tee342 commited on
Commit
84f511d
ยท
verified ยท
1 Parent(s): 9d21f18

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +45 -46
app.py CHANGED
@@ -26,13 +26,21 @@ import json
26
  # Suppress warnings
27
  warnings.filterwarnings("ignore")
28
 
 
29
  try:
30
- import pyrubberband as pyrb
31
  except ImportError:
32
- print("Installing pyrubberband...")
33
  import subprocess
34
- subprocess.run(["pip", "install", "pyrubberband"])
35
- import pyrubberband as pyrb
 
 
 
 
 
 
 
36
 
37
  # === Helper Functions ===
38
  def audiosegment_to_array(audio):
@@ -441,6 +449,26 @@ def visualize_spectrum(audio_path):
441
  buf.seek(0)
442
  return Image.open(buf)
443
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
444
  # === Main UI โ€“ With Studio Pulse Branding ===
445
  with gr.Blocks(css="""
446
  body {
@@ -473,7 +501,6 @@ with gr.Blocks(css="""
473
  border-radius: 12px;
474
  padding: 15px;
475
  box-shadow: 0 0 10px #1f7bbd44;
476
- border: none;
477
  transition: all 0.3s ease;
478
  }
479
 
@@ -497,21 +524,6 @@ with gr.Blocks(css="""
497
  box-shadow: 0 0 15px #3b82f6aa;
498
  }
499
 
500
- .gr-tabs button {
501
- font-size: 16px;
502
- padding: 10px 20px;
503
- border-radius: 8px;
504
- background: #1e293b;
505
- color: white;
506
- transition: all 0.3s ease;
507
- }
508
-
509
- .gr-tabs button:hover {
510
- background: #3b82f6;
511
- color: black;
512
- box-shadow: 0 0 10px #3b82f6aa;
513
- }
514
-
515
  input[type="text"], input[type="number"], select, textarea {
516
  background-color: #334155 !important;
517
  color: white !important;
@@ -643,6 +655,16 @@ with gr.Blocks(css="""
643
  clear_btn=None
644
  )
645
 
 
 
 
 
 
 
 
 
 
 
646
  # --- Harmonic Saturation / Exciter โ€“ Now Included ===
647
  with gr.Tab("๐Ÿงฌ Harmonic Saturation"):
648
  gr.Interface(
@@ -699,8 +721,7 @@ with gr.Blocks(css="""
699
  ],
700
  outputs=gr.Audio(label="Pitch-Corrected Output", type="filepath"),
701
  title="AI Auto-Tune",
702
- description="Correct vocal pitch automatically using AI",
703
- allow_flagging="never"
704
  )
705
 
706
  # --- Frequency Spectrum Tab โ€“ Real-time Visualizer ===
@@ -753,7 +774,8 @@ with gr.Blocks(css="""
753
  ],
754
  outputs=gr.File(label="Project File (.aiproj)"),
755
  title="Save Everything Together",
756
- description="Save your session, effects, and settings in one file to reuse later."
 
757
  )
758
 
759
  gr.Interface(
@@ -767,29 +789,6 @@ with gr.Blocks(css="""
767
  description="Load your saved session"
768
  )
769
 
770
- # --- Preset Cards Gallery โ€“ Visual Selection ===
771
- with gr.Tab("๐ŸŽ› Preset Gallery"):
772
- gr.Markdown("### Select a preset visually")
773
- preset_gallery = gr.Gallery(value=[
774
- ("https://picsum.photos/id/10/150/100", "Pop"),
775
- ("https://picsum.photos/id/20/150/100", "EDM"),
776
- ("https://picsum.photos/id/30/150/100", "Rock"),
777
- ("https://picsum.photos/id/40/150/100", "Hip-Hop"),
778
- ("https://picsum.photos/id/50/150/100", "R&B"),
779
- ("https://picsum.photos/id/60/150/100", "Soul"),
780
- ("https://picsum.photos/id/70/150/100", "Funk")
781
- ], label="Preset Cards", columns=4, height="auto")
782
-
783
- preset_name_out = gr.Dropdown(choices=preset_names, label="Selected Preset")
784
- preset_effects_out = gr.CheckboxGroup(choices=list(preset_choices["Default"]), label="Effects")
785
-
786
- def load_preset_by_card(evt: gr.SelectData):
787
- index = evt.index % len(preset_names)
788
- name = preset_names[index]
789
- return name, preset_choices[name]
790
-
791
- preset_gallery.select(fn=load_preset_by_card, inputs=[], outputs=[preset_name_out, preset_effects_out])
792
-
793
  # --- Prompt-Based Editing Tab โ€“ Added Back ===
794
  def process_prompt(audio, prompt):
795
  return apply_noise_reduction(audio)
 
26
  # Suppress warnings
27
  warnings.filterwarnings("ignore")
28
 
29
+ # Try importing speechbrain for remastering
30
  try:
31
+ from speechbrain.inference.denoisers import SEEnhance
32
  except ImportError:
33
+ print("Installing SpeechBrain...")
34
  import subprocess
35
+ subprocess.run(["pip", "install", "speechbrain"])
36
+ from speechbrain.inference.denoisers import SEEnhance
37
+
38
+ # Load denoiser model
39
+ try:
40
+ enhancer = SEEnhance.from_hparams(source="speechbrain/reverb-wpe-wsj0-wsj0mix", savedir="remaster_model")
41
+ except Exception as e:
42
+ enhancer = None
43
+ print("SpeechBrain model not available:", str(e))
44
 
45
  # === Helper Functions ===
46
  def audiosegment_to_array(audio):
 
449
  buf.seek(0)
450
  return Image.open(buf)
451
 
452
+ # === AI Remastering Tab โ€“ New Feature ===
453
+ def ai_remaster(audio_path):
454
+ try:
455
+ # Step 1: Noise reduction
456
+ audio = AudioSegment.from_file(audio_path)
457
+ samples, sr = audiosegment_to_array(audio)
458
+ reduced = nr.reduce_noise(y=samples, sr=sr)
459
+ cleaned = array_to_audiosegment(reduced, sr, channels=audio.channels)
460
+
461
+ # Step 2: Vocal isolation
462
+ isolated_path = apply_vocal_isolation(cleaned.export(os.path.join(tempfile.gettempdir(), "cleaned.wav"), format="wav")
463
+
464
+ # Step 3: Apply mastering chain
465
+ mastered_path = ai_mastering_chain(isolated_path, genre="Pop", target_lufs=-14.0)
466
+
467
+ return mastered_path
468
+
469
+ except Exception as e:
470
+ return None
471
+
472
  # === Main UI โ€“ With Studio Pulse Branding ===
473
  with gr.Blocks(css="""
474
  body {
 
501
  border-radius: 12px;
502
  padding: 15px;
503
  box-shadow: 0 0 10px #1f7bbd44;
 
504
  transition: all 0.3s ease;
505
  }
506
 
 
524
  box-shadow: 0 0 15px #3b82f6aa;
525
  }
526
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
527
  input[type="text"], input[type="number"], select, textarea {
528
  background-color: #334155 !important;
529
  color: white !important;
 
655
  clear_btn=None
656
  )
657
 
658
+ # --- AI Remastering Tab โ€“ New Feature ===
659
+ with gr.Tab("๐Ÿ”ฎ AI Remastering"):
660
+ gr.Interface(
661
+ fn=ai_remaster,
662
+ inputs=gr.Audio(label="Upload Low-Quality Recording", type="filepath"),
663
+ outputs=gr.Audio(label="Studio-Grade Output", type="filepath"),
664
+ title="Transform Low-Quality Recordings to Studio Sound",
665
+ description="Uses noise reduction, vocal isolation, and mastering to enhance old, low-quality audio."
666
+ )
667
+
668
  # --- Harmonic Saturation / Exciter โ€“ Now Included ===
669
  with gr.Tab("๐Ÿงฌ Harmonic Saturation"):
670
  gr.Interface(
 
721
  ],
722
  outputs=gr.Audio(label="Pitch-Corrected Output", type="filepath"),
723
  title="AI Auto-Tune",
724
+ description="Correct vocal pitch automatically using AI"
 
725
  )
726
 
727
  # --- Frequency Spectrum Tab โ€“ Real-time Visualizer ===
 
774
  ],
775
  outputs=gr.File(label="Project File (.aiproj)"),
776
  title="Save Everything Together",
777
+ description="Save your session, effects, and settings in one file to reuse later.",
778
+ allow_flagging="never"
779
  )
780
 
781
  gr.Interface(
 
789
  description="Load your saved session"
790
  )
791
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
792
  # --- Prompt-Based Editing Tab โ€“ Added Back ===
793
  def process_prompt(audio, prompt):
794
  return apply_noise_reduction(audio)