catiR commited on
Commit
8827531
·
1 Parent(s): 7066115

run clustering

Browse files
Files changed (1) hide show
  1. scripts/clusterprosody.py +8 -8
scripts/clusterprosody.py CHANGED
@@ -423,8 +423,8 @@ def get_audio_part(start_time, end_time, id, path):
423
 
424
  def plot_pitch_tts(speech_data,tts_data, tts_align,words,seg_aligns,cluster_id, voice):
425
  colors = ["red", "green", "blue", "orange", "purple", "pink", "brown", "gray", "cyan"]
426
- i = 0
427
- fig = plt.figure(figsize=(6, 5))
428
  plt.title(f"{words} - Pitch - Cluster {cluster_id}")
429
  for k,v in speech_data.items():
430
 
@@ -447,12 +447,12 @@ def plot_pitch_tts(speech_data,tts_data, tts_align,words,seg_aligns,cluster_id,
447
  if len(word_times)>2:
448
  for i in range(1,len(word_times)-1):
449
  bound_line = np.mean([word_times[i][2],word_times[i+1][1]])
450
- plt.axvline(x=bound_line, color=colors[i], linestyle='--', linewidth=1, label=f"Speaker {spk} -> {word_times[i+1][0]}")
451
 
452
- plt.scatter(pitch_xvals, pitches, color=colors[i], label=f"Speaker {spk}")
453
- i += 1
454
- if i >= len(colors):
455
- i=0
456
 
457
  tpitches = [p for p,e in tts_data]
458
  t_xvals = [x*0.005 for x in range(len(tpitches))]
@@ -469,7 +469,7 @@ def plot_pitch_tts(speech_data,tts_data, tts_align,words,seg_aligns,cluster_id,
469
  plt.scatter(t_xvals, tpitches, color="black", label=f"TTS {voice}")
470
 
471
 
472
- plt.legend()
473
  #plt.show()
474
 
475
 
 
423
 
424
  def plot_pitch_tts(speech_data,tts_data, tts_align,words,seg_aligns,cluster_id, voice):
425
  colors = ["red", "green", "blue", "orange", "purple", "pink", "brown", "gray", "cyan"]
426
+ cc = 0
427
+ fig = plt.figure(figsize=(10, 5))
428
  plt.title(f"{words} - Pitch - Cluster {cluster_id}")
429
  for k,v in speech_data.items():
430
 
 
447
  if len(word_times)>2:
448
  for i in range(1,len(word_times)-1):
449
  bound_line = np.mean([word_times[i][2],word_times[i+1][1]])
450
+ plt.axvline(x=bound_line, color=colors[cc], linestyle='--', linewidth=1, label=f"Speaker {spk} -> {word_times[i+1][0]}")
451
 
452
+ plt.scatter(pitch_xvals, pitches, color=colors[cc], label=f"Speaker {spk}")
453
+ cc += 1
454
+ if cc >= len(colors):
455
+ cc=0
456
 
457
  tpitches = [p for p,e in tts_data]
458
  t_xvals = [x*0.005 for x in range(len(tpitches))]
 
469
  plt.scatter(t_xvals, tpitches, color="black", label=f"TTS {voice}")
470
 
471
 
472
+ #plt.legend()
473
  #plt.show()
474
 
475