Kangarroar commited on
Commit
95add38
1 Parent(s): 1ea2ba0

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +23 -3
app.py CHANGED
@@ -25,9 +25,10 @@ from infer import *
25
  import logging
26
  from infer_tools.infer_tool import *
27
  import io
 
28
 
29
  clip_completed = False
30
- def render_audio(ckpt_temp_file, config_temp_file, audio_temp_file, title, title2, choice, noise_step, use_mel_as_base):
31
  logging.getLogger('numba').setLevel(logging.WARNING)
32
  title = int(title)
33
  title2 = int(title2)
@@ -43,12 +44,31 @@ def render_audio(ckpt_temp_file, config_temp_file, audio_temp_file, title, title
43
  key = title # 音高调整,支持正负(半音)
44
  # 加速倍数
45
  pndm_speedup = 20
46
- wav_gen='queeeeee.wav'
47
 
48
  # Show the spinner and run the run_clip function inside the 'with' block
49
  with st.spinner("Rendering Audio..."):
50
  f0_tst, f0_pred, audio = run_clip(svc_model,file_path=wav_fn, key=key, acc=title2, use_crepe=choice, use_pe=False, thre=0.05,
51
- use_gt_mel=use_mel_as_base, add_noise_step=noise_step,project_name=project_name,out_path=wav_gen)
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
52
  clip_completed = True
53
  if clip_completed:
54
 
 
25
  import logging
26
  from infer_tools.infer_tool import *
27
  import io
28
+ import parselmouth
29
 
30
  clip_completed = False
31
+ def render_audio(ckpt_temp_file, config_temp_file, audio_temp_file, title, title2, title3, choice, noise_step, use_mel_as_base):
32
  logging.getLogger('numba').setLevel(logging.WARNING)
33
  title = int(title)
34
  title2 = int(title2)
 
44
  key = title # 音高调整,支持正负(半音)
45
  # 加速倍数
46
  pndm_speedup = 20
47
+ wav_gen='que.wav'
48
 
49
  # Show the spinner and run the run_clip function inside the 'with' block
50
  with st.spinner("Rendering Audio..."):
51
  f0_tst, f0_pred, audio = run_clip(svc_model,file_path=wav_fn, key=key, acc=title2, use_crepe=choice, use_pe=False, thre=0.05,
52
+     ##PRAAT
53
+     formant_shift_ratio_str = title3
54
+     formant_shift_ratio = float(formant_shift_ratio_str)
55
+     
56
+     # If the formant shift ratio is not equal to 1.0, change the gender of the sound using parselmouth
57
+     if formant_shift_ratio != 1.0:
58
+         sound = parselmouth.Sound(wav_gen)
59
+         print(wav_gen)
60
+         Audio(data=sound.values, rate=sound.sampling_frequency)
61
+         sound.get_power()
62
+         sampling_rate = sound.sampling_frequency
63
+         print(sampling_rate)
64
+         resampled_sound = sound.resample(sampling_rate)
65
+         print(resampled_sound)
66
+         factor = formant_shift_ratio
67
+         print(factor)
68
+         manipulated_sound = call(sound, "Change gender", 75, 500, factor, 0, 1, 1)
69
+         print(manipulated_sound)
70
+         manipulated_sound.save("que.wav", "WAV")
71
+         print("Gender correct!") use_gt_mel=use_mel_as_base, add_noise_step=noise_step,project_name=project_name,out_path=wav_gen)
72
  clip_completed = True
73
  if clip_completed:
74