What to show in a documentary when no footage exists
History, true crime and disaster channels tell stories with no camera anywhere near them. Nobody filmed most of the moments they describe, and nobody ever will. The question isn't whether to find footage of the exact moment - you won't - it's what to put on screen instead, in an order that keeps the story honest and the viewer oriented.
What do you show in a YouTube documentary when there is no footage?
Show what the line names, one step removed: the place, the date, the object, the document. Use real photos and period documents first, maps for journeys, era details for objects, and labeled recreations for actions. When photos of a person run out, change the kind of shot instead of repeating the face.
That's the short version. The rest of this page is the order to look in, why the order matters more than any single source, and what it looks like measured on two real channels.
What order should you look for material in?
Working down a fixed list, instead of grabbing whatever comes up first in a search, is what keeps a video from feeling like a slideshow of unrelated stock. In roughly this order:
- Event archive. A photo, newsreel frame or document from the actual event, if one exists. This is rarer than people expect, especially for anything that happened quietly.
- Place and era archive. The same city, building or type of vehicle, photographed in the right period, even if it isn't the exact moment. A street in the right city from the right decade is closer to the truth than a modern stock clip of an unrelated street.
- Dated documents and newspapers. Archival documents and newspapers go on screen as what they are, with their real dates, so a headline shown for a moment is a headline from that time, not a prop.
- Maps. When the story moves from one city to another, or the geography itself is part of the story (a route, a border, a flight path), an animated map shows it instead of asking the viewer to picture it from the narration alone.
- Era-matched objects. A radio of the right period, a notebook, a suitcase, a desk. These carry a scene without claiming to be the actual object owned by the actual person, which is an honest kind of detail rather than a guess dressed up as fact.
- Labeled recreation, last. When nothing above fits the line, a recreation can stand in for an action - but it should read as a recreation, not be passed off as the real thing.
Each step down this list trades a little documentary weight for coverage. Skipping straight to recreation because it's fast and available is how a history channel starts to look like a thinner genre. For the legal side of pulling from real archives - what's licensed to reuse and what to credit - see archival footage and copyright in YouTube documentaries.
How do you pick a picture for what the voice says in that second?
The test isn't "who is talking" - it's what the line names right now. If the narration says a name, a picture of that person is the obvious choice, but names are a small share of what a script actually says. Most of a documentary script is describing a place, a time of day, an object or an action, and those deserve just as much attention as the names do.
If the line says a city at night, the shot should be that city, at night - not a daytime stock photo of a generic skyline. If the line mentions a door, there should be a door on screen, not twenty seconds of a face while the door goes unseen. If the line names an object - a notebook, a radio, a suitcase - that object, or a period-correct equivalent, belongs in the frame. This is the same rule a spy-history series I edit follows: choose the picture for the place, the time of day, the object and the action the line names, before reaching for a portrait. More on how that plays out across a full episode in the Spy Show documentary series case study.
What do you do when you run out of photos of a person?
Every real person's photographic record runs out eventually, often after two or three usable images. The instinct is to reach for a stock photo of someone who looks roughly right and use it as a stand-in. Don't. A stranger's face next to a real person's name tells the viewer something false about who that person was - the kind of thing an attentive viewer notices, and stops trusting the rest of the video for.
Change the type of shot instead. Move to an object tied to that person - a jacket over a chair, a case on a desk, hands on a table. Move to the room, the street, the vehicle, or a document with their name on it. None of these claim to be a likeness, so none of them can misrepresent one, and a run of varied detail shots reads as more careful than the same portrait recycled every few seconds. The general workflow this fits into, from script and voiceover to a finished upload, is in how a faceless YouTube video gets edited.
This is also where labeled recreation earns its place lower on the list above: a recreation that's honest about being one can carry an action a photo never will, without asking a stranger's face to do a real person's job. What it shouldn't do is stand in as generic, unoriginal filler dressed up as insight - that's closer to the repetitive, templated content YouTube's inauthentic content policy targets, a separate problem from sourcing but worth keeping in mind before leaning on recreation too heavily.
How much of a real documentary is actually archive footage?
Less than most people assume, and it's worth measuring rather than guessing. Two channels, measured frame by frame and asset by asset:
| Channel | Material mix |
|---|---|
| The Interpreter (client) | Archival photos 66% · video b-roll 24% · drawn cards 3.6% · maps and diagrams 3% · date cards 1.8% |
| Philip Thompson (reference) | Real moving footage about 20.6% of runtime; archival footage alone about 12.5%. The rest is still images and text |
On the client channel, drawn cards, maps and date cards together are still under 9% of the material, but they cover exactly the moments - geography, chronology - that a photo can't. On the reference channel, footage that actually moves is roughly a fifth of the runtime; the rest is stills doing the carrying, which is normal for this genre, not a sign of a thin archive. For how these two channels pace those images once on screen, see the frame-by-frame numbers in how often should you cut in a YouTube video. I measured The Interpreter's mix on the channel's own videos before cutting a Tenerife 1977 deep dive for it; the breakdown is in the Interpreter Tenerife documentary case study.
How do you license each piece?
Every asset in the edit should be traceable back to where it came from: source page, author, license, and whether it needs a credit in the description. That's a plain per-asset log, not a memory to keep in your head, and it's the log a copyright claim or a licensing question gets answered from later. The legal detail is covered in archival footage and copyright in YouTube documentaries. For what this looks like end to end on a documentary channel, see documentary YouTube editor for history, disaster and true crime deep dives.
No-footage documentary FAQ
What do you show in a YouTube documentary when there is no footage?
Show what the line names, one step removed: the place, the date, the object, the document. Use real photos and period documents first, maps for journeys, era details for objects, and labeled recreations for actions. When photos of a person run out, change the kind of shot instead of repeating the face.
Is it okay to use stock footage in a history documentary?
For generic scenes only, such as a hallway, a radio dial or a night street, and never labeled or implied as archive. Real archive comes first for anything the narration names specifically; stock fills the gaps around it.
How do you show a person when no photo of them survives?
Change the type of shot rather than putting a stranger's face on screen. Move to an object tied to them, a document, the room, or a wide shot that doesn't ask a face to stand in for someone real.
How much of a real documentary is actual archive footage?
Less than most people assume. Measured on The Interpreter, a client channel, archival photos make up 66% of the material and moving video b-roll 24%; on the reference channel Philip Thompson, footage that actually moves is about 20.6% of runtime and archival footage alone about 12.5%. The rest is stills, cards, maps and diagrams.