Editing

Video Export Settings for Social Media: A Cheat-Sheet Mindset

Every platform re-encodes your upload, so the goal is not a perfect file — it is a file that survives the re-encode. Here is the practical export logic for vertical, square, and horizontal delivery, plus color and loudness traps that quietly ruin uploads.

JML Tech Studios7 min read
Editor comparing vertical, square, and horizontal export presets on a color-calibrated monitor

Here is the fact that simplifies every export decision: no social platform shows viewers your file. Every platform re-encodes every upload into its own compressed versions. Your export is not the deliverable — it is the source material for the platform's encoder. Once you internalize that, the goal stops being "maximum quality file" and becomes "clean, correctly-shaped, generously-encoded file that gives the platform's encoder the best possible starting point."

That mindset replaces memorizing fifty preset numbers, which is good, because platform specs change quietly and often. What follows is the durable logic — shape, bitrate philosophy, color, loudness — plus a delivery table with safe ranges rather than false precision.

Shape first: aspect ratio and resolution per placement

Placement decides shape, and shape is the one thing the platform cannot fix for you. Ship the wrong aspect and the platform crops or letterboxes — both are quality losses you chose by accident.

  • Vertical 9:16 (1080×1920): Reels, TikTok, Shorts, Stories. This is the native shape of phone-first placements; anything else gets padded or cropped.
  • Square 1:1 (1080×1080): feed posts where you want maximum frame height in a scrolling feed without committing to full vertical. Increasingly a legacy shape, but still safe in feeds.
  • Portrait 4:5 (1080×1350): the feed-post workhorse on platforms that support it — taller than square, fills more screen than 16:9 without being a Reel.
  • Horizontal 16:9 (1920×1080 or 3840×2160): YouTube proper, embedded web players, TV. On phone-first platforms, 16:9 shrinks into a letterboxed strip; use it only where horizontal viewing genuinely happens.

Plan this at the edit stage, not export. Cropping a finished 16:9 timeline to 9:16 in the export dialog throws away most of your frame and usually beheads your subjects. If a project needs multiple shapes, duplicate the sequence per aspect and re-frame shot by shot — tedious, and visibly better than auto-crop. Building reframing into your project template is one of the habits covered in editing videos faster.

Bitrate philosophy: feed the re-encoder

Because the platform re-encodes, obsessing over exact bitrates is wasted energy — but starving the upload is a real error. Compression artifacts compound: whatever blockiness your export contains, the platform's second compression pass makes worse. The philosophy is simple: export comfortably above what the platform will serve, from your highest-quality source, in one generation.

In practice, for H.264 delivery: roughly 8–16 Mbps for 1080p and roughly 35–60 Mbps for 4K sits comfortably in the safe zone for social uploads — high enough that the platform's encoder has clean material, low enough that uploads are quick. VBR with two-pass encoding when time allows, high profile, standard frame-rate matching your timeline. Going far higher mostly costs upload time; going far lower is where visible damage starts. H.265/HEVC uploads are accepted by major platforms and can carry the same quality at smaller sizes, but H.264 remains the compatibility default that never surprises you.

One genuinely useful trick: on platforms that visibly crush 1080p uploads, uploading a 4K master of the same edit often gets you a better-looking result, because many platforms allot their higher-quality encoding paths to higher-resolution sources. It costs nothing to try on an important piece.

Color space gotchas: log stays home

The most common "my video looks washed out on my phone" complaint has one of two causes, and both are color-management mistakes, not platform mysteries.

First: never export ungraded log. Log profiles are capture formats, flat and desaturated by design, meant to be graded to a display standard before delivery. Social delivery should be standard dynamic range in Rec.709 with a standard gamma unless you are deliberately and knowledgeably shipping HDR. If grading is new territory, color grading basics covers the path from flat log to a delivered image.

Second: tag what you ship. Mismatched or missing color metadata is why an export looks right in your NLE and wrong in a browser or phone player. Keep the pipeline boring: Rec.709 timeline, Rec.709 export, correct color tags, then check the actual uploaded result on an actual phone — the device your audience uses is the only monitor that gets the final vote. HDR delivery is a separate, deliberate workflow; accidentally shipping HDR-flagged files to SDR viewers produces dim, lifeless playback and is worth avoiding until you choose to do it on purpose.

Audio loudness: the forgotten spec

Platforms normalize loudness — they turn down uploads that are too hot, and some quietly do nothing for uploads that are too quiet, leaving them limp next to everything else in the feed. Mixing to a consistent loudness target is the fix. General streaming practice clusters around integrated loudness in the range of roughly −16 to −13 LUFS, with −14 LUFS a widely used reference point; landing anywhere in that band with true peaks kept below about −1 dBTP puts you in safe territory on every major platform.

Two practical rules do most of the work. Use a loudness meter, not your ears alone — headphone listening at midnight lies. And keep dialogue as the anchor: music beds that fight the voice lose more viewers than any bitrate decision ever will. Export audio as AAC at a standard high-quality setting; audio bitrate is cheap, so never be the person who saved half a megabyte by making the interview sound underwater.

The delivery table

PlacementAspect / typical resolutionCodecVideo bitrate (safe range)Notes
Reels / TikTok / Shorts9:16 — 1080×1920H.264 (H.265 ok)~8–16 MbpsKeep text in the middle band; expect aggressive re-encode
Instagram / Facebook feed4:5 — 1080×1350 (1:1 ok)H.264~8–16 MbpsTaller shapes earn more screen in scroll
YouTube (standard)16:9 — 1920×1080H.264~10–20 MbpsTwo-pass VBR from a clean master
YouTube (4K master)16:9 — 3840×2160H.264 / H.265~35–60 MbpsHigher-res uploads often receive better encoding paths
Stories9:16 — 1080×1920H.264~8–12 MbpsShort runtime; UI overlays top and bottom heavily
Client review / archiveNative timelineHigh-bitrate master or mezzanine codecAs high as practicalThis is your one true master; socials are derivatives of it

Treat the ranges as floors of comfort, not ceilings of law. Platforms revise specs; the philosophy — right shape, generous clean encode, Rec.709 tagged correctly, normalized loudness — does not.


Build presets once, verify quarterly

  1. Export one master at your timeline's native resolution and highest sensible quality. Archive it.
  2. Build named presets per placement — shape, bitrate range, loudness-checked audio — and derive every social version from the master, never from another compressed export.
  3. Upload, then check each platform on a phone: sharpness, color, safe zones, loudness against neighboring content.
  4. Once a quarter, skim each platform's current recommended specs and nudge presets if anything moved.

That is the whole system. One master, a handful of presets, a phone check after upload. Exports stop being a research project and go back to being the ten quiet minutes at the end of the edit — where they belong.

Frequently asked questions

What is the best export format for social media?
H.264 in an MP4 container at a generous bitrate remains the universal safe choice — every platform accepts it and re-encodes it predictably. H.265/HEVC is widely accepted too and carries similar quality in smaller files, but H.264 is the default that never causes compatibility surprises.
What bitrate should I use for 1080p social uploads?
Roughly 8–16 Mbps H.264 is a comfortable range for 1080p social delivery — high enough that the platform's re-encoder gets clean source material, low enough that uploads stay fast. Exact numbers matter less than exporting from your highest-quality master in a single generation.
Why does my video look washed out after uploading?
Almost always a color management issue on the export side: ungraded log footage, an HDR-flagged file shown to SDR viewers, or wrong/missing color tags. Deliver graded Rec.709 with correct metadata, then verify the uploaded result on a phone rather than trusting a desktop player preview.
How loud should audio be for social video?
Mix to an integrated loudness in the general streaming band of roughly −16 to −13 LUFS — −14 LUFS is a widely used reference — with true peaks under about −1 dBTP. Platforms normalize loud uploads downward, so mixing hotter than that buys nothing and costs headroom.