What to do after losing photos is decided almost entirely by the first hour, and the common mistakes are not exotic: letting the computer repair the card, taking test shots to see if the camera still works, and running three recovery applications in sequence on the original media. Every one of those writes to the card, and every write reduces what can be recovered.
The plan below divides the first hour into four blocks you can follow against a clock, because panic produces improvisation and improvisation destroys evidence. Total time to reach a position where nothing further can be lost: about sixty minutes, most of it waiting.
Minutes 0 to 5: stop the machine, not yourself
The goal in the first five minutes is to stop anything from writing to the card. Nothing else matters yet, and nothing you do in this window is wasted.
- Remove the card from the camera. Do this before switching anything off if the camera is already off. If the body is mid-write, and you can see a card access light, let it stop before you pull the card.
- Set the write-protect switch to LOCK if the card has one. That physical tab is the cheapest possible guarantee that nothing can be written for the rest of the session.
- Do not format anything, anywhere. Not in the camera, not in Windows, not in a camera menu that offers to “repair” the card. A card reporting an unreadable file system is reporting that its directory structure cannot be parsed, not that its data is gone.
- Do not insert the card into the camera again. A camera writes to a card on insertion: it maintains its own folder structure, updates its image database, and may write thumbnail indexes. Each of those is a write that can land on top of the data you are trying to recover.
- Do not take test shots with a different card in the same body yet. If the camera is the suspect, you will want to test it eventually, but not on a card holding anything you need.
There is one exception to the five-minute freeze. If the device is a phone or a camera with internal storage and the footage is inside it, power the device down rather than pulling anything. Internal storage cannot be separated from the device, and the correct approach is a different branch covered in the internal memory recovery guide.
Minutes 5 to 15: write down what you know before you forget it
This block feels unproductive and is one of the highest-value steps in the whole hour. A recovery decision depends on facts that are almost impossible to reconstruct from memory two days later.
Record these on paper or in a note on your phone, not on the card:
| Fact to record | Why it changes the recovery |
|---|---|
| Camera model and firmware version | Determines file container and whether internal storage is involved |
| Card brand, capacity, and model text on the back | Identifies the controller and file system behavior |
| What you did immediately before the loss | Deletion and formatting respond to different scanning modes |
| Whether the loss followed a disconnect, a power loss, or a drop | Separates logical failure from physical or electrical failure |
| Approximate number of files and total size expected | Sets a benchmark for whether a scan found everything |
| Whether the card has seen water, heat, or impact | Decides whether software is even the right tool |
| Any error message, verbatim | Error text maps directly onto diagnostic branches |
Photograph the card and, if the camera still shows anything at all, photograph the screen. If this is a paid shoot, this record is also the beginning of the conversation with a client, and having it ready changes the tone of that conversation considerably.
Two of those entries deserve emphasis because they are commonly misread. A card that failed after an unexpected disconnect is usually a logical problem with a very good recovery rate, and the failure pattern behind power loss during recording explains why the data is often still intact. A card that failed after a drop or a submersion is a physical problem where running software is not merely useless but sometimes harmful, which is why the emergency guide for physical and water damage exists as a separate branch.
Minutes 15 to 30: make the image, because the original will not wait
A card that reads today may not read tomorrow. Corrosion continues after water exposure, a hairline board crack propagates with each insertion, and a controller on the edge of failure gets no better with use. The only way to make a recovery repeatable is to stop using the original.
Make a byte-for-byte copy of the entire card to a file on a separate drive. Not a copy of the files, which is a different operation entirely, but an image of the raw device:
- Windows: a raw-image tool that can open a physical drive, or
ddfrom a Linux live environment. - Linux or macOS:
dd if=/dev/sdX of=/path/card.img bs=4M status=progress. - Any platform: a disk-imaging utility that verifies the image against the source after writing.
Three rules apply. Write the image to a different physical drive than the one you will save recovered files to. Keep the write-protect switch locked for the whole operation. And verify the image, because an unverified image is a guess.
If the card cannot be imaged because nothing detects it, that is a diagnosis in itself. Do not keep trying different readers in a loop; one attempt in a known-good reader is the test. If you need the decision logic for a silent card, the card not detected decision tree walks through it branch by branch.
Minutes 30 to 60: classify the failure and choose a lane
By the end of the hour you should be able to place the failure in one of four categories, and that placement decides the entire rest of the job.
| Category | Typical signs | Right lane | Wrong lane |
|---|---|---|---|
| Accidental deletion | Files missing, card healthy, folder structure intact | Recovery software on the image | Formatting, reusing the card |
| Format or directory loss | Card mounts, capacity wrong or empty, files gone | Recovery software on the image | Running a repair utility |
| Interrupted write | One clip or image truncated, rest fine | Repair or re-encode of that file | Scanning the whole card repeatedly |
| Physical or electrical failure | Not detected in any reader, heat, water, impact, crack | Data recovery laboratory | Any software at all |
The classification is usually obvious once the facts from the 5-to-15-minute block are in front of you, and it is worth saying plainly that the fourth row is not a bad-luck category. Cards in that category have a genuinely high laboratory success rate precisely because the memory chips are usually undamaged, while cards in the first two rows are the ones that software handles well.
If the card mounts and the files are simply absent, the loss is probably recoverable and the step-by-step recovery workflow is the direct next read. If the loss happened because the card was cleared in the camera, note that this is a much less destructive event than most people assume; the mechanics of recovering photos after an in-camera format explains why, and the reasons deleted files remain on a card is the underlying explanation.
The one reflex that destroys the most photographs
The single most expensive action in the first hour is accepting Windows’ offer to scan and fix the drive.
That prompt appears when a card’s file system cannot be parsed. It looks like help. It is not. A scan-and-fix operation writes a new file system to the card, and in doing so it can overwrite the very directory entries and allocation tables that a recovery scan needs in order to reconstruct the original folder structure. Users who click yes frequently lose the file names and folder layout permanently, and are left recovering a pile of unnamed files.
The same warning applies to camera-side repair functions, which some bodies offer when they detect an unreadable card, and to any utility described as a partition repair or file system fixer. Look at the card and make an image first; repair is what you consider only after the image is safe on another drive.
A second counterintuitive point is about the clock itself. The window that matters is not measured in minutes but in writes. A card that has been sitting in a drawer untouched for a week has had no writes and is in essentially the same state as one that failed an hour ago. A card that has been reinserted into a camera six times, tested in three readers, and scanned twice has been written to repeatedly and is in a materially worse state. The urgency in this plan is about preventing writes, not about beating a timer.
If you are on Windows and the card is still detected by a computer, 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. Software reads the card; it cannot repair a cracked card or a burned controller, and working from the image rather than the original keeps that option open.
If a client deadline is running in the background while you work through this, the tradeoff between speed and caution is worth making explicitly rather than by default. The triage logic for urgent client work covers how to split a card into what must be recovered first and what can wait. And when the hour is over, the habits that prevent a repeat are the ones around card handling: the case for ejecting a card properly and the list of recovery mistakes to avoid both take less time to apply than this plan took to execute.
Frequently Asked Questions
Is it already too late if the card has been used since the photos were lost?
Not necessarily, but every write since the loss has overwritten some of the space the deleted files occupied. The realistic question is how much was written, not how much time passed. Stop using the card now, image what is left, and scan the image rather than the card.
Should I run recovery software directly on the card or on a copy?
On a copy. A byte-for-byte image makes the recovery repeatable and protects the only original you have. Running a scan directly on the card gives you one attempt, and a scan that writes any temporary data to the card can reduce what a second attempt would find.
The camera says the card needs formatting. Should I let it?
No. That message means the camera cannot read the file system, which is a different condition from the data being gone. Decline, remove the card, and work from a computer. Accepting the format writes a new file system over the old one and permanently removes the folder structure you would otherwise recover.
How long do I have before the data is gone for good?
There is no fixed window. Data remains recoverable until the space it occupies is overwritten, which depends on how much new material is written to the card, not on elapsed time. A card left untouched is not degrading; a card being tested repeatedly in a camera is losing ground with every insertion.