Data Recovery Case File · Memory Cards · The Customer With a Manifest

Thirty-two videos, all named in advance: a corrupted microSD recovered against the owner's own checklist

Most recoveries end with an estimate of completeness; this one ended with arithmetic, because the customer made it possible. His three-year-old, lightly-used 128GB Samsung microSD had failed on two fronts — Windows offering only a format, with the fuller error spelling out that "the volume does not contain a recognised file system," while the action camera it lived in reported "No SD card" at all. He'd attempted nothing: no format, no software. And then the detail that elevated the enquiry: he supplied the card's entire contents as a list — 32 short videos, with filenames, lengths from one to ten minutes, and resolution. A recovery with a manifest is a recovery that can be scored.

Reading the two devices' different verdicts

Windows and the camera appeared to disagree — one saw a volume it couldn't parse, the other denied the card existed — but the split is standard and informative. A computer meeting a card whose filesystem is destroyed still sees the device underneath and reports the truth in its own idiom: unrecognised volume, offer to format. A camera is a narrower reader: firmware built to find exactly the filesystem it expects, and finding gibberish instead, many models round the verdict down to "no card." Two witnesses, one fault: the card's electronics alive, its filing system dead. That's the benign half of the card-failure spectrum — the presenting-but-corrupt state, where imaging is possible and everything turns on what survives beneath the ruined index. His declined format was, as ever, the decision that kept the answer good.

The recovery — and the fragmentation problem the manifest solved

The card was imaged once, write-blocked, and the work ran on the copy. The filesystem proved damaged beyond honest repair, so the videos were recovered by what they are: the image swept for video-stream structures, each recording's chain of segments located and rebound. Here the manifest earned its keep twice. Video files of minutes' length on a well-used card are routinely fragmented — laid down in pieces wherever free space offered — and reassembling pieces into correct, complete recordings is the genuinely hard part of card-video recovery: get it wrong and playback stutters, jumps timelines, or dies mid-file. Knowing there should be exactly 32 recordings, and how long each should run, converted reassembly from estimation into verification — every rebuilt file checked against its expected duration, every join confirmed by continuous playback. And at delivery, the arithmetic the enquiry deserved: 32 of 32, each verified start-to-finish at full length, returned on new media alongside the retired card.

Outcome

A perfect score, honestly earned — and a customer habit this archive would love to see spread. You don't need his level of record-keeping to benefit: even a rough list — "the wedding folder, about 400 photos; the two client videos, roughly an hour each" — gives a recovery something to be measured against, turns "we got most of it" into specifics, and lets priorities aim the work. Card-side, the familiar closing pair: a card that's thrown a filesystem error retires regardless of how a reformat later flatters it, and recordings worth keeping come off the card the same day they're made. He'll do both. The list suggests he was always going to.

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.