Data Recovery Case File · Cameras, Drones & Cards · The Half-Gray Photographs

The gray block is where the addressing stops, not always where the photo does: half-images decoded, her two hypotheses put to the test — and most of the missing halves found intact, filed under the wrong addresses

Her enquiry deserves framing as the template it is. The media: an xD picture card from a digital camera — the veteran format of an earlier camera generation. The symptoms, itemised like an engineer's ticket: "most of the photos stored on the card are incomplete — only the top portion of each image is visible, while the rest appears as a solid gray block"; onset "after transferring or viewing photos"; the card "recognised by both the camera and computer", yet the files defeat standard recovery tools. Her hypotheses, both offered: "file corruption during writing, or a possible physical issue with the card's memory sectors." And the closing requests — recover as many as possible, "ensure no further damage is done to the data" — that mark an owner who understands stakes. Both hypotheses are respectable, and this page's whole purpose is the fork between them: because one means the gray halves are gone, and the other means they are misfiled — present on the card, pointed at by nothing.

The decode: why images die bottom-up, and the fork her hypotheses drew

The anatomy of a half-photo: the common photograph format decodes strictly top-down — a viewer renders from the file's first byte until the data stops making sense, then gives up, painting the remainder as flat gray. So every half-gray image is a message with a precise meaning: good data ran out at exactly this line. What it does not say is why — and that is the fork. If the file was truncated at writing, or its later territory sits on dead memory, the gray is real absence: her "physical sectors" hypothesis, and an honest per-photo loss. But if the card's allocation chains were damaged — the filing system's linked map of which cluster follows which — then the viewer follows the map off a cliff mid-file: the top renders from correctly-addressed clusters, and the "missing" remainder sits intact elsewhere on the card, pointed at by nothing. Her "corruption during writing" hypothesis, near enough — and the recoverable branch. The tell in her own report leaned that way: consistent, clean breaks across most photos at once is the signature of a damaged map, not of media dying photo by photo; and the onset around transfer/viewing fits an interruption catching the filing system mid-update.

Her handling, credited: the systematic symptom list, the twin hypotheses, and above all the request that nothing worsen the data — she had already stopped using the card, declined further tool experiments, and asked for exactly the right thing. On a veteran-format card whose remaining readers and remaining endurance are both finite, that restraint preserved the evidence the re-chaining would need.

The recovery: the map rebuilt, the halves reunited

The card was imaged write-blocked, first and completely — her no-further-damage instruction made structural. On the image, the analysis ran the fork to ground: the allocation chains showed the crossing and orphaning of a filing system caught mid-amendment, while the underlying territory read overwhelmingly clean — the misfiled branch, mostly. Reassembly then did the patient work: each photograph's fragments traced by structure and content, chains rebuilt, the orphaned halves reunited with their tops — and the majority of the collection rendered whole, gray blocks replaced by the bottoms of pictures that had existed all along. The honest remainder was itemised per photo: a minority whose later clusters sat on genuinely degraded memory, delivered as the recoverable portions they were, named in the ledger rather than promised around.

On the bench

xD is nearly bare NAND — the format predates smart onboard controllers, the camera doing the housekeeping — so the card read naturally on the PC-3000 Flash, raw pages captured with ECC correction applied. The fork this page turned on was then settled visually: in the Rusolut VNR, whose bitmap view renders allocation structures as patterns the eye can audit, the FAT chains showed their crossing — files linked into foreign clusters mid-photo, exactly where every image went gray. Re-chaining on the image reunited tops with their misfiled bottoms; the minority sitting on genuinely failed pages was itemised per photo, as the honest ledger required.

The outcome

The collection delivered with a per-photo manifest — whole, largely; partial, honestly, where the media itself had failed — and her two hypotheses returned graded: both plausible, the kinder one mostly true, the distinction settled by analysis rather than assumption. Free assessment, one fixed written figure including VAT, no recovery, no fee. The veteran card retired with honours; its successor format ships in her next camera, and its contents now live in two places, as photographs should.

Our case files are drawn from genuine enquiries received by our laboratory over the past ten years, anonymised to protect client confidentiality. Each one describes the diagnostic and recovery procedure our engineers apply to that fault, using the equipment listed.