Exporting video for portrait and 4:5 feeds

You have a finished edit in 16:9 and you need it vertical. You crop it, and now the presenter's forehead is missing and the product is half out of frame.

This is not a problem you fix at export. It is one you avoid earlier, and the fix is mostly about knowing which shapes you actually need.

The three shapes worth knowing

9:16 — full-screen vertical. 1080 × 1920. Reels, TikTok, Shorts, Stories. Fills a phone screen completely, which is why it dominates.

4:5 — tall but not full screen. 1080 × 1350. The tallest shape a feed post can occupy on Instagram. It takes up substantially more vertical space than a square without going full-screen, which makes it the best-performing shape for in-feed content that people scroll past.

1:1 — square. 1080 × 1080. Safe everywhere, optimal nowhere. Still useful when one asset has to work across several platforms and you cannot make variants.

16:9 remains correct for YouTube proper, for anything that will be watched on a laptop, and for embeds on your own site.

The practical answer for most people: 9:16 for stories and short-form, 4:5 for feed posts, 16:9 for long-form. That is two or three exports from one edit, not three separate edits.

Safe areas: the reason your text gets covered

Platforms draw their interface on top of your video. On a vertical video, the bottom fifth is occupied by captions, username and buttons; the right edge has the action column; the top has a status bar and sometimes a progress indicator.

Anything you want people to read needs to live in the middle 60–70% vertically and avoid the right-hand strip. This is the single most common avoidable mistake in short-form video: perfectly good on-screen text, sitting underneath a Follow button.

The same applies to your subject. Compose so the face or product sits in the upper-middle, not dead centre and definitely not low.

Shoot so you can reframe later

If you know a clip will end up in more than one shape, the cheapest insurance is to shoot wider than you need and leave headroom. A subject framed tightly in 16:9 cannot be recovered in 9:16; the same subject framed with air around them can be cropped to anything.

Shooting in 4K and delivering at 1080 gives you the same freedom for free: you can crop and reposition within the frame without losing resolution, because you are throwing away pixels you were never going to deliver.

If you are shooting for vertical specifically, shoot vertical. Rotating a landscape phone video does not make it vertical, it makes it a landscape video with black bars or a heavy crop.

Reframing without ruining it

When you take a wide edit to vertical, resist the temptation to crop the whole thing identically. Each shot usually wants its own position — a two-person interview needs to follow whoever is talking, a product shot needs the product centred, an establishing shot may work better scaled down with a background than cropped at all.

Doing this shot by shot takes ten minutes on a sixty-second video and is the difference between "reformatted" and "made for this".

Export settings that avoid disappointment

• Resolution: 1080 on the short edge is enough for every feed. 4K vertical is a large file that the platform will re-encode down anyway.
• Frame rate: match your source. Converting 24 to 30 introduces judder for no benefit.
• Codec: H.264 remains the safest for upload compatibility. HEVC produces smaller files at the same quality but is less universally accepted by upload pipelines.
• Bitrate: higher than you think. Every platform re-compresses what you upload, so you are giving it the best possible input, not the final output. A visibly soft upload becomes a very soft post.

And export the captions burned in for feeds — platform auto-captions are inconsistent and often off by default.

Doing this in Garfi

Garfi Video Editor works in landscape, portrait, square and 4:5 canvases, so the reframe is a property of the project rather than a crop you fight with at the end. You can crop, rotate and flip footage per clip to reposition each shot inside the new shape.

It exports MOV or MP4 in H.264 or HEVC, up to 4K where the hardware supports it. Text, transitions and editable automatic captions are on the same timeline, so burning captions in for a feed is part of the export rather than a second tool.

Projects save as portable .gvideo packages, which means producing a 4:5 variant of an edit you finished last month is opening a file, not rebuilding it. There is no watermark on what you export, no account, and no generative AI in the app.