How to fix the empty boxes and stray lines in a news style video

Last updated 2026-09-15 · 8 min read

Point Vizard Agent at each empty element by where it is on screen, and say whether you want it filled or gone. It gives the panels real labels, ties the ticker text to what the narration is actually saying at that moment, removes the rules and placeholder tags, and checks each frame afterwards at full size.

What is the short version?

A news look is mostly furniture: a header bar, a lower third, a side panel, a ticker, a footage window. Generate one and you get the shapes, but a shape with nothing in it reads as a mistake rather than as broadcast design.

Tell Vizard Agent where each empty element is.
Say whether to fill it or remove it.
Ask for the ticker to follow the script.

What do you need before you start?

The video and a willingness to be specific about position. Vizard Agent cannot act on "it looks unfinished", but it can act on "the rectangle under the presenter at twelve seconds", and a timecode with a location is the whole instruction.

The video. The version with the empty furniture.
Where each thing is. Timecode and position.
What each panel should say. Or that it should go.
The script. So the ticker can follow it.
What the footage actually shows. For the labels.

What do you type into Vizard Agent?

Go item by item. A single instruction to "make it look properly finished" gives Vizard Agent no way to know which of the empty shapes you want filled and which you want removed, and those are opposite fixes.

Prompt

At twelve seconds there is a red horizontal line under the header that goes nowhere — remove it from the opening, the header, the lower band and the outro. The rectangle under the presenter is empty: put a moving ticker in it whose text follows what the narration is saying at that point. And there is a placeholder label sitting over the footage window — cover it without covering the footage.

Variants worth knowing:

"Remove it from the opening and the outro too." These repeat.
"Follows what the narration is saying." A ticker with real content.
"Without covering the footage." The constraint on every patch.

What does Vizard Agent actually do?

Here is the sequence on a real news-style report that came back full of empty furniture. Much of it is iterative, because each patch has to be checked at full size — a cover that works at playback size can be a visible grey block up close.

Zooms into the frame you named to see the element exactly.
Decides what to move, shorten or blank in that area.
Covers the stale text and empties the panel that is no longer relevant.
Shrinks the cover so only the old text goes and the presenter stays clear.
Gives the panel a short, relevant label so it reads as real graphics.
Removes the stray rule from the opening, header, band and outro.
Replaces the opening background so no old graphics show inside the letters.
Adds a lower third and a moving ticker synced to the narration's content.
Covers the placeholder tag on each footage window without hiding the image.
Drops a picture-in-picture that had no material and keeps the presenter.

Step four is the correction that always follows step three. The first cover is too big, because covering a thing generously is the safe first guess, and then the presenter is half behind a grey rectangle.

Step eight is what separates a finished news graphic from a template. A ticker with lorem text is still empty furniture; one carrying the actual story at the moment it is being told is the thing the format exists for.

Step ten is worth asking for explicitly. An empty picture-in-picture window is the most obviously unfinished element on a news screen, and removing it and staying on the presenter is almost always better than filling it with something unrelated.

What does the result look like?

Broadcast furniture that carries content. Every panel says something, the ticker tracks the story, the stray rules are gone, and Vizard Agent shows you full-size frames of each place it touched so you can check the patches rather than trust them.

When does this not work well?

A patch is a cover, not a rebuild. Where an element sits over moving footage or over the presenter, covering it means covering part of the picture, and past a certain size the honest fix is to regenerate the graphic rather than to patch it.

Elements over moving footage. A static cover will show.
Furniture behind the presenter. Covering it covers them.
Very large empty areas. Rebuild rather than patch.
Animated furniture. The cover has to move too.
Text baked into a transition. Hardest case; may need regenerating.

How do you fix a result that came back wrong?

Give a timecode and say what you can see there. Vizard Agent inspects the frame at full resolution rather than at playback size, so a report like "there is still a grey block at 1:17" is something it can go and look at directly, rather than something it has to guess the location of.

"The cover is visible." It is shrunk and matched to the background.
"It covers the presenter." The patch is moved or reduced.
"The line is still in the outro." The same fix is applied there too.
"The ticker overlaps the badge." The ticker is re-laid out.

How does Vizard Agent compare to doing it yourself?

By hand you patch the one you noticed, and three more appear on the next watch — the same rule in the outro, the same placeholder on the fourth footage window. Each is a two-minute job and there are nine of them.

	By hand	Vizard Agent
Finding them	One per watch	Surveyed across the whole video
Repeated furniture	Fixed where you spotted it	Fixed everywhere it appears
Patch size	Generous, then visible	Shrunk and checked at full size
Empty panels	Left empty or hidden	Filled with real content
The ticker	Placeholder text	Tied to the narration

Common questions

Why are the boxes empty in the first place? A generated news look produces the layout. Vizard Agent fills it with your content.

Can it write the ticker text? Yes, from your script. Vizard Agent keeps it in step with what is being said.

Will the patches be visible? Not if they are checked. Vizard Agent inspects each one at full resolution.

What if the element moves? Then the cover has to move. Tell Vizard Agent and it tracks rather than patches.

Can it remove the picture-in-picture entirely? Yes, and Vizard Agent keeps the presenter full frame instead.

What about labels over the footage? Vizard Agent covers those without covering the image underneath.

Does the same fix apply to the outro? It should. Ask Vizard Agent to sweep the whole video for that element.

Can it match the panel to the house style? Give Vizard Agent an example and it will follow the colours and type.

What if the footage does not match the story? That is a content problem, not a graphics one. Raise it with Vizard Agent separately.

Will the presenter stay clear? Yes. Vizard Agent checks the presenter's position after every patch.

Can I keep some panels empty? Yes, if they are meant to be. Say which, and Vizard Agent leaves them.

How many rounds does this take? Usually two or three, because each patch is verified before the next.

Does the audio change? No. Vizard Agent works on the picture unless you ask otherwise.

Should I just regenerate the graphics? Sometimes cheaper. Vizard Agent will tell you when patching is not the better route.