Guides · How-to

Stop keyframing every punch-in by hand

Premiere doesn't have an auto-zoom feature. It has the Motion effect's Scale and Position properties, an Ease In / Ease Out curve, and however many keyframe pairs you're willing to set by hand. Here's the manual method end to end — and what to do once punch-ins are the twenty minutes you dread on every single video.

· 7 min read

What a punch-in is doing, mechanically

A punch-in is a scale-and-reposition move: the frame gets larger and recentres on the speaker for a few frames to a couple of seconds, then holds or eases back out. On a single-camera talking-head shot it does the job a second camera angle would do on a multicam set — it breaks up a static frame and puts a visual accent under the words that carry the point. Premiere has no button for this. What it has is the Motion effect's Scale and Position properties, keyframed by hand, one punch at a time.

Doing it by hand in Premiere Pro

This is real work, and it is worth doing once so you know what the automated version is actually saving you. For a ten-minute talking-head video with a punch every minute or two, budget somewhere between twenty and forty minutes once you include the time spent finding where each one should land.

  1. Select the clip on the timeline and open Effect Controls (Window → Effect Controls), then twirl down Motion.
  2. Move the playhead to the frame just before the line you want to punch in on, and click the stopwatch next to Scale — and Position, if you're also reframing — to set a starting keyframe at the current value.
  3. Move the playhead forward — a handful of frames for a snap punch, up to a second for a slower push — and raise Scale to the punched-in value. 110–130% is the usual range; how far depends on how much resolution headroom your source gives you.
  4. Select both new keyframes, right-click, and set Temporal Interpolation to Ease In / Ease Out, or drag the Bezier handles in the graph, so the move isn't linear. A linear scale keyframe looks like a stutter, not a punch.
  5. If the punch also needs to recentre the frame — a wide two-shot tightening to a single — keyframe Position the same way, timed to land together with the Scale change.
  6. Scrub at quarter speed and nudge the in-point until it lands on the stressed word, not a frame or two after it. This is the step that separates a punch that reads as intentional from one that reads as a glitch.
  7. Repeat for every line worth the emphasis. Copy the two keyframes and paste them onto the next clip's Motion property to avoid re-setting the same values from scratch.

Watch how far you push Scale relative to your footage. A 4K source cut into a 1080p sequence has real headroom — 130% is usually clean. A 1080p source pushed past about 115–120% starts to look soft, and there's no fixing that afterward; you're stretching pixels that don't exist.

Finding the punch without guessing

The hard part isn't the keyframing, it's picking where to punch. Scrubbing a waveform and eyeballing the loud parts finds volume, not emphasis, and those are not the same thing — a punch on the wrong word reads as random. Transcribe the sequence first (Window → Text → Transcript, then Transcribe) and read it instead of scrubbing it. The words a speaker leans on when they land a point are obvious on the page in a way they aren't in a waveform, and the transcript hands you the exact timestamp to keyframe against instead of hunting for it by ear.

What to do about the parts that are still manual

There are two real ways to speed this up, and they trade off differently.

Buy a zoom-transition preset pack

Motion Array, Envato and similar sell scale-and-ease presets you drop onto a clip — they save you the keyframe-and-Bezier setup in the steps above, and they're worth it if your punches are stylistically fussy. What they don't do is decide where the punch goes or how long to hold it; you're still finding the moment and setting the in-point yourself, which — per the point above — is the half actually worth doing off a transcript rather than a waveform.

Generate the keyframes off the transcript inside Premiere

This is what AutoZoom in Backstage Cut does. It reads the same transcript your captions come from, picks the moments worth a punch based on what's being emphasised in the speech, and writes ordinary Scale keyframes onto a non-destructive adjustment layer above your clip — not a burned-in render, not a new clip, just keyframes you can nudge, retime or delete like anything else in Premiere. Reject a punch it suggests and your timeline is back to exactly what it was before.

Which one to pick

  • One video a week, one or two punches in it: do it by hand. Four or five keyframe pairs is fifteen minutes, not worth a subscription.
  • Punches are frequent but your style is fixed: a preset pack removes the easing-curve fiddling and leaves you to find the moment, which is the faster half already.
  • Every talking-head video this week needs punches timed to what's actually being said: generating them off the transcript removes the finding-the-moment step too — the one that actually eats the time.

Backstage Cut starts with 30 free minutes and no card — enough to run AutoZoom across a couple of real edits and see whether the timing holds up on your own voice and your own pacing.

Try it on the video
you're editing tonight

30 free minutes is a whole video, start to finish, on your own footage — no card, and nothing to undo if you don't like it.