Vertical video is footage shot and delivered in a 9:16 aspect ratio — taller than it is wide, sized for a phone held upright. If you are asking what is vertical video in a stock context, the short answer is that it is a separate deliverable, not a crop. Agencies treat it as its own format, with its own list of accepted resolutions, its own duration guidance, and its own keywording rules. Clips that were rotated or cropped out of 16:9 footage after the shoot get rejected.
That last point is where most contributors lose files. This guide covers the exact resolutions Adobe Stock and Shutterstock accept, the frame rate and codec constraints, how vertical changes your metadata, and an honest read on where buyer demand actually sits.
What vertical video is — and what it is not
A vertical clip uses the full sensor in a portrait orientation. The camera is turned, the composition is built for a tall frame, and the footage leaves the card already at 9:16. That is what agencies mean by "native" vertical, and it is the only version they want.
The alternative — cropping a horizontal clip — fails for two reasons. First, the math: taking a 9:16 slice out of a 1920x1080 frame at full height leaves you 608x1080. Roughly two-thirds of the horizontal frame is discarded, and the result sits well under the 1080x1920 floor every major agency enforces. Second, it looks like what it is. Framing built for a wide shot rarely survives the crop, and reviewers notice heads clipped at the top and subjects shoved against an edge.
Up-scaling does not rescue it either. Adobe Stock's video requirements state plainly that footage should not be up-converted — submit it as shot, or smaller if necessary. A 608x1080 crop stretched to 1080x1920 is a technical rejection waiting to happen.
Vertical video dimensions agencies actually accept
Adobe Stock publishes four accepted vertical resolutions. Everything else is a rejection:
- 1080x1920 — the floor. Standard vertical HD.
- 2160x3840 — vertical UHD, the most common 4K vertical output.
- 2160x4096 — vertical DCI 4K.
- 2304x4096 — the widest vertical variant accepted.
Upload in the clip's native frame size. Adobe is explicit that you should not down-convert or up-convert from the original camera resolution unless the source is larger than 4K. Duration runs 5 to 60 seconds across the board, but for vertical specifically Adobe recommends 10 to 30 seconds per clip — shorter than the 15-second-plus guidance it gives for horizontal footage. Maximum file size is 3900 MB.
Shutterstock's technical envelope is close but not identical. It accepts .mov and .mp4 containers only, holds the same 5-to-60-second window, and caps files at 4 GB. Its accepted codec list includes PhotoJPEG, Apple ProRes 422 (HQ, LT and 4444), H.264, PNG, MPEG-4 Visual and Apple Animation. H.265 is accepted but not currently recommended, because of a known HDR compatibility issue.
Adobe accepts MOV, MPG and MP4, recommends ProRes or H.264, and also takes MPEG-2, MPEG-4, Motion JPEG and PNG. Frame rates must be standard — 23.98, 24, 25, 29.97, 30, 50, 59.94 or 60 — and you should not convert away from your native rate unless the clip is deliberately slow or fast motion. If you keep audio, it has to be 48 kHz, 16-bit uncompressed. Submitting silent clips is perfectly acceptable, and often simpler than chasing releases for every identifiable voice.
For AI-generated video, Adobe sets a minimum of 1080 pixels on both sides — so 1080x1920 vertical — inside the same 5-to-60-second window, with the additional requirement that the clip meets its generative AI content rules and shows no artifacts or temporal inconsistency.
The detail almost everyone misses: your thumbnail is frame five
Adobe generates a clip's preview image automatically from the frame five seconds in. Nobody chooses it for you. On a 10-second vertical clip that is the midpoint, and it is the single image a buyer scanning search results will judge the whole file on.
This changes how you cut. If your strongest composition lands at second two and the subject has walked out of frame by second five, the thumbnail sells an empty room. Plan the edit so the five-second mark holds the shot you would have picked yourself. Leave one to five seconds of handle at each end, cut cleanly so the clip does not open on the tail of another take, and trim out focus hunts and zoom adjustments before they reach the timeline.
Metadata for stock videos follows different rules than photos
Here is a rule that contradicts itself unless you read carefully, and it costs contributors acceptances. Adobe's general keywording guidance tells you not to use content types or camera specifications as keywords — no "video", no "4K", no camera brand names. But its vertical video guidance asks you to do the opposite: include the keywords "vertical" and "video" in your titles and keywords.
The exception exists because buyers searching for 9:16 assets have no other way to filter for them. Format is the search intent. So on vertical clips, and only on vertical clips, those two terms belong in your metadata. Everywhere else the ban still applies.
Beyond that, technique and camera-movement terms carry real weight on video in a way they do not on stills. If a clip is a timelapse, an aerial, a slow push-in or a handheld follow, say so in both the title and the keywords — those are the words motion buyers type. Trademarks, logos, brand names, artist names and references to in-copyright creative works remain prohibited, and video is scrutinised harder here than photography because signage and screens tend to wander into moving frames.
Because video metadata and photo metadata diverge like this, tooling built purely for stills tends to produce keyword sets that read wrong on motion files. If you shoot both, it is worth using something that handles photo and video metadata under one system, rather than running two disconnected workflows and hand-patching the differences.
Where the demand for vertical stock actually is
The honest position: vertical is a supply gap, not a guaranteed sale. Demand comes almost entirely from social-first advertising — short-form feeds, in-app ad units, story placements and the paid social work agencies produce at volume. Those buyers need footage that fills a phone screen edge to edge, and the stock libraries they search were built over two decades around 16:9.
That imbalance is visible from the contributor side too. Adobe's own community forums carry standing requests for stronger 9:16 support in search and submission, which is the sort of friction that only surfaces when buyers are looking for something the catalogue is thin on.
What that does not mean is that any vertical clip sells. Format scarcity gets your file into a smaller results page; the footage still has to earn the licence once it is there. The categories that move are the ones paid social actually buys: people using products, hands and close work, food, fitness, workspace scenes, travel and city motion with room for text overlay at the top and bottom thirds. Leaving that overlay space deliberately is one of the cheapest things you can do to make a clip more licensable, because the buyer will be putting a caption there whether you planned for it or not.
Building a vertical workflow that scales
Shooting vertical properly means shooting it as its own pass. If a setup is worth capturing, capture it twice — once wide, once tall — rather than hoping one file serves both. Keep the two orientations in separate bins from the card onwards, because the specs, the durations and the keyword sets diverge and mixing them is how converted clips end up in a vertical batch by accident.
Delivery is where the time goes. Each agency wants its own field mapping, its own category structure and its own release handling, and video metadata in particular rewards automation — motion files are larger, slower to move, and more expensive to re-upload when a batch is rejected on a metadata technicality.
This is the problem Rastock is built for. It generates titles, descriptions and keywords for photo and video, then delivers to 10+ agencies over FTP/SFTP — Adobe Stock, Shutterstock, Magnific, Alamy, Depositphotos, Pond5, Canva, 123RF, VectorStock and Vecteezy — with per-agency field mapping and upload status tracking in one place. You can add any other agency that accepts FTP or SFTP yourself. Just as importantly, it gives you control rather than just speed: policy-aware tagging, banned and mandatory keyword lists, and rollback when a batch goes out wrong. There is no revenue share, no ownership claim over your files, and IPTC and CSV export stay open so nothing is locked in. See how it works, or check the pricing — the trial runs 14 days and does not ask for a card.
A pre-upload checklist for vertical clips
- Shot natively at 9:16 — never rotated or cropped from horizontal.
- Resolution matches one of the four accepted sizes, at native frame size, with no up-res.
- Duration between 10 and 30 seconds, with 1-5 second handles at each end.
- The frame at five seconds is one you would happily use as the thumbnail.
- Standard frame rate, accepted codec, audio at 48 kHz 16-bit or removed entirely.
- "Vertical" and "video" present in the title and keywords, plus any technique or camera-movement terms.
- No logos, trademarks or brand names anywhere in frame or metadata; releases attached for every identifiable person.
Vertical is one of the few places in stock right now where the format itself is a differentiator. That advantage lasts exactly as long as the catalogue stays thin — which is an argument for shooting it properly now rather than converting badly later.
Frequently asked questions
What is vertical video in simple terms?
Vertical video is footage where the frame is taller than it is wide, in a 9:16 aspect ratio, matching how a phone is held upright. In stock terms it means the clip was shot natively in that orientation rather than cropped from a horizontal file. Agencies list separate accepted resolutions for it, starting at 1080x1920, and reject clips that were converted from 16:9 after the shoot.
Can I crop a 16:9 clip to make a vertical version?
No. Adobe Stock explicitly asks contributors to shoot vertical natively and not to convert from 16:9. There is also a hard technical problem: a 9:16 slice from a 1920x1080 frame is only 608x1080, well below the 1080x1920 minimum, and up-scaling it back up is separately prohibited. Shoot a dedicated vertical pass instead.
What resolutions does Adobe Stock accept for vertical video?
Four: 1080x1920, 2160x3840, 2160x4096 and 2304x4096. Upload at your camera's native frame size without converting up or down, unless your source is larger than 4K. For AI-generated vertical video the minimum is 1080 pixels on both sides, meaning 1080x1920, and the clip must also satisfy Adobe's generative AI content requirements.
Should I put the word "video" in my keywords?
Normally no — Adobe treats content-type words like "video" and camera specs like "4K" as prohibited keywords. Vertical clips are the documented exception: Adobe asks you to include both "vertical" and "video" in the title and keywords, because format is what buyers are searching for. Use those terms on vertical files only, and keep the general ban in place everywhere else.
How long should a vertical stock clip be?
Adobe recommends 10 to 30 seconds for vertical clips, inside a hard limit of 5 to 60 seconds. That is shorter than its general guidance for horizontal footage. Keep one to five seconds of edit handle at each end, and remember that the preview thumbnail is pulled automatically from the frame five seconds in — so a 10-second clip needs its strongest moment near the midpoint.
Is vertical stock video actually selling?
Demand comes mainly from social-first advertising, where every placement is a phone screen and stock libraries built around 16:9 are thin on supply. That scarcity helps a well-shot clip surface, but it does not sell weak footage on its own. The categories that move are what paid social buys — people with products, food, fitness, workspaces, travel — ideally composed with clear space in the top and bottom thirds for caption overlays.