alibabasglab commited on
Commit
341eb54
·
verified ·
1 Parent(s): 9f61737

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +11 -10
app.py CHANGED
@@ -6,21 +6,22 @@ from clearvoice import ClearVoice
6
  import os
7
 
8
  @spaces.GPU
9
- def fn_clearvoice_sr(input_wav, sr):
10
- if sr == "16000 Hz":
11
- myClearVoice = ClearVoice(task='speech_enhancement', model_names=['FRCRN_SE_16K'])
12
- fs = 16000
13
- else:
14
- myClearVoice = ClearVoice(task='speech_enhancement', model_names=['MossFormer2_SE_48K'])
15
- fs = 48000
 
16
  output_wav_dict = myClearVoice(input_path=input_wav, online_write=False)
17
  if isinstance(output_wav_dict, dict):
18
  key = next(iter(output_wav_dict))
19
  output_wav = output_wav_dict[key]
20
  else:
21
  output_wav = output_wav_dict
22
- sf.write('enhanced.wav', output_wav, fs)
23
- return 'enhanced.wav'
24
 
25
  @spaces.GPU
26
  def fn_clearvoice_se(input_wav, sr):
@@ -86,7 +87,7 @@ sr_demo = gr.Interface(
86
  fn=fn_clearvoice_sr,
87
  inputs = [
88
  gr.Audio(label="Input Audio", type="filepath"),
89
- gr.Checkbox(["Apply Enhancement"], label="Apply Speech Enhancement"),
90
  ],
91
  outputs = [
92
  gr.Audio(label="Output Audio", type="filepath"),
 
6
  import os
7
 
8
  @spaces.GPU
9
+ def fn_clearvoice_sr(input_wav, apply_se):
10
+
11
+ myClearVoice = ClearVoice(task='speech_super_resolution', model_names=['MossFormer2_SR_48K'])
12
+ fs = 48000
13
+ if apply_se:
14
+ myClearVoice_se = ClearVoice(task='speech_enhancement', model_names=['MossFormer2_SE_48K'])
15
+ output_wav_dict = myClearVoice_se(input_path=input_wav, online_write=True, output_path='enhanced.wav')
16
+ input_wav = 'enhanced.wav'
17
  output_wav_dict = myClearVoice(input_path=input_wav, online_write=False)
18
  if isinstance(output_wav_dict, dict):
19
  key = next(iter(output_wav_dict))
20
  output_wav = output_wav_dict[key]
21
  else:
22
  output_wav = output_wav_dict
23
+ sf.write('enhanced_high_res.wav', output_wav, fs)
24
+ return 'enhanced_high_res.wav'
25
 
26
  @spaces.GPU
27
  def fn_clearvoice_se(input_wav, sr):
 
87
  fn=fn_clearvoice_sr,
88
  inputs = [
89
  gr.Audio(label="Input Audio", type="filepath"),
90
+ gr.Checkbox(["Apply Speech Enhancement"], label="Apply Speech Enhancement"),
91
  ],
92
  outputs = [
93
  gr.Audio(label="Output Audio", type="filepath"),