3
|
|
|
|
|
|
|
|
|
|
|
2023-07-23 03:47:53 +00:00
|
|
|
|
|
|
|
|
# praatEXE = join('.',os.path.abspath(os.getcwd()) + r"\Praat.exe")
|
|
|
|
|
|
2023-04-15 20:44:24 +09:00
|
|
|
|
2023-07-21 20:36:54 +07:00
|
|
|
def load_audio(file, sr, DoFormant, Quefrency, Timbre):
|
2023-07-26 23:01:10 +07:00
|
|
|
converted = False
|
2023-03-31 17:54:38 +08:00
|
|
|
try:
|
2023-07-25 05:53:47 +07:00
|
|
|
conn = sqlite3.connect('TEMP/db:cachedb?mode=memory&cache=shared', check_same_thread=False)
|
|
|
|
|
cursor = conn.cursor()
|
2023-03-31 17:54:38 +08:00
|
|
|
# https://github.com/openai/whisper/blob/main/whisper/audio.py#L26
|
|
|
|
|
# This launches a subprocess to decode audio while down-mixing and resampling as necessary.
|
|
|
|
|
# Requires the ffmpeg CLI and `ffmpeg-python` package to be installed.
|
2023-04-15 20:44:24 +09:00
|
|
|
file = (
|
|
|
|
|
file.strip(" ").strip('"').strip("\n").strip('"').strip(" ")
|
|
|
|
|
) # 防止小白拷路径头尾带了空格和"和回车
|
2023-07-23 03:47:53 +00:00
|
|
|
file_formanted = file.strip(" ").strip('"').strip("\n").strip('"').strip(" ")
|
2023-07-25 05:53:47 +07:00
|
|
|
cursor.execute("SELECT Quefrency, Timbre, DoFormant FROM formant_data")
|
|
|
|
|
Quefrency, Timbre, DoFormant = cursor.fetchone()
|
2023-07-26 06:24:23 +07:00
|
|
|
#print(f"dofor={bool(DoFormant)} timbr={Timbre} quef={Quefrency}\n")
|
2023-07-25 05:53:47 +07:00
|
|
|
if bool(DoFormant):
|
2023-07-26 06:24:23 +07:00
|
|
|
numerator = round(random.uniform(1,4), 4)
|
2023-07-23 03:47:53 +00:00
|
|
|
# os.system(f"stftpitchshift -i {file} -q {Quefrency} -t {Timbre} -o {file_formanted}")
|
|
|
|
|
# print('stftpitchshift -i "%s" -p 1.0 --rms -w 128 -v 8 -q %s -t %s -o "%s"' % (file, Quefrency, Timbre, file_formanted))
|
2023-07-25 00:23:56 +07:00
|
|
|
|
|
|
|
|
if not file.endswith(".wav"):
|
2023-07-26 06:24:23 +07:00
|
|
|
|
|
|
|
|
if not os.path.isfile(f"{file_formanted}.wav"):
|
2023-07-26 23:01:10 +07:00
|
|
|
converted = True
|
2023-07-26 06:24:23 +07:00
|
|
|
#print(f"\nfile = {file}\n")
|
|
|
|
|
#print(f"\nfile_formanted = {file_formanted}\n")
|
|
|
|
|
converting = (
|
|
|
|
|
ffmpeg.input(file_formanted, threads = 0)
|
|
|
|
|
.output(f"{file_formanted}.wav")
|
|
|
|
|
.run(
|
|
|
|
|
cmd=["ffmpeg", "-nostdin"], capture_stdout=True, capture_stderr=True
|
|
|
|
|
)
|
2023-07-25 00:23:56 +07:00
|
|
|
)
|
2023-07-26 06:24:23 +07:00
|
|
|
else:
|
|
|
|
|
pass
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
file_formanted = f"{file_formanted}.wav" if not file_formanted.endswith(".wav") else file_formanted
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
print(f" · Formanting {file_formanted}...\n")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-07-23 03:47:53 +00:00
|
|
|
os.system(
|
2023-07-26 06:24:23 +07:00
|
|
|
'stftpitchshift.exe -i "%s" -q "%s" -t "%s" -o "%sFORMANTED_%s.wav"'
|
|
|
|
|
% (file_formanted, Quefrency, Timbre, file_formanted, str(numerator))
|
2023-07-23 03:47:53 +00:00
|
|
|
)
|
2023-07-26 06:24:23 +07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
print(f" · Formanted {file_formanted}!\n")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-07-23 03:47:53 +00:00
|
|
|
# filepraat = (os.path.abspath(os.getcwd()) + '\\' + file).replace('/','\\')
|
|
|
|
|
# file_formantedpraat = ('"' + os.path.abspath(os.getcwd()) + '/' + 'formanted'.join(file_formanted) + '"').replace('/','\\')
|
2023-07-26 06:24:23 +07:00
|
|
|
#print("%sFORMANTED_%s.wav" % (file_formanted, str(numerator)))
|
2023-07-24 23:59:28 +07:00
|
|
|
|
2023-07-21 20:36:54 +07:00
|
|
|
out, _ = (
|
2023-07-26 06:24:23 +07:00
|
|
|
ffmpeg.input("%sFORMANTED_%s.wav" % (file_formanted, str(numerator)), threads=0)
|
2023-07-21 20:36:54 +07:00
|
|
|
.output("-", format="f32le", acodec="pcm_f32le", ac=1, ar=sr)
|
2023-07-23 03:47:53 +00:00
|
|
|
.run(
|
|
|
|
|
cmd=["ffmpeg", "-nostdin"], capture_stdout=True, capture_stderr=True
|
|
|
|
|
)
|
2023-07-21 20:36:54 +07:00
|
|
|
)
|
2023-07-23 03:47:53 +00:00
|
|
|
|
2023-07-26 06:24:23 +07:00
|
|
|
|
2023-07-25 05:53:47 +07:00
|
|
|
|
2023-07-21 20:36:54 +07:00
|
|
|
else:
|
|
|
|
|
out, _ = (
|
|
|
|
|
ffmpeg.input(file, threads=0)
|
|
|
|
|
.output("-", format="f32le", acodec="pcm_f32le", ac=1, ar=sr)
|
2023-07-23 03:47:53 +00:00
|
|
|
.run(
|
|
|
|
|
cmd=["ffmpeg", "-nostdin"], capture_stdout=True, capture_stderr=True
|
|
|
|
|
)
|
2023-07-21 20:36:54 +07:00
|
|
|
)
|
2023-04-01 16:42:19 +08:00
|
|
|
except Exception as e:
|
|
|
|
|
raise RuntimeError(f"Failed to load audio: {e}")
|
2023-07-25 05:53:47 +07:00
|
|
|
|
2023-07-26 06:24:23 +07:00
|
|
|
try: os.remove("%sFORMANTED_%s.wav" % (file_formanted, str(numerator)))
|
2023-07-26 23:01:10 +07:00
|
|
|
except Exception: pass; print("couldn't remove formanted type of file")
|
|
|
|
|
|
|
|
|
|
if converted:
|
|
|
|
|
try: os.remove(file_formanted)
|
|
|
|
|
except Exception: pass; print("couldn't remove converted type of file")
|
|
|
|
|
converted = False
|
2023-07-26 06:24:23 +07:00
|
|
|
|
2023-07-25 05:53:47 +07:00
|
|
|
conn.close()
|
2023-04-24 20:35:56 +08:00
|
|
|
return np.frombuffer(out, np.float32).flatten()
|