Can you change the aspect ratio in the middle of a video?

Yes, visually. A video file keeps one resolution for its whole length, so a mid-video change is the visible frame closing in to the new shape inside that canvas. In Valmera you say “go vertical for this part, then open back out,” and the agent animates the frame there and back.

How do I switch from 16:9 to 9:16 and back in one video?

Name both moments: “go 9:16 when the phone demo starts and back to full width at the reveal.” The agent places one shift at each. The frame narrows to a centered 9:16 window over 0.8 seconds by default and pushes in as it narrows.

Will TikTok or YouTube show the vertical section full-screen?

No. Players size themselves to the file once and don’t resize during playback. Export 9:16 for TikTok, Reels or Shorts and shift to a wide window inside it, or export 16:9 for YouTube and shift to a vertical window inside that.

What happens to captions when the frame narrows?

Caption timing never changes. The bars sit above captions and titles, so text outside the narrowed window is hidden while the frame is closed. Place captions for the narrowest frame in the edit.

Does the shift stay in place if I cut the video later?

Yes. Each shift is anchored to the moment in the footage it was placed on. Cut earlier material and it moves with its moment; cut the moment itself and the shift is removed, and the agent tells you.

How is this different from resizing the whole video?

Resizing sets one shape for every frame of the export by cropping or fitting. An aspect shift keeps the export’s shape and changes the visible frame for part of the runtime, as an editorial beat.

Change the Aspect Ratio in the Middle of a Video

To change the aspect ratio partway through a video, upload it to Valmera and name the moment: “go vertical when the phone demo starts, then open back out at the reveal.” The AI agent morphs the visible frame to the new shape over 0.8 seconds, holds it, and opens it back to the full frame where you asked.

A video file has one resolution for its whole length, so the shift is the frame closing in inside that canvas, with bars around the new shape. Nothing is retimed: audio, cuts and captions stay in sync.

Make the Frame Part of the Edit

Name the moment, watch the frame move, export. Creating an account and uploading are free. Editing requires a subscription.

Create an account →

See pricing →

How to Change the Aspect Ratio Mid-Video in Valmera

1
Pick the delivery canvas
Upload your footage (MP4, MOV, M4V, MKV or WebM, up to 100 GB and 3 hours) and set the shape you’ll publish in: “make this 9:16 for TikTok,” or leave it 16:9 for YouTube. That shape is fixed for the whole export; shifts happen inside it.
2
Name the moment and the new shape
“Go 9:16 when he picks up the phone, then back to full width when he puts it down.” The agent finds both moments in the word-level transcript and the frames and places a shift at each.
3
Tune the move
Change the morph length (0.1–4 seconds), the bar color, or the push-in: “take two seconds,” “make the bars white,” “don’t push in.”
4
Preview, then export
The agent renders a preview and checks the frames where each shift lands. When it looks right, click Export in Studio for a full-resolution MP4.

PROMPTS TO PASTE

Go 9:16 when the phone demo starts, then back to full width at the reveal.

Squeeze to square at 0:42 and hold it until 1:10.

Make the frame close in slower, over two seconds, with white bars.

Don't push in when it goes vertical.

What Happens to the Frame

A 9:16 window inside a 16:9 canvas keeps the full height and 31.6% of the width. With the push-in on, the picture grows 34% as the bars close, so the move reads as a deliberate reframe.

The canvas never changes size. Bars slide in from the edges and leave a centered window in the new shape: a 9:16 window inside 16:9 keeps 31.6% of the width, and a 1:1 window keeps 56.25%. Inside a 9:16 canvas, a 16:9 shift is a letterbox instead. The move eases in and out, and the frame holds its new shape until the next shift.

With the push-in on (the default), the picture zooms in as the frame narrows, in proportion to the area it loses: 34% for 16:9 to 9:16, 22% for 16:9 to 1:1, never more than 35%. Turn it off and the picture stays put behind the bars.

Settings

Setting	Options	Ask for it like this
Shape	16:9, 9:16, 1:1, 4:5, 4:3, or source (back to the full frame)	“Go square at 0:42.”
Morph length	0.1–4 seconds; default 0.8	“Take two seconds to close in.”
Push-in	On by default; scales with the area the frame loses, up to 35%	“Don’t push in.”
Bar color	Any hex color; default black	“Make the bars white.”
Timing	Anchored to a moment in the footage	“Move it to when she says ‘here’s the trick.’”

Pick the Delivery Canvas First

Players size themselves to the file once and never resize during playback, so choose the export shape first, then shift inside it.

Vertical canvas, wide moment. Export 9:16 for TikTok, Reels or Shorts and open to a 16:9 window for an establishing shot or a screen recording.
Wide canvas, vertical moment. Export 16:9 for YouTube and squeeze to 9:16 for a phone demo or a quoted vertical clip.
The mistake to avoid. Building in 16:9 and expecting TikTok to show the vertical section full-screen. The file is 16:9, so the whole upload is letterboxed and the vertical section plays as a small strip.
Need both shapes full-screen? That’s two exports: change the whole frame with resize video or auto reframe and export again.

What Stays in Sync

A shift changes only what is visible, so no segment is retimed and no caption moves by a frame. Each shift is anchored to the moment in the footage it was placed on: cut something earlier and the shift moves with its moment, and the agent says where it went. Cut the moment itself and the shift is removed rather than left over the wrong sentence.

The bars sit above captions and titles, so text placed for the wide frame can be hidden while the frame is narrow. Size and place captions for the narrowest frame in your edit.

Limits

The window is always centered on the canvas. For an off-center subject, add a zoom aimed at them over the same stretch (“push in on him while it’s vertical”); see zoom effects. Bars are a flat color, with no blurred backdrop for a mid-video shift. Two morphs can’t overlap, and a shift to the shape the frame is already in is refused, because it would change nothing on screen. The export keeps one resolution, so no platform shows a shifted section full-screen.

Related Workflows

The reframing docs cover whole-video ratios and the framing panel. Using Claude, ChatGPT or Cursor? The same add_aspect_shift and remove_aspect_shift tools are available through Valmera's MCP server.

Creating an account and uploading are free. Editing requires a subscription — see plans.