ShortsPacer — Viral Voice Tightener & Studio De-Breath

Stop losing 70% of your short-form audience in the first 3 seconds.
Turn loose AI voiceovers, raw phone takes, and talking-head videos into punchy, broadcast-grade, high-retention audio in under 0.4 seconds.

🥊 ShortsPacer vs. Traditional DAWs (Adobe Audition)

"Stop spending 30 minutes manually slicing audio in Adobe Audition for a 60-second TikTok. ShortsPacer does it cleaner and automatically in 0.3 seconds."

⚡ 4 All-in-One Killer Features

1. 🎬 Direct Video Support (.mp4 / .mov / .webm)

No more separating audio tracks in your video editor. Drag and drop talking-head .mp4 or .mov files directly; ShortsPacer extracts the audio stream, surgically trims pauses, and exports ready-to-cut mastered tracks in sub-second speeds without re-encoding video.

2. 🔊 Built-in -14 LUFS Broadcast Mastering (TikTok & YouTube Standard)

Weak, quiet, or inconsistent audio gets buried by social algorithms. ShortsPacer applies ITU-R BS.1770 / EBU R128 loudness normalization (-14 LUFS, -1.5 dB True Peak), giving your voiceover rich acoustic authority without digital distortion.

3. ✂️ The Ultimate Pre-Processor for CapCut & Premiere Auto-Captions

When you run Auto-Captions in CapCut or Premiere on raw recordings, awkward gasps and dragging pauses cause the AI to generate split, messy 1-word subtitle cards. ShortsPacer eliminates dead space beforehand, resulting in:

• Zero Fragmented Captions: Perfectly grouped phrases.
• Flawless Word-Level Timing: Dynamic bouncing text and karaoke effects snap right to the beat.
• 3x Faster Subtitle Editing: No more manual word merging on the timeline.

4. 🎛️ 3 Creator Presets (Zero Guesswork)

• viral (TikTok / Shorts / Reels): Aggressive 20%+ pacing compaction. High-tempo delivery designed for maximum viewer retention.
• podcast (Interviews / Long-Form): Removes heavy gasping breaths while maintaining natural conversational warmth and breathing room.
• natural (Audiobooks / Documentary): Gentle room-tone smoothing for slow narrative pacing.

📊 Real Benchmark: Before vs. After

Verified production run on a raw creator voiceover track:

🎯 Who This Is For (And Who It Isn't)

✅ PERFECT FOR:

1. AI Voiceover Creators: Using ElevenLabs, MiniMax, or OpenAI TTS who need to eliminate awkward robotic pauses, gasping breaths, and boost 3-second retention.
2. Talking-Head Video Creators: Editors who record raw video takes and need a broadcast-tightened master audio track to drop into CapCut or Premiere as their pacing anchor.
3. Podcasters & Interviewers: Creators wanting clean, warm vocal tracks without spending 45 minutes manually slicing waveforms in a DAW.
4. CapCut / Premiere Auto-Caption Users: Creators who want crisp, rhythmically grouped subtitles without fragmented, single-word clutter cards.

⚠️ IMPORTANT BOUNDARIES & EXPECTATIONS:

1. Output Deliverables: ShortsPacer outputs broadcast-grade Master Audio (.mp3 / .wav) and Pacing Manifests (.json). It does not re-encode or jump-cut raw video frames, ensuring zero video compression degradation and allowing you to naturally place B-roll and graphics over vocal transitions in your video editor.
2. Recommended File Sizes: Optimized for short-form clips (TikTok, Shorts, Reels) under 200MB. For massive multi-hour 4K raw camera footage, please extract the vocal audio track first for instant sub-second processing.
3. Vocal Clarity Recommendation: Designed for voice recordings. If your track contains overpowering background music louder than the voice, process the clean vocal take before your final BGM mixdown for best de-breathing precision.

🚀 Usage & CLI Commands

# 1. Viral TikTok Mode (Default - direct video or audio)
python3 scripts/shortspacer_engine.py -i input_video.mp4 --preset viral

# 2. Warm Podcast De-Breath Mode
python3 scripts/shortspacer_engine.py -i podcast_episode.wav --preset podcast

# 3. Batch Process Entire Folder to Lossless WAV
python3 scripts/shortspacer_engine.py -i ./raw_takes/ --preset viral --format wav

Options

• -i, --input: Single audio file (.mp3, .wav), video file (.mp4, .mov), or folder.
• -o, --output: Target folder (defaults to ./cleaned_audio).
• -p, --preset: viral (default), podcast, or natural.
• --format: Output format, mp3 (320 kbps) or wav (lossless PCM).
• --no-norm: Disable -14 LUFS broadcast normalization if mastering separately.

📋 Generated Manifest (pacing_manifest.json)

Every run produces an audit manifest for your video editor:

{
  "status": "success",
  "preset": "viral",
  "loudness_standard": "-14.0 LUFS (TikTok/YouTube Shorts Standard)",
  "capcut_auto_captions_optimized": true,
  "processed_count": 1,
  "items": [
    {
      "filename": "voiceover.mp3",
      "input_type": "audio",
      "preset_used": "viral",
      "original_duration_sec": 47.21,
      "cleaned_duration_sec": 36.37,
      "time_saved_sec": 10.84,
      "retention_pacing_boost": "+23.0%",
      "loudness_mastered": "-14.0 LUFS",
      "capcut_caption_ready": true,
      "output_path": "/workspace/cleaned_audio/voiceover_tight.mp3"
    }
  ]
}

💡 Pro Tips for Creators

• Short-Form Focus: ShortsPacer is built for short-form video & audio clips. For the fastest sub-second experience, upload audio files (.mp3, .wav) or clips under 200MB.
• Massive Raw Footage: If you have multi-hour 4K raw camera footage, simply export the vocal/audio track first before running ShortsPacer for instant processing.
• Consonant Protection: Every cut point includes a 50ms voice-head acoustic buffer and 10ms micro-crossfading, ensuring explosive consonants (p, t, k) never get clipped.