How to stop info cards covering the speaker's face

Last updated 2026-09-10 · 8 min read

What is the short version?

A card in a fixed position and a person who moves will collide eventually. It looks like an intermittent bug and it is not — the card never moved, you did. Fixing it means placing the cards against where your head actually goes rather than against one frame.

1. Go to Vizard Agent with the cut.
2. Say the cards cover your face, sometimes.
3. Say where they should sit instead.

What do you need before you start?

The video and a sentence about the collision. The most useful detail is the one people leave out — that it does not happen every time — because that is what tells Vizard Agent this is a position problem rather than a design problem.

• The cut with the cards in it.
• That it is intermittent. Not every card, not every time.
• Where they should go. "Above my head", "top right".
• Whether they may shrink. Usually yes, a little.
• What must stay readable. The card text, at full size.

What do you type into Vizard Agent?

Say "sometimes". A note that says the cards cover your face invites a redesign; a note that says they cover your face from time to time points straight at the fact that the cards are anchored and you are not.

Prompt

Variants worth knowing:

• "From time to time." Identifies the fault as positional.
• "Above my head." A zone defined by you, not by the frame.
• "Shrink them a little." Often needed as well as moving them.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real long talking-head film whose information cards intermittently sat across the presenter. Notice that the proposed position is tested against every card before any of them are moved.

1. Inspects the current card layout and how the cards are built.
2. Measures where your head sits in the frame across the video.
3. Samples every window a card appears in with the proposed zone marked on it.
4. Reviews that zone against your position in each of those windows.
5. Repositions all the cards to the agreed zone.
6. Previews the new placement before rendering.
7. Shrinks the cards and re-anchors them to the same zone.
8. Checks the text is still sharp at full size after shrinking.
9. Renders and screens for any remaining overlap with you.
10. Has the whole film checked for face coverage, not a sample.
11. Splits a card whose list is too long into two shorter ones.
12. Checks the changeover between the two new cards.
13. Confirms the card area is not left black anywhere.

Step three is the step that makes one pass enough. Marking the proposed zone on every moment a card is actually on screen answers "will this work everywhere" before anything is rebuilt, instead of discovering the third card still collides after everything has been rendered.

Step eleven is the honest answer to a card that will not fit. A long list shrunk until it clears your head becomes unreadable, so it gets split into two cards that each fit — which costs a little screen time and keeps the text legible.

What does the result look like?

Cards that sit clear of you for the whole film: small enough to leave room around your head, large enough to read at full size, and any over-long one split by Vizard Agent into two rather than squeezed down until nobody can read it.

Nothing about the design changed. Only where the cards live.

When does this not work well?

Some frames have nowhere for a card to go. A tight close-up, a subject who moves across the whole width, or a vertical crop with very little margin can leave no zone that is clear for the entire time a card is up.

• Tight framing. No clear zone exists.
• A subject who moves a lot. Any fixed zone gets crossed.
• Vertical crops. Far less room at the sides.
• Very long card text. Splitting is the only fix.
• Cards that must sit centre-frame. Then the framing has to change.

How do you fix a result that came back wrong?

Point at the one card that still collides and say roughly when. Vizard Agent keeps the head measurements and the per-card windows it sampled, so a fix moves or splits that single card rather than redesigning the whole set of them.

• "That one still clips my head." Re-anchored against the measurement.
• "Now they are too small to read." Split instead of shrunk further.
• "They feel cramped in the corner." Zone widened and re-tested.
• "One card flashes past." Its timing extended after the split.

How does Vizard Agent compare to doing it yourself?

By hand the card gets positioned on the frame that happens to be showing, and checked on that frame. The collisions then appear at the moments nobody looked at, which is exactly the pattern that reads as an intermittent glitch.

Common questions

Why does it only happen sometimes? Because the card is anchored and you move. Vizard Agent measures both.

Can they go above my head? Yes. Vizard Agent checks that zone is clear whenever a card is up.

Will shrinking make them unreadable? Vizard Agent checks the text at full size and splits rather than over-shrink.

Can it check the whole video? Yes, and it should. Ask Vizard Agent for a face-coverage screen.

What if a card has too much on it? Vizard Agent splits it into two shorter cards.

Will the timing change? Only where a split needs it. Vizard Agent extends that section.

Can the cards move during the video? Yes, if you want them to follow you rather than sit still.

Does the design change? No. Vizard Agent moves and resizes; it does not redesign.

What about vertical versions? Less room, so Vizard Agent re-measures the zone per format.

Can it keep one card centred? Yes. Say which and Vizard Agent leaves that one alone.

Will it check for other overlaps? Yes — captions and lower thirds too, if you ask.

Can I see the zone before it renders? Yes. Vizard Agent marks it on real frames for you.

What if I move out of frame entirely? Then that window is clear, and Vizard Agent will say so.

Why measure rather than just move them up? Because "up" is a direction and your head is a position, and only one of them can be checked against every card in the film.