跳转到主要内容
Workflows 5 min read

One Video, Many Platforms: The Derivative-Copy Strategy

One video, four platforms: one wants vertical, one wants landscape, one caps the duration. You do not edit four videos — you keep one master and derive four spec copies. The right pipeline makes it a ten-minute job.

Master first, derivatives second

Build the master at the most information-complete spec: landscape 1080p, high bitrate, full duration — the single source every copy derives from. Platform copies are derived, not re-edited: same content, different specs, and content fixes happen only on the master.

Before deriving, check three things per platform: aspect direction (vertical platforms need recomposition), duration cap (trim or split if over), size cap (compress if over). Those three answers define the derivation actions.

Three action types for derivatives

Aspect adaptation: horizontal-to-vertical is not a center crop by default — the subject must land inside the vertical safe area, so re-frame with the crop tool and move captions into the bottom safe zone. Resolution unification: copies target the platform recommended resolution (mostly 1080p); mixed-source masters unify first.

Size compression: platforms with hard caps get target-size encodes; the rest get a standard CRF 26 pass. Name each copy by platform and version (douyin_v3) — never "final_2_new" — and keep the master untouched so any copy can be re-derived at any time.

Batch discipline

Finish every picture-changing action (crop, scale) before any compression — compression last means one quality tax. Burn captions before compression; post-compression captions get double-softened. Record each platform parameter set as a checklist: the next video re-runs the same list and specs stay identical.

After publishing, spot-check the first 10 seconds on each platform: did the crop cut key content, are captions hidden under platform UI? Fixes go back to the master, then re-derive.

Frequently asked questions