how to hand off rough cuts to captions without losing hook timing beats

# The Rough Cut Handoff That Actually Works

I used to think the edit was done when the visuals locked. I’d export a file called “FINAL_FINAL_v3.mp4,” send it to my captioner, and then get the SRT back with all the timing just… off. The punchline landed in the middle of a cut. The hook felt sluggish. I’d have to go back, tweak the caption file myself, or worse, ask for a revision and wait another day. This broke my entire weekly rhythm.

The mistake was assuming a “final” video was the right deliverable for captions. It’s not.

## What You’re Actually Handing Off Isn’t the Video

It’s the **timing beats**. The captioner doesn’t need your color grade or your perfect sound mix. They need to see where words start and stop with pixel-perfect accuracy relative to your cuts. Sending a fully rendered, compressed MP4 introduces a tiny bit of audio drift, enough to throw off frame-accurate timing. I was wrong. The polished file is the enemy of good caption timing.

My embarrassing moment? I blamed three different freelancers for “sloppy work” before I realized I was the common denominator, sending them a format that was fundamentally unsuited for the task.

## The “Captioning Proxy” Workflow

Here’s what I stopped doing: sending finished videos for captioning.

Now, I have a sequence in my editor called “ROUGH CUT - CAPTION READY.” Before I do any fine-tuning of visuals or audio, I make sure the narrative flow is there. Then, I do two things:

1. **Freeze the audio track.** I bounce the audio to a single, clean WAV file and lay it back onto the timeline, replacing all the raw clips. This eliminates any possibility of playback jitter or drift. 2. **Export a low-res reference with burned-in timecode.** Not the fancy SMPTE kind, just the simple hours:minutes:seconds:frames counter right in the image. Most editors can do this on export. The file is tiny, exports in seconds, and is useless to anyone but me and my captioner.

**This is the file I send.** The captioner uses the burned-in timecode to mark in/out points. When they send back the SRT, every single cue line is referenced to that timecode. When I import it into my *actual* final timeline, it snaps into perfect alignment. The hook lands on the beat. The pauses feel intentional.

## The Realization

The blunt truth? You can’t hand off creative timing. You have to engineer it.

What actually changed was that I stopped thinking of captions as a post-production polish and started treating them as a core layer of the edit, like sound design. The handoff became a technical spec, not a creative review.

It shaved a full day off my turnaround for every short-form piece. No more back-and-forth. No more manual tweaking of caption files. The workload reduction was immediate and stupidly simple in hindsight. The time I saved went straight into finding the next client.