How to get shots of your presenter that you never filmed

Last updated 2026-09-22 · 8 min read

What is the short version?

A title sequence is a small film about a show, made almost entirely of shots the show itself does not contain. Everything in it still has to look like the same programme, and the only reference Vizard Agent has for that is the episodes you have already filmed.

1. Give Vizard Agent an episode and the credits as they should read.
2. Say what shots the sequence should contain.
3. Say the length and where the logo lands.

What do you need before you start?

An episode, the credit list and the logo file. Vizard Agent takes the set, the colours and the host's own appearance directly from that footage, so the more representative the episode you give it, the closer the finished sequence will look to the programme itself.

• An episode. Or several, for more reference.
• The credits. Names and roles, spelled exactly.
• The tagline. If there is one.
• The logo. A file, or the show's typeface.
• The length. Twenty to thirty seconds.

What do you type into Vizard Agent?

Describe the shots you want, not just the credits. A credits list produces cards; a sentence like "her turning to face the camera, smiling, and one of her laughing" produces a sequence that is about the person who presents the show.

Prompt

Variants worth knowing:

• "From the shoulders up, turning to camera." A shot, described.
• "The logo coming up at the end." Where the sequence lands.
• "25 seconds." Title sequences are short, and the length shapes everything.

What does Vizard Agent actually do?

Here is the Vizard Agent sequence on a title sequence built for an existing talk show. Note how much of the work is spent looking at the programme itself before a single frame of new material is generated from it.

1. Checks the project files and samples frames from the show footage.
2. Looks closely at the host, the logo on set and the network's style.
3. Checks the available generators, effect treatments and fonts.
4. Checks generation prices before committing to shots.
5. Scans the first ten minutes for clean shots of the host's face.
6. Pulls closer frames and picks the cleanest reference images.
7. Prepares reference crops and uploads them.
8. Generates the two host shots, two backgrounds and the theme music in one pass.
9. Reviews the generated clips before building anything around them.
10. Rebuilds the show's logo in its own typeface with its glowing ring.
11. Generates the animated credit and tagline overlays, and previews them.

Step five is the step that makes the host look like herself. A generated shot built from a description produces a plausible presenter; one built from clean frames of the actual host produces her, in her own lighting, in the show's colours.

Step ten matters because a logo is not an image file to a viewer, it is a shape they recognise. Rebuilding it in the show's typeface means it can animate properly rather than being a flat picture that scales.

Step four is worth copying on any generative job. Knowing what the shots cost before generating four of them turns an open-ended bill into a decision.

What does the result look like?

Twenty-five seconds that belong to the programme: shots of the host that look like the host, backgrounds in the show's palette, animated credits with the names spelled as you gave them, a tagline, and the logo arriving at the end with its own motion.

When does this not work well?

Some shows have nothing usable to build from. If the footage is a low-resolution stream capture, if the host is never in clean light, or if the brand has no settled typeface or colours, then Vizard Agent has to invent a look rather than express the one you already have.

• Low-resolution source. No usable reference frames.
• No clean shots of the host. Everything is wide or moving.
• No brand to follow. The sequence becomes a guess.
• Very long credit lists. Twenty-five seconds fills quickly.
• Rights-restricted music. A theme has to be generated or licensed.

How do you fix a result that came back wrong?

Say which element is wrong. A title sequence is made of separable parts — the host shots, the backgrounds, the credits, the logo and the music — and Vizard Agent can rebuild any one of them without touching the rest of the sequence.

• "That does not look like her." New reference frames are pulled and the shot regenerated.
• "The logo is wrong." It is rebuilt from the typeface rather than scaled.
• "A name is misspelled." That card is regenerated with the correct spelling.
• "It feels too long." The sequence is retimed, usually by shortening the backgrounds.

How does Vizard Agent compare to doing it yourself?

By hand a title sequence is a specialist job, which is why most independent shows do not have one — they have a logo on a black card. The part that stops people is exactly the part here: you need shots of your host that you never filmed.

Common questions

Do I need to film anything new? No. Vizard Agent works from the episodes you already have.

Will the host look like themselves? That is what the reference frames are for — ask Vizard Agent to show them first.

Can it use our theme music? Yes. Supply it, or Vizard Agent generates a theme to fit.

How long should titles be? Twenty to thirty seconds. Vizard Agent builds to whatever you set.

Can we reuse it every episode? Yes. Vizard Agent keeps the elements so later changes are cheap.

What if the cast changes? Vizard Agent regenerates just the affected card or shot.

Can it make a shorter version? Yes. Vizard Agent cuts a ten-second version from the same elements.

Will the credits be readable? Vizard Agent checks them against the backgrounds and adds a scrim if needed.

Can it match another show's style? It can take direction from a reference, though the look should be yours.

What about an end-credits sequence? Yes, built from the same design so the pair match.

Can it animate our existing logo file? Yes, though rebuilding it in the typeface animates better.

How much does generating the shots cost? Ask Vizard Agent first — it checks the prices before generating.

Can we get the elements separately? Yes, so an editor can rebuild the sequence later.

What if we do not have a tagline? Then the sequence runs credits into logo, which is often cleaner.