How to decide how much of your film stays on the presenter

Last updated 2026-09-22 · 8 min read

What is the short version?

Serious journalism keeps the person talking on screen; explainers cut away to diagrams instead. The ratio between those two modes is the entire register of the film, and it is worth deciding on purpose rather than discovering afterwards what the footage happened to allow.

1. Tell Vizard Agent the share of the film that stays on the presenter.
2. Say the narration must be followed exactly.
3. Ask for each graphic to be cued to the sentence it explains.

What do you need before you start?

The presenter footage, the narration and a view on the register. Vizard Agent builds the graphics, but it needs to know whether this is cinematic journalism, where the person is the anchor, or an explainer, where they are a voice.

• The presenter footage. As recorded.
• The ratio. Roughly two thirds is a common choice.
• The narration. Fixed, or open to trimming.
• Which facts need graphics. Numbers, flows, structures.
• The language. It affects caption space and spelling.

What do you type into Vizard Agent?

State the split as a percentage and say what the graphics are for. The number sets the rhythm of the film; the list of facts decides where the cutaways land, and together they remove most of the guesswork.

Prompt

Variants worth knowing:

• "About 62 to 65 per cent." A checkable target.
• "Follow the narration exactly." The audio is the specification.
• "Cue each graphic to the sentence." Placement rule, not decoration.

What does Vizard Agent actually do?

Here is the Vizard Agent sequence on a financial journalism piece built to a stated presenter ratio. The graphics turn out to be nearly all of the work, and three separate things go wrong with them before the film is finished.

1. Inspects the presenter footage and the narration before building anything.
2. Checks framing, speech boundaries and the exact wording for the diagrams.
3. Inspects the presenter's framing so captions and window graphics stay clear of them.
4. Maps the narration to the requested graphics without changing any spoken material.
5. Keeps the original voice with no music, since the source is already full-frame 1080p.
6. Checks safe subtitle space at full resolution before setting the caption layout.
7. Finds a spoken inconsistency — a percentage stated before the figure it refers to — and preserves both rather than silently correcting.
8. Sets exact cue points for every number and diagram, and marks unverified material for removal.
9. Builds one consistent graphic system across the film rather than a style per chart.
10. Corrects names and transcription errors in the captions.
11. Re-runs two graphics that failed as separate, uniquely named jobs, and resets the caption preset that sized text too large.
12. Checks each graphic at full resolution, and the subtitles against the presenter's hands and microphone.
13. Removes unnecessary labels and unsupported wording the review found in the generated graphics.

Step seven is the one worth thinking about. The narration contains a slip, and the brief says do not change the narration — so Vizard Agent keeps both statements and flags the inconsistency rather than quietly editing what the presenter said.

Step thirteen is the discipline that keeps a graphic honest. Generated charts arrive with plausible extra labels that nobody said, and every one of those is a claim the film is making on the presenter's behalf.

Step three is why the film looks composed rather than assembled. Knowing where the presenter sits in frame decides where a caption or a window graphic can go without covering them.

What does the result look like?

A film that holds on the presenter for roughly the share you asked for, cuts to a graphic exactly where the narration needs one, carries a single consistent chart style, keeps the spoken track untouched, and has captions that sit clear of the presenter's hands.

When does this not work well?

Some material cannot hold a high presenter ratio. If the footage is a single locked-off take with poor light, or the narration is dense with figures, sitting on the presenter for two thirds of the film is a long time for a viewer.

• One static angle. Two thirds is a long time on one shot.
• Dense figures. The graphics want more room than the ratio allows.
• Poor presenter footage. The anchor has to carry the film.
• Very long films. The ratio needs varying by section.
• Unverified claims. They cannot be put on screen at all.

How do you fix a result that came back wrong?

Say whether the problem is the overall balance or one specific graphic. Those are two different repairs, and Vizard Agent can shift the ratio without rebuilding any of the charts, or replace a single chart without disturbing the balance you approved.

• "Too long on the presenter." The ratio is shifted and cutaways added.
• "That chart says something he did not." The unsupported labels are removed.
• "The captions cover his hands." The layout is reset against the framing.
• "A graphic is missing." That failed job is re-run on its own.

How does Vizard Agent compare to doing it yourself?

By hand the ratio is always an accident. You cut away whenever you happen to have something to cut to, which produces the graphics-heavy middle and the long talking-head ending that every self-edited explainer on the internet has.

Common questions

What ratio should I choose? Two thirds on the presenter reads as journalism; tell Vizard Agent and it works to that.

Can it vary by section? Yes. Tell Vizard Agent which sections need more graphics.

Will it change my narration? Not unless you allow it. Vizard Agent treats the spoken track as fixed.

What if I misspoke? Vizard Agent flags it and asks rather than correcting it for you.

Can it build the charts? Yes, in one consistent system — ask Vizard Agent for the style first.

Will the numbers be right? They come from your narration, and Vizard Agent removes anything it cannot support.

Can it add music? Yes, though for this register Vizard Agent will often recommend none.

How does it place captions? Against the presenter's framing, checked at full resolution.

What if a graphic fails to generate? Vizard Agent re-runs it on its own rather than repeating the whole batch.

Can it use static cards instead of animation? Yes, and for a serious piece they often look better.

Does it check the spelling of names? Yes, and in a second language that is worth asking for explicitly.

Can I see the cue points first? Yes. Ask Vizard Agent for the list before it builds anything.

Can it hit an exact runtime? Yes, if the narration allows — otherwise the narration wins.

How do I check the ratio? Ask Vizard Agent for the measured split in the finished film.