tee342 commited on
Commit
ff7c58e
Β·
verified Β·
1 Parent(s): 86dccd9

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +80 -3
app.py CHANGED
@@ -23,6 +23,8 @@ import base64
23
  import pickle
24
  import json
25
  import soundfile as sf
 
 
26
 
27
  # Suppress warnings
28
  warnings.filterwarnings("ignore")
@@ -549,7 +551,7 @@ def clone_voice(source_audio, reference_audio):
549
  mixed.export(out_path, format="wav")
550
  return out_path
551
 
552
- # === AI Remastering – Now Fixed! ===
553
  def ai_remaster(audio_path):
554
  try:
555
  audio = AudioSegment.from_file(audio_path)
@@ -574,6 +576,50 @@ def ai_mastering_chain(audio_path, genre="Pop", target_lufs=-14.0):
574
  audio.export(out_path, format="wav")
575
  return out_path
576
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
577
  # === Main UI – With Studio Pulse Branding ===
578
  with gr.Blocks(css="""
579
  body {
@@ -990,7 +1036,7 @@ with gr.Blocks(css="""
990
  ],
991
  outputs=gr.Audio(label="Stretched Output", type="filepath"),
992
  title="Match Track to BPM",
993
- description="Stretch audio to match a specific tempo"
994
  )
995
 
996
  # --- A/B Compare ===
@@ -1067,7 +1113,7 @@ with gr.Blocks(css="""
1067
  ],
1068
  outputs=gr.Audio(label="Natural-Sounding Vocal", type="filepath"),
1069
  title="Preserve Vocal Quality During Pitch Shift",
1070
- description="Make pitch-shifted vocals sound more human"
1071
  )
1072
 
1073
  # --- Voice Swap / Cloning – New Tab ===
@@ -1104,4 +1150,35 @@ with gr.Blocks(css="""
1104
  description="Perfect for sharing with producers or archiving"
1105
  )
1106
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1107
  demo.launch()
 
23
  import pickle
24
  import json
25
  import soundfile as sf
26
+ from moviepy.editor import VideoFileClip, AudioFileClip
27
+ from fpdf import FPDF
28
 
29
  # Suppress warnings
30
  warnings.filterwarnings("ignore")
 
551
  mixed.export(out_path, format="wav")
552
  return out_path
553
 
554
+ # === AI Remastering – Now Fixed & Working ===
555
  def ai_remaster(audio_path):
556
  try:
557
  audio = AudioSegment.from_file(audio_path)
 
576
  audio.export(out_path, format="wav")
577
  return out_path
578
 
579
+ # === Generate PDF Report ===
580
+ class PDFReport(FPDF):
581
+ def header(self):
582
+ self.set_font('Arial', 'B', 12)
583
+ self.cell(0, 10, 'Studio Pulse AI - Session Report', ln=True, align='C')
584
+
585
+ def export_session_report(session_data):
586
+ pdf = PDFReport()
587
+ pdf.add_page()
588
+ pdf.set_font("Arial", size=10)
589
+ pdf.multi_cell(0, 10, txt=session_data)
590
+ out_path = os.path.join(tempfile.gettempdir(), "session_report.pdf")
591
+ pdf.output(out_path)
592
+ return out_path
593
+
594
+ # === Video Sync Mode – Extract and re-sync audio ===
595
+ def extract_audio_from_video(video_path):
596
+ clip = VideoFileClip(video_path)
597
+ out_path = os.path.join(tempfile.gettempdir(), "video_audio.wav")
598
+ clip.audio.write_audiofile(out_path)
599
+ return out_path
600
+
601
+ def re_sync_audio(video_path, edited_audio_path):
602
+ original_clip = VideoFileClip(video_path)
603
+ edited_audio = AudioSegment.from_file(edited_audio_path)
604
+ temp_audio_path = os.path.join(tempfile.gettempdir(), "temp_edited.mp3")
605
+ edited_audio.export(temp_audio_path, format="mp3")
606
+ final_clip = original_clip.set_audio(AudioFileClip(temp_audio_path))
607
+ out_path = os.path.join(tempfile.gettempdir(), "final_video.mp4")
608
+ final_clip.write_videofile(out_path, codec="libx264", audio_codec="aac")
609
+ return out_path
610
+
611
+ # === Sidechain Compression / Ducking ===
612
+ def sidechain_ducking(background, voice, threshold=-16, ratio=4):
613
+ bg_samples, bg_sr = audiosegment_to_array(background)
614
+ voice_samples, v_sr = audiosegment_to_array(voice)
615
+ if bg_sr != v_sr:
616
+ voice_samples = librosa.resample(voice_samples, orig_sr=v_sr, target_sr=bg_sr)
617
+ bg_rms = np.sqrt(np.mean(bg_samples ** 2))
618
+ voice_rms = np.sqrt(np.mean(voice_samples ** 2))
619
+ if voice_rms > bg_rms * threshold / 100:
620
+ bg_samples = bg_samples * ratio
621
+ return array_to_audiosegment(bg_samples.astype(np.int16), bg_sr, channels=background.channels)
622
+
623
  # === Main UI – With Studio Pulse Branding ===
624
  with gr.Blocks(css="""
625
  body {
 
1036
  ],
1037
  outputs=gr.Audio(label="Stretched Output", type="filepath"),
1038
  title="Match Track to BPM",
1039
+ description="Stretch or shrink audio to match a specific tempo"
1040
  )
1041
 
1042
  # --- A/B Compare ===
 
1113
  ],
1114
  outputs=gr.Audio(label="Natural-Sounding Vocal", type="filepath"),
1115
  title="Preserve Vocal Quality During Pitch Shift",
1116
+ description="Make pitch-shifted vocals sound more natural"
1117
  )
1118
 
1119
  # --- Voice Swap / Cloning – New Tab ===
 
1150
  description="Perfect for sharing with producers or archiving"
1151
  )
1152
 
1153
+ # --- Session Report Export (PDF) – New Tab ===
1154
+ with gr.Tab("πŸ“„ Session Report"):
1155
+ gr.Interface(
1156
+ fn=export_session_report,
1157
+ inputs=gr.Textbox(label="Session Data (JSON)", lines=10),
1158
+ outputs=gr.File(label="Download PDF Report"),
1159
+ title="Export Detailed Session Report",
1160
+ description="For professional use or sharing with collaborators"
1161
+ )
1162
+
1163
+ # --- Effect Chains Marketplace – New Tab ===
1164
+ with gr.Tab("🧩 Effect Chain Marketplace"):
1165
+ gr.Interface(
1166
+ fn=publish_effect_chain,
1167
+ inputs=[
1168
+ gr.Textbox(label="Effect Chain (JSON)", lines=2),
1169
+ gr.Textbox(label="Name", value="My Chain"),
1170
+ gr.CheckboxGroup(choices=["music", "vocals", "podcast", "radio", "edm"], label="Tags")
1171
+ ],
1172
+ outputs=gr.JSON(label="Published Chain Info"),
1173
+ title="Publish Your Effect Chain",
1174
+ description="Share your setup with others"
1175
+ )
1176
+
1177
+ with gr.Tab("🧩 Search Chains"):
1178
+ gr.Interface(
1179
+ fn=search_effect_chains,
1180
+ inputs=gr.Textbox(label="Search Query", value="vocal"),
1181
+ outputs=gr.JSON(label="Matching Chains")
1182
+ )
1183
+
1184
  demo.launch()