How to make a community tribute video from public archive footage

Last updated 2026-08-31 · 8 min read

What is the short version?

An anniversary piece needs two kinds of picture: what it looked like then, which you almost certainly do not have, and what it looks like now, which you do. The first half comes from public archives, and where it comes from matters as much as what it shows.

1. Go to Vizard Agent and upload your recording and any present-day photos.
2. Say which event and which year the tribute marks.
3. Say the arc — where it starts and where it should end.

What do you need before you start?

Your own audio is the spine of this, and everything else hangs off it. Vizard Agent builds the picture around what is being said, so a recorded conversation or a piece of narration gives the montage its shape far better than a shot list would.

• Your recording. A host, a conversation, a piece to camera.
• The event and year. Specific enough to search.
• Present-day photos. What was rebuilt, what stands now.
• The arc. Loss to recovery, usually.
• The end card. What it should leave people with.

What do you type into Vizard Agent?

Say what happened, when, and what the piece is honouring. Vizard Agent uses that to search public archives, so the more precise the event and the date, the better the material it comes back with — and the less chance of imagery from somewhere else entirely.

Prompt

Variants worth knowing:

• Say "public archival only." It changes where it searches.
• Give the recovery. The second half needs your own photos.
• Ask for facts on screen. Named, dated, checkable.

What does Vizard Agent actually do?

Here is the order Vizard Agent worked in on a real ten-year community tribute. Notice steps nine and ten: it lists what the archive holds and reads the descriptions before downloading anything, because an archival photograph without provenance is not usable in a published piece.

1. Inspects your clip, samples its frames and transcribes the audio.
2. Researches the event, the location and where archival photographs are held.
3. Lists the archival photographs available on public repositories.
4. Reads each photograph's description before selecting any.
5. Looks for present-day imagery of the same places.
6. Fetches candidates for review and builds a contact sheet of them.
7. Downloads the selected archival set through the proper downloader.
8. Views your uploaded present-day photos.
9. Searches the stock library and news archives for authentic footage of the event.
10. Generates a reflective piano underscore and measures the layout.
11. Builds animated captions from the transcript of your recording.
12. Renders the montage, extracts frames and reviews them, then re-renders with corrections.
13. Searches again for more authentic footage and re-renders using only that.
14. Rebuilds the mix with the music ducking under the dialogue, verifies loudness and checks the waveform.

Step thirteen is worth dwelling on. The first render used some material that was close but not right, and Vizard Agent went back to look for genuinely authentic footage rather than accepting an approximation — which is the correct instinct when the subject is a real event that people lived through.

What does the result look like?

In the run this page is written from, the host's recorded clip ran 62.4 seconds and the finished tribute extended to about 66 seconds in vertical 1080x1920 to allow an end card. The full dialogue stayed intact as the narration spine.

The visual arc was built deliberately: desaturated archival imagery for the disaster, warming into present-day footage of what was rebuilt. Facts appeared as lower thirds, and the music sat under the speech with proper ducking rather than fighting it.

When does this not work well?

A tribute is a piece about real people and a real event, and that constrains it in ways an advertisement is not constrained. Vizard Agent will tell you what it could not verify instead of using something that merely looks right.

• Licensing is yours to confirm. Public archives have terms.
• Facts on screen must be checked. Figures and dates especially.
• Small local events. The archive may hold nothing.
• Generic stock reads as generic. Viewers who were there notice.
• Tone is not automatic. Say what the piece should feel like.

How do you fix a result that came back wrong?

Say which image or which fact is wrong. Vizard Agent keeps the sourced material, the transcript, the captions and the mix settings, so a replacement image or a corrected figure changes one part rather than triggering a rebuild.

• "That photo is not from here." Replaced from the reviewed set.
• "The music covers the speech." Ducking increased under the dialogue.
• "That figure is wrong." The lower third is corrected and re-rendered.

How does Vizard Agent compare to doing it yourself?

By hand this means hours in archive catalogues reading rights statements, then a montage assembled by eye, then a mix where the music keeps swallowing the narration. Vizard Agent does the searching and the reading, and mixes to a measured loudness rather than by ear.

Common questions

Where does the old footage come from? Public archives and stock libraries. Vizard Agent reads each item's description before it uses anything.

Do I need my own photos? For the present-day half, yes. Vizard Agent uses your photos for what stands now.

Can it check the facts? It researches them and shows its sources. Vizard Agent expects you to confirm anything going on screen.

Who owns the archival images? Not you, and not us. Vizard Agent tells you where each item came from so you can check the terms.

Can it write the narration? It can, but a recorded voice from the community is better. Vizard Agent builds the captions from whatever you record.

How long should it be? Around a minute for a feed. Vizard Agent extends slightly for an end card.

Does the music matter? Enormously. Vizard Agent generates a reflective underscore rather than using something with a beat.

Will the speech stay audible? Yes. Vizard Agent ducks the music under the dialogue and verifies the loudness afterwards.

Can it add captions? Yes. Vizard Agent builds them from the transcript of your own recording, word for word.

What about the colour treatment? Deliberate. Vizard Agent desaturates the archival half and warms the present-day half so the arc is visible.

What if the archive has nothing? It says so. Vizard Agent reports an empty search rather than substituting stock footage of somewhere else, which is the failure that would actually embarrass you.

Can I use this for a memorial? Yes, and the same cautions apply. Vizard Agent handles the sourcing and the assembly; the judgement about what belongs in it stays with you.

Does it work in landscape? Yes. Vizard Agent builds the same structure at whatever aspect ratio the piece is going to be shown in.