Most camera data loss is logical damage, not physical destruction. A delete, an in-camera format, a card pulled mid-write, or a corrupted directory entry breaks the map to your files without touching the files themselves, so the footage is often still on the card right now. It also explains why the question why do camera files get corrupted has a dozen different answers that produce the same message on the camera’s rear screen yet behave very differently in recovery.
Find the row below that matches what you saw, note its recoverability band, then confirm the diagnosis with the signal-based triage before writing anything else to the card.
Logical damage and physical failure look identical on the camera
A memory card is three layers stacked on top of each other: the NAND flash chips that hold charge, a controller that runs the Flash Translation Layer (FTL), and a file system that the camera can actually read. The camera only ever sees the third layer, and it is not capable of telling you anything about the first two. When the file system is broken but the chips are healthy, the camera reports exactly what it reports when the card is genuinely dead: no images, cannot read this card, format it now. That is why the on-screen message is close to worthless as a diagnostic.
Logical damage means the directory entries are damaged or gone. A directory entry is the record that says “DSC_0142.JPG begins at cluster 4096 and is 8,214,528 bytes long.” When that record is lost, the payload sitting in the flash cells is completely unchanged. The bytes are still there, in the same blocks, in the same order, waiting to be re-indexed.
Physical failure means the payload itself can no longer be read reliably. Cells that have lost their charge, a controller that cannot build a valid block map, a substrate that has cracked, or oxidized contact pads. No scanning utility repairs that layer, because the problem is below the abstraction the utility operates on. This is the layer where a data recovery laboratory with chip-off equipment is the only realistic route. If you want the background on how the first layer is actually organized, the anatomy of how cameras store photos is worth ten minutes before you start triage.
The practical consequence is short and useful: a card that will not mount is not automatically a dead card. Check whether the computer can see the device as a raw physical drive with an unreadable or unrecognized file system. That is a logical symptom, and it is the good kind.
Every cause of camera data loss, ranked by frequency and recoverability
The table below is the core of this page. Frequency estimates are drawn from the pattern of consumer reports and service-desk tickets rather than from a single controlled study, but the relative ordering is stable. “Recoverability” assumes you stop writing to the card immediately and do not run a repair utility that rewrites the file system.
| Cause | How often | First signal you notice | Recoverability |
|---|---|---|---|
| Accidental in-camera delete | Very common | File count drops; thumbnail turns into a placeholder icon | High, near-complete |
| In-camera quick format | Very common | Camera asks to initialize or reads as empty | High, near-complete if untouched |
| Card pulled or power cut mid-write | Common | Last clip missing, 0 bytes, or will not play | High for earlier clips; variable for the in-progress clip |
| File system or directory corruption | Common | “Cannot read card”, folder appears empty, RAW files vanish | High with a capable scanner |
| Counterfeit or under-capacity card | Common | Recording stops at a repeatable point; old files corrupt first | Medium to low, depends on how much was written past true capacity |
| Interrupted import on the computer | Common | Import stalls; half a folder copied | High, the source card is usually untouched |
| Controller or FTL failure | Less common | Card goes read-only, or keeps disappearing and returning | Low, laboratory territory |
| NAND wear and retention loss | Less common | Random single unreadable files spread over weeks | Low for the worn blocks only |
| Firmware write bugs | Uncommon | Corruption starts right after a firmware update | Medium, usually recoverable |
| Physical damage | Uncommon | Card will not seat or is not detected by any reader | Only via chip-off work at a lab |
| Water exposure and corrosion | Uncommon | Card works, then fails one to three weeks later | Low without a lab; failure is delayed |
| Heat and thermal stress | Uncommon | Camera shuts down on long takes; final clip lost | Medium |
| X-ray, magnets, static discharge, EMP | Rare | No consistent field symptom | Negligible for X-ray; physical handling is the real risk |
The three causes behind most real-world camera data loss
Deletion and in-camera formatting
Formatting in the camera is a quick format, not a secure erase. It writes a fresh file system skeleton, meaning a boot sector, empty allocation tables, and an empty root directory, and then marks the previously used clusters as available. It does not touch the flash cells that hold your footage. Until the camera records new material into those same blocks, the old files are intact, and their filenames, timestamps, and EXIF are usually recoverable as well because the directory area is typically only partially reused. The mechanics are covered in more depth in accidentally deleted photos from camera.
The reason recovery still requires a scanner rather than a simple undelete is that the camera also destroyed the index. A recovery tool works in two passes: first it parses the surviving file system fragments, then it carves by signature for anything the directory no longer describes. The second pass is why footage can often be recovered even when the directory is completely gone, and it is also why recovered clip names rarely match the originals. For still images the underlying reason recovery is possible at all is worth understanding separately, since it comes down to how flash memory and file systems interact rather than to anything the tool does: why deleted photos can be recovered.
Unsafe removal and power loss during a write
A camera does not write a file in one atomic operation. It buffers, allocates clusters, writes the payload, and only then updates the directory. Interrupt the sequence between payload and directory, and you get clusters that contain a complete clip with no entry pointing to them, which the camera reports as a file that disappeared. Interrupt during the directory update itself and you get a partial entry, which is the classic zero-byte file or the thumbnail that renders but refuses to open.
The clips finished before the interruption are normally untouched, because their directory entries were committed earlier. The in-progress clip is the variable one, and its outcome depends on which of the four stages was running. The specific failure modes of an interrupted write are collected in power loss during recording, and the user-behavior version of the same problem is in safe to remove memory card.
File system corruption
exFAT and FAT32, the two file systems that cover almost every consumer camera card, both keep redundant copies of the allocation table and a root directory. Damage any of those structures and the entire volume becomes unreadable even though sixty gigabytes of footage is sitting there undamaged. The triggers are mundane: the card was pulled before the write cache flushed, the battery dropped at the wrong moment, the card was moved between two devices that disagree about the volume’s state, or the card sat in a drawer for a year and then Windows offered to “repair” it with a tool that rewrote the very structures a recovery scan needed.
The common causes of SD card corruption are worth reading as a group, because the mitigation is identical for all of them: never let the operating system repair a card that holds the only copy of something you care about. Mount it read-only, image it, and work from the image.
Common causes that are usually misdiagnosed
Counterfeit and under-capacity cards
The failure pattern here is backwards from what everyone expects, which is why it gets misdiagnosed so often. A counterfeit 512 GB card built on a 32 GB chip does not simply stop working when you cross 32 GB. Its controller firmware has been altered to wrap around and overwrite from the start of the address space. The result is that your most recent clip is perfectly fine and your oldest files are the corrupted ones. Users read that as “the card suddenly corrupted older photos” and blame a bad batch, a bad camera, or a bad reader, then buy another card from the same source. The warning signs, testing procedure, and the reason a capacity test is the only reliable check are set out in fake SD card.
NAND wear and end-of-life behavior
Flash cells tolerate a finite number of program and erase cycles. The controller hides this behind wear leveling and a pool of spare blocks, so the card behaves normally until the spare pool runs dry. At that point the symptom changes from failure to data retention loss: files write successfully, verify successfully, and then turn out to be unreadable three weeks later because the charge in the cells has leaked away. A card that has been rewritten for several years, or that lives in a hot car, is the usual profile. The failure curve and what the remaining spare blocks mean for your data are covered under SD card lifespan.
One structural detail explains why camera recovery succeeds as often as it does. Camera cards and the great majority of card readers never issue a TRIM command, and the cards themselves generally do not implement it. On a modern SSD, deleting a file lets the drive proactively erase the underlying cells, so deleted data can be gone in minutes. On a camera card, a deleted photo can remain physically intact for months. That is not a feature anyone designed for recovery; it is a side effect of the interface, and it is the single biggest reason this class of recovery works at all.
Firmware write bugs
Camera firmware occasionally writes files in a way that violates its own file system conventions, most often by mishandling a directory update when a buffer fills at a specific size, or by producing a container with a broken index. The tell is timing: corruption that begins on the first card written after a firmware update and affects one file type or one recording mode consistently. This is one of the few causes where the manufacturer may already have a fix, and where the affected cards are usually still recoverable because the media itself is healthy. See camera corrupts memory card.
Interrupted imports on the computer
In the overwhelming majority of “the photos vanished during import” cases, the card is fine and the destination was the problem: a hub that dropped, a drive that filled, a cable that moved, or an import utility that reported success on a partial copy. The recovery is trivial as long as the user did not then reformat the card to “clean it up.” If you already did, that reformat is the actual loss event, and the card’s contents are back in the same situation as any quick-formatted card. The failure patterns are catalogued in photos lost during import.
Physical and environmental causes
Physical damage to the card
Cracked substrate, bent or oxidized contacts, a delaminated microSD-to-SD adapter, gold fingers worn through by repeated insertion. If a card is not detected at all in two different readers on two different machines, that is the signal to stop experimenting. Repeated insertion attempts into a reader that is supplying power to a card with a shorted controller can turn a recoverable card into a dead one. The distinctions between damage a lab can work around and damage nothing can fix are in broken SD card data recovery. One frequent false alarm: microSD cards are blamed for adapter failures constantly. Test the card naked in a slot that accepts microSD directly before concluding anything.
Water exposure and delayed corrosion
Water rarely kills flash memory immediately, and this is the counterintuitive part. The realistic sequence is that the card is dried, works perfectly for a week, and then fails because corrosion has been migrating across the controller leads the entire time. Powering a wet card is what converts a survivable event into an unrecoverable one, because it drives electrolytic damage while moisture is still present. The full timeline, including why rice is the wrong tool, is in does airport X-ray damage SD card.
Heat and thermal stress
Long 4K or high-bitrate takes in direct sun push cards past their rated operating temperature. The card throttles writes, and in some cameras the final clip is lost when the camera shuts down before committing its directory entry. Cards that have been repeatedly heat-cycled also show accelerated retention loss later, which is why a card that survived one overheated shoot can still produce mysterious unreadable files months afterward.
X-ray, magnets, static discharge and EMP
Airport X-ray does not erase flash memory. The energy in a baggage or CT scanner is not sufficient to flip the charge state of the cells in a meaningful and organized way, and there is no credible mechanism for a scanner to rewrite your file system. The real risk at airport security is entirely physical: a card loose in a tray, a screener handling a bag roughly, or a static discharge from a dry plastic bin. Magnets are similarly oversold as a threat; a permanent magnet strong enough to matter would have to be held against the card, and it would damage the reader’s mechanics before it damaged the data. Bulk erasure of flash requires a field strength that does not occur in luggage.
Identifying your cause from the signals you saw
Work through these in order, and stop at the first one that matches. The point of the sequence is to separate causes that require different actions, not to produce a perfect diagnosis.
- Did the card ever mount after the loss? If yes, the trouble is logical and recovery odds are good. If no, test in a second reader and a second computer before concluding anything about the card’s health.
- Does the capacity shown by the computer match the label? If it does not, or if the card reports a capacity far below the printed figure, stop and treat everything on it as suspect until you have read about counterfeit cards.
- Is the loss limited to the newest files, or does it hit the oldest ones? Newest-only points to an interrupted write or a power event. Oldest-first points to wrap-around overwriting on a fake card.
- Did the failure appear immediately after a firmware update or a change of device? That narrows it to a write bug or a volume-state conflict between two hosts.
- Does the card go read-only, or repeatedly disappear and reappear? That is controller and FTL behavior, and it is the one pattern that should go to a lab rather than a scanner.
- Has the card been getting slower for weeks, or has it been running hot? Wear and thermal stress fit here, and the practical response is to stop using the card rather than to keep testing it.
The urgency framing for the first hours after a loss, including which actions to avoid, sits in camera data loss what to do. Once you have a diagnosis, the procedural side is in how to recover camera data.
What recoverable actually means in practice
There are three outcomes, not two, and conflating them is what produces most of the disappointment around data recovery.
The first is that the data was never overwritten. The file system lost track of it, but the blocks still hold the original bytes, and a scan that reads the raw device and rebuilds the index can return the files in full. This covers deletion, quick format, most file system corruption, and most interrupted writes.
The second is that the files come back but will not play or open. Here the payload is largely present but the container index is broken, which for video means missing or damaged index atoms, and for photos means a truncated or misordered data stream. The recovered material can be made playable by re-encoding the recovered portion, which rebuilds the indexes from the decoded frames rather than trying to repair the original structure.
The third is that the data was overwritten. Once new writes have landed on the blocks that held your files, no tool recovers them, and no laboratory recovers them either, because the previous charge state is physically gone. This is why the correct first instruction is always to stop writing to the card, and why continuing to shoot on a card you intend to recover is the one mistake that makes a good situation into a hopeless one. The boundary between the second and third outcome, and how to tell which side a specific file is on, is the subject of unrecoverable data.
On Windows, if the card is still detected and the loss is logical, a purpose-built recovery tool such as XRecovery (xappset.com) can scan SD, microSD, and other memory cards for deleted or formatted photos and videos. As with any recovery tool, stop writing to the card first and always save recovered files to a different drive, never back to the card you are scanning.
Frequently Asked Questions
Why do my camera files keep getting corrupted on the same card?
Repeat corruption on one card points to the media rather than the workflow unless you can name a specific behavioral trigger. Test the card’s real capacity first, because wrap-around overwriting on a counterfeit card produces exactly this pattern. If capacity is genuine and the card is several years old or has been heavily rewritten, suspect wear and retention loss. If corruption started immediately after a firmware update, suspect the camera.
Can a corrupted memory card be fixed, or do I need a new one?
The card can often be made readable again by rebuilding its file system, but that is a repair of the index, not of the data, and some repair operations make the original files harder to recover. Recover first, verify the recovered files, and only then decide whether to repair or retire the card. A card that has produced one logical failure is usually safe to reuse after a full format and a capacity test. A card that has produced two should be replaced.
Does formatting in the camera erase my photos permanently?
No. The in-camera option is a quick format: it writes a new, empty file system and marks the old clusters as available without erasing the flash cells. Your photos remain physically present until the camera writes new data over those blocks. The practical implication is that you should stop using the card immediately after an accidental format, and that every additional shot reduces what remains recoverable.
How long do I have before lost camera files become unrecoverable?
There is no clock, only a write counter. Files stay recoverable as long as nothing overwrites their blocks, which on a card left in a drawer can be months, because camera cards do not issue TRIM commands the way SSDs do. Files become permanently unrecoverable the moment new data lands on those blocks, which on an actively used card can happen within a single shoot.