Why Video-Use Adds Padding to Cut Edges During Rendering

Padding cut edges in video-use applies 30 ms audio fade‑in/fade‑out to every segment, eliminating audible pops and ensuring smooth transitions between clips.

The browser-use/video-use rendering pipeline deliberately pads each cut edge with a micro-fade. This quality‑control step prevents abrupt audio truncation artifacts that would otherwise mar the final output. According to the source code in helpers/render.py, this 30‑millisecond buffer is baked into every segment before lossless concatenation.

---

The Technical Implementation in helpers/render.py

The core logic resides in the segment extraction function. Lines 87‑90 construct an FFmpeg audio filter string that applies symmetric fades at both boundaries:

# helpers/render.py (simplified excerpt from L87-L90)

fade_out_start = max(0.0, duration - 0.03)
af = f"afade=t=in:st=0:d=0.03,afade=t=out:st={fade_out_start:.3f}:d=0.03"

This filter chain executes two operations:

1. afade=t=in:st=0:d=0.03 – ramps audio from silence to full volume over the first 30 ms.
2. afade=t=out:st={fade_out_start}:d=0.03 – ramps back to silence over the final 30 ms.

The fade_out_start calculation ensures the fade‑out never begins before the segment starts, handling edge cases where a clip is shorter than 30 ms.

---

What Padding Cut Edges Accomplishes

Pop‑Free Audio Transitions

Raw audio streams truncated at arbitrary sample boundaries produce click and pop artifacts. The 30 ms fade smooths the waveform discontinuity, yielding professionally clean cuts. This is especially critical when adjacent segments originate from different source files with varying DC offsets or levels.

Consistent Timing for Draft Verification

The padding cut edges mechanism supports the pipeline's draft mode (--draft flag). Because each segment starts and ends at mathematically predictable points, the timeline verification in helpers/timeline_view.py can precisely evaluate cut accuracy without compensating for arbitrary audio latency.

Lossless Concatenation Safety

Per‑segment MP4s include padded audio; the final concatenation step joins these pre‑smoothed streams. This avoids applying global fades that would distort cross‑segment timing.

---

Rendering Commands That Use Edge Padding

All rendering paths invoke the same extract_segment function with identical fade logic:

# Full‑resolution final output (30 ms fades applied per segment)

python helpers/render.py edl.json -o final.mp4

# Quick draft at 720p, ultrafast preset — padding preserved

python helpers/render.py edl.json -o draft.mp4 --draft

# Preview at 1080p, medium preset — edge padding active

python helpers/render.py edl.json -o preview.mp4 --preview

Regardless of resolution or encoding speed, the 30 ms audio fade‑in/fade‑out is non‑negotiable.

---

Summary

• Padding cut edges in video-use refers to mandatory 30 ms audio fades applied during segment extraction.
• The implementation lives in helpers/render.py at lines 87‑90, constructing FFmpeg afade filters.
• This eliminates audible pops, enables precise draft verification, and supports lossless concatenation.
• All rendering modes (--draft, --preview, default) preserve this padding behavior.

---

Frequently Asked Questions

What does "padding cut edges" mean in video processing?

Padding cut edges refers to adding a brief buffer — here, 30  milliseconds of audio fade — at the boundary where one clip ends and another begins. In video-use, this prevents waveform discontinuities that cause clicks or pops in the final mix.

Does padding cut edges affect video duration or sync?

No. The 30 ms padding applies only to audio and is symmetric (fade‑in + fade‑out). Visual content remains frame‑accurate, so lip‑sync and action timing are preserved. The helpers/timeline_view.py module verifies that padded cuts align with the original edit decision list.

Can I disable the 30 ms fade padding?

Not through official CLI flags. The fade logic is hardcoded in helpers/render.py because it is considered essential for output quality. Modifying the source would require editing the af filter string construction around line 87.

Does the draft rendering mode skip edge padding?

No. Both --draft and --preview modes retain the full 30 ms fade logic. These flags only adjust resolution, encoding preset, and bitrate — they do not alter the audio processing pipeline defined in extract_segment.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}