# I DID NOT TAKE THIS PHOTO

> JPEG artefacts cluster exactly where the eye is already looking for something in the dark. A compression flaw became a horror language completely by accident.

August 18, 2026 · https://justog.club/stories/creepypasta-visual-language

An empty hallway, a doorway at the far end. Blown out by a flash that was too close for the room,
blocky compression artefacts smeared through the dark corners.

`I DID NOT TAKE THIS PHOTO`, small, underneath.

Nothing in the picture is doing the work. The sentence is doing all of it.

## Why does a bad amateur photo scare more than a composed one?

Because a composed shot announces itself as fiction, and a bad snapshot doesn't.

Even lighting, deliberate framing and a clean exposure are all signals of a photographer at work,
which means a viewer's guard is already up — somebody made this, somebody chose this angle. A flash
photo taken too close, at a bad angle, with the exposure blown out in the middle of the frame reads
as an accident: nobody composed it, somebody just pointed a phone and pressed the button, possibly
without meaning to. The technical failure is what makes it look unstaged.

## Why do the artefacts cluster in the dark?

Because that's exactly where the compression algorithm has the least information to work with, and
exactly where the eye is already looking hardest.

JPEG compression is harshest in low-detail, low-light regions — the dark corners of a badly exposed
photo, block artefacts and colour banding gather there because there's less original data to
preserve. That happens to be precisely where a viewer scanning the image for something wrong is
already staring. The format degrades in the same place the fear was already pointed, and it did that
by accident, not by design — nobody engineered JPEG to be frightening in the dark.

## Where does the actual horror live — the image or the caption?

The caption. The photo is only ever evidence for a claim somebody else is making.

`I DID NOT TAKE THIS PHOTO` doesn't describe what's in frame. It asserts a fact about the photo's
origin that the image itself can't confirm or deny — and that gap between the claim and the
unremarkable-looking picture underneath it is the entire narrative device. Strip the caption and the
photo is just a badly lit hallway. The one line is doing more work than the whole composition.

## Does the rights question work differently for a shared tradition like this?

Slightly — because creepypasta lives on wikis with their own publishing terms, not just in a single
author's hands.

An individual invented photo and caption are entirely yours to make, the same as any other
document you write yourself. But the wider creepypasta tradition runs on shared-universe wikis that publish under
specific licences, which is a different kind of permission question than a single artwork or a
single name — the format itself belongs to a community with rules about reuse.
[The detail lives in the containment-form design](/stories/scp-style-documentation), where
that licence actually determines what you can and can't do with a specific published entry.

## Does the format work outside a hallway?

Any mundane setting works the same way, provided the photo looks accidental and the caption makes a
claim the picture can't back up. A car park caught mid-frame, the underside of a staircase, the
inside of a cupboard door left open. The location matters less than the two conditions holding: bad
exposure, and a sentence doing the actual claiming.

## Does it print well?

Yes, and the compression artefacts translate to fabric better than a clean photograph would.

Block artefacts are already flat rectangular shapes with hard edges — the compression has done the
work of turning a photograph into something printable, since [fine photographic detail below the
weave's floor fills in on fabric anyway](/stories/bold-shapes-survive-fabric). The blown-out
flash area holds as a near-white block; the artefact clusters hold as flat grey-black shapes. Off-white
or pale grey for the shirt, one or two inks for the photo, keeps the caption legible in plain black
type beneath it.

## How do I get one made?

Describe the photo at [JustOG](/) — what the flash caught, how bad the compression is, and what the
caption claims.
Pick a direction, drag the crop frame, see it composited on the real garment, and it's made to
order and shipped.

Designs other people have published are in the [shop](/shop).

Somebody uploaded it in the middle of the night, with no other post before or since.