top of page

The Master Guide / 14 Finishing and delivery · Graphics, captions, versions and delivery

Versions

Every version is its own edit, made from the same media. The approved master is never changed to make one.

Every version is its own edit, made from the same media. The approved master is never changed to make one.

The version table lists every deliverable: duration, aspect, language, captions (burned in or sidecar), music cue, loudness target, codec, file name and status (planned, rendering, checked, sent, approved). Ask each TV channel for its own technical sheet at the start of finishing; it overrides every default in this section.

Each version is its own sequence, cut list and timeline (decided), from the same media.

Ratios (decided):

  • The main deliverables, 16:9 and 9:16, were generated natively, each designed and blocked for its own frame.

  • 1:1 from either master and 4:5 from the 9:16 master are crops, at the offsets the plan set (see Planning, budgets and takes → The asset list and versions). A crop is fine only when the action, faces, hands and product stay inside the new frame on every frame.

  • A version that would need enlarging is refused and generated natively: new takes, priced, made in a production batch you review at once.

  • Higgsfield's Reframe comes last, as a new take. Use it only once Seedance 2.5 is confirmed as the model behind it and several creators show it works, or when you choose it for that shot (see Tools at a glance → Each job, its tool, its label and its chapter).

Online versions: a feed edit, or the same story, shorter (decided). You ask the client which one at the brief (see The brief and research → Opportunities and meeting questions).

The same story, shorter, uses the master's shots with new windows.

A feed edit is rebuilt for the feed. The treatment writes each version's first 5 s, and the client approves them. For a vertical version, those opening shots are generated natively in 9:16 and priced, and the rest is cropped from the 16:9 master at the plan's offsets.

In the plan, each native opening shot is its own generation, with the ratio after its ID (S01_9x16) and its price in the estimate. The cropped rest gets a crop row per shot (see Planning, budgets and takes → The asset list and versions).

The platforms' guidance for those first 5 s [official guidance]:

  • brand and message inside 5 s;

  • more than two shots in them;

  • a tight face or product to open.

A 6-second version gets its own one-line idea in the treatment, never a trim of the 30 s film.

Cut-downs (30 to 15 to 6): each a new edit from the approved takes, following the treatment's priority beats, at exact length, with the music edited on bar lines. Each is approved like the master.

Languages: the picture is shared. Each language gets its own voice-over and dialogue, lip-synced takes for its speaking shots, captions, graphics renders and sequence.

Graphics per version: one render per aspect, language and frame rate, from the same grid and timings, with positions re-set from that aspect's grid; never scale a 16:9 layout into 9:16. A longer line in another language gets a longer hold, approved with the cut.

The script is a core element of the brief, so use Higgsfield's Ad Multiplier for a client version only with your explicit approval.

Sources

bottom of page