Data Recovery Case File · Cameras, Drones & Cards · Why the App Could Not Help

Files Present, Pictures Missing, and an App That Did Not Help

Her enquiry described a symptom precisely and then a reasonable attempt that failed. Pictures downloaded from a microSD card onto a tablet and then a laptop: "they are there but won't open. The JPEG file is there but they won't show the pictures. I've tried a data recovery app but that hasn't worked either, and if I put the card back in my camera it says no playback." Three devices agreeing places the fault in the card. But the more useful thing to explain is why the app was never going to help — because recovery software does one job well and a completely different job not at all, and knowing which is which saves people from buying several of them in succession.

The decode: what is actually happening, and what an app can and cannot do

Why the files exist and the pictures do not: a card holds two different things in two different places. The directory — small, concentrated, holding filenames, sizes and dates — and the image content, spread across the rest of the card. Those are separate reads. When the directory region reads cleanly and the content regions do not, you get exactly what she describes: a perfect list of JPEGs with correct names and plausible sizes, all of which display nothing. The card is not lying; it is returning what it can read.

What a recovery app is for: finding and extracting files whose listing has been lost — deleted entries, a formatted card, a damaged directory. It scans, identifies files by their internal signatures or from surviving records, and writes them out. That is a genuinely useful job and these tools do it well.

Why it could not help her: because her listing is not the problem. Her files are already visible and already copy across. What is failing is the reading of the content — and no recovery application can make a card return data it will not return. Faced with unreadable sectors, an app does what her copy did: it takes what it gets, pads what it cannot read, and writes out a file of the correct size that will not open. So running one produces the same unopenable images by a longer route. That is not a fault in the software; it is the wrong tool for this fault, and buying a second and a third will produce identical results.

What actually changes the outcome: patience with the reads themselves, which no consumer tool offers. Imaging under per-sector timeouts gives each failing sector a strict budget and defers it rather than retrying until something stalls; the healthy expanse is captured first while the card is at its strongest; and the difficult regions are revisited on later passes, where repeated patient reads frequently recover sectors that failed a dozen hurried attempts. The images are then rebuilt from sectors genuinely read rather than from padding.

What to stop doing: copying, and running further software. Every attempt is a prolonged session of a struggling card being asked repeatedly for exactly the regions it cannot return, and that is spending margin for no benefit. The camera's no playback message is the same finding from a device with a simpler vocabulary — and its next offer will be to format.

On the bench

Copying and scanning stopped, which was the most valuable single step available. The card was imaged write-blocked behind the DeepSpar USB Stabilizer 10Gb with per-sector timeouts enforced — the decisive difference from what her devices and the app had been doing — and the sweep prioritised so the readable expanse was banked first. The content regions that had returned nothing were then revisited on later, patient passes, recovering sectors no ordinary read had obtained. The photographs were rebuilt from those recovered sectors, validated by rendering at full size rather than counted, and delivered ordered by embedded capture date.

The outcome

The card imaged under timeout control with the content regions re-read, and the photographs rebuilt and validated by rendering. Free assessment, one fixed written figure including VAT; where a chip has to be removed, 50% of parts and labour is payable upfront with the balance only on success — otherwise no recovery, no fee. The decode, for anyone whose files exist and show nothing: the directory and the content are separate reads from different areas, so a perfect file list over blank images means the small concentrated region reads and the large one does not; recovery apps find files whose listing has been lost, which is a real and useful job and not your problem — your listing is fine, so an app has nothing to add and will pad the unreadable parts exactly as your copy did; what changes the outcome is patient re-reading under timeout control; and stop copying and scanning, because each attempt spends margin.

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.