Most short-form editing advice is about hooks and pacing. Those matter, and they're covered in our pacing and retention guide. This is the other half — the craft and technical decisions that determine whether a well-conceived video actually looks good on a phone.
These are unglamorous and they're where most amateur edits lose.
Frame for the safe zone, not the canvas
Vertical video is 1080×1920, and a meaningful portion of that frame is covered by interface. Captions, usernames, buttons, and the description sit on top of your video, and they differ per platform.
Practical approach: keep anything important — faces, text, product detail — within roughly the centre 80% vertically and 90% horizontally. The top strip and bottom third are contested space.
Set up guides in your editor once and leave them on. The most common amateur mistake is burned-in captions positioned at the bottom, where TikTok's own interface sits directly on top of them.
If you're repurposing one edit across platforms, frame for the most aggressive interface and accept some dead space elsewhere. Re-framing per platform is better but rarely worth the time.
Captions: burned in, and readable
A large share of viewing happens with sound off. Captions aren't an accessibility extra — they're how most of your audience receives the content.
Burn them in. Platform auto-captions are inconsistent, positioned unpredictably, and often wrong on technical terms and names.
What makes them readable:
- High contrast against any background. A solid or semi-transparent backing shape, or a heavy stroke. White text on unpredictable footage will be illegible somewhere.
- Large. Bigger than feels right on a desktop editing monitor. Check on a phone.
- Few words on screen at once. Three to six. Full sentences force reading rather than glancing.
- Timed to speech, not to a fixed interval. Words appearing before they're spoken pull attention away from the speaker.
- Sans-serif, heavy weight. Fine strokes disappear at phone size and under compression.
Emphasis colour on key words is a real technique — one accent colour, used sparingly. Every word highlighted means no word is highlighted.
Audio is half the video
Viewers tolerate mediocre picture and abandon bad audio. If you fix one production variable, fix this one.
Levels. Aim for dialogue peaking around −6dB with no clipping. Loudness normalisation on the platforms means wildly loud exports get turned down anyway, usually badly.
Music well under speech. Music should sit clearly beneath dialogue — audible, never competing. Duck it under speech rather than setting one level throughout.
Clean the source. Noise reduction on a hissy recording, and a high-pass filter to remove low-frequency rumble, both do more for perceived quality than any amount of colour work.
Sound effects punctuate, they don't decorate. A transition sound on every cut is exhausting within ten seconds. Use them to mark genuine structure — a topic change, a reveal.
Check licensing. Platform-provided audio libraries are safe for organic posting. Trending commercial tracks generally are not cleared for business or advertising use, and this catches brands out — the video gets muted or taken down after it starts performing.
Colour: consistency over intensity
Colour work is where editors overspend time relative to impact for short-form. What actually matters:
Consistency across shots. Clips that shift in temperature or exposure between cuts look amateur regardless of how good any individual grade is. Match first, style second.
Mobile-appropriate contrast. Phones are viewed in bright environments. Slightly increased contrast and saturation reads better than a flat cinematic grade that looks great on a calibrated monitor and muddy in daylight.
Don't crush shadows. Detail lost to compression is unrecoverable, and platform compression will take more.
Skin tones first. If faces look right, most viewers perceive the whole image as correct.
If you're shooting log footage, a correct conversion to Rec.709 plus a light grade beats an elaborate look. If you're shooting standard profiles on a phone, minimal correction is usually right.
Export settings that survive compression
Every platform re-encodes your upload. Your job is to give the encoder the cleanest possible source.
Reasonable targets:
- Resolution: 1080×1920. Higher offers little benefit after re-encoding.
- Frame rate: match your source — 30 or 60fps. Don't convert unnecessarily.
- Codec: H.264, high profile.
- Bitrate: generous. Around 10–20 Mbps for 1080p vertical. It gets re-encoded regardless; a higher-quality source produces a better result.
- Audio: AAC, 192kbps or above, 48kHz.
Upload the highest quality file you can. Do not upload a version that's already been compressed once — re-encoding compressed footage compounds artefacts badly.
Never screen-record or re-download to repost. Cross-posting by downloading from one platform and uploading to another stacks compression and adds watermarks that platforms actively deprioritise. Export fresh from your project each time.
A workflow that scales
If you're producing volume, the editing time is in setup, not creativity:
- A project template with safe-zone guides, caption style, and export presets already configured.
- Caption styles saved as presets so they're never rebuilt.
- An organised audio library with licensing status noted per track.
- A consistent file naming and folder structure, so finding a project from three months ago takes seconds.
- A checklist before export — captions inside the safe zone, audio levels checked, no clipping, correct aspect, watched once on a phone.
That last step catches more problems than everything before it. Always review on the device people will watch on, not on your editing monitor.
Want video production handled?
We produce short-form video for brands — edited, captioned, and exported correctly for each platform. Talk to us about video editing.
