Data Recovery Case File · Cameras, Drones & Cards · Charge Fades
Flash Memory Is a Working Medium, Not an Archive
This enquiry describes a partial failure rather than a total one, and the pattern is the diagnosis. A camcorder card holding around 32GB of family video where the recordings "will not play back consistently. Symptoms are start, stop, pixelly. The card is 64GB and has been in use from 2013. No signs of physical defects, card only used in the camcorder." Nothing is broken and nothing was deleted. The card has simply been holding those recordings for a decade, and flash memory does not hold anything indefinitely.
| Media | 64GB microSDXC card from a camcorder, in continuous use since 2013 — holding approximately 32GB of video; playback intermittent with visible corruption |
| Reported situation | Card in use since 2013 · used only in the originating camcorder · approximately 32GB of family video and photographs held · playback starting and stopping · visible pixelation during playback · no physical damage evident |
| Fault class | Charge decay in stored flash cells over an extended retention period — progressive; error correction margins exhausted in affected blocks |
| Equipment used | Card removed from use immediately · imaged write-blocked with per-block timeouts and multiple read passes (DeepSpar USB Stabilizer 10Gb) · marginal blocks re-read and best-result assembled · video validated by playback and reported by affected duration |
The decode: why stored data fades, and what the stutter actually is
How flash memory holds data: as electrical charge trapped in an insulated cell. There is no magnetism and no mechanical state — just a quantity of charge, and a threshold that separates one stored value from another. That charge leaks. Slowly, through the insulation, over years — faster when warm, faster in cells that have been written many times, and faster in the denser memory used in higher-capacity cards.
What happens as it drifts: nothing, for a long time. Every card corrects errors as a matter of course, and a cell that has drifted slightly still reads correctly because the correction absorbs it. As drift increases, more correction is needed. Eventually a block accumulates more errors than the correction can fix, and it returns data that is wrong rather than data that is missing.
Why that produces exactly his symptom: video is decoded continuously, and a decoder handed corrupted bytes does not stop — it does its best. The pixelation he is watching is the decoder rendering wrong data, and the starting and stopping is it losing synchronisation and recovering. Files that fail completely would be easier to interpret; files that half-work are the signature of degradation rather than damage.
Why "no physical defects" and "only used in the camcorder" both fit: nothing has happened to this card. It has not been abused, dropped, moved between devices or filled repeatedly. It has been doing exactly what it was asked to do, which was to hold data for eleven years — and that is the one thing this medium is not designed for.
The doctrine, and it applies to every card and stick anybody owns: flash memory is a working medium, not an archive. A card or stick left in a drawer holding the only copy of something is quietly degrading the whole time, with no symptom until the day it produces this. Nor does occasionally powering it help — reading does not restore charge. Only rewriting the data does, which in practice means copying it somewhere else.
Why this is urgent rather than merely unfortunate: the process continues. Blocks currently marginal will become unreadable, and every month reduces what can be recovered. The card should come out of use today and be imaged, with marginal blocks read repeatedly — a block that fails one pass frequently succeeds on another, since the outcome sits close to a threshold.
On the bench
The card was removed from use immediately, charge decay being progressive and every further month reducing what can be read. It was imaged write-blocked behind the DeepSpar USB Stabilizer 10Gb with per-block timeouts and multiple read passes — marginal blocks sitting close to the correction threshold and frequently succeeding on a later attempt where they failed on an earlier one. Best results were assembled across passes, and video validated by playback with the affected duration reported rather than a file count.
The outcome
The card taken out of use, imaged across multiple read passes with best results assembled, and playback validated by duration. 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, and the doctrine worth carrying: flash memory stores data as trapped electrical charge, and charge leaks over years. Video that starts, stops and pixelates is a decoder rendering wrong bytes as correction margins run out. Nothing happened to this card except time. Cards and sticks are working media, not archives — and reading one does not refresh it.
Old card whose videos stutter and pixelate
Stop using it and get the contents copied off now, because this gets worse every month. Nothing has happened to your card except time. Flash memory stores data as electrical charge trapped in an insulated cell, and that charge leaks slowly over years — faster when warm, faster in high-capacity cards. For a long while nothing shows, because error correction absorbs the drift. Eventually a block needs more correction than is available and returns wrong data rather than no data, and a video decoder handed wrong bytes doesn't stop, it renders them. That pixelation is literally what you're watching. The lesson for everything else you own: cards and sticks are working media, not archives, and simply reading one occasionally doesn't refresh it — only copying it elsewhere does.
Stop using it today — call Newcastle Data Recovery on 0191 406 1051; imaged write-blocked across multiple read passes with marginal blocks re-read, best results assembled and playback validated by duration.
Request a quote online →
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.