A wedding is the one job where "we can reshoot it" does not exist. That single fact should change how you handle storage, and it definitely changes what you do in the first hour after something fails.
This is written professional to professional: where the chain actually breaks across a wedding workflow, why ingest is the most dangerous step of the whole job, what modern CFexpress and UHS-II cards do when they fail, what a clean RAW recovery really returns, and how to protect a delivery deadline without turning a recoverable problem into a permanent one.
First: stop writing to the card or drive
If a card, drive or array holding wedding files has just failed, what you do in the next few minutes either preserves your options or removes them. Work through this before anything else.
- Eject it and put it in a labelled sleeve. Do not put the card back in the camera. Do not keep shooting to it. If the problem is on a working drive, quit Lightroom or Capture One, unmount the volume and shut the machine down.
- Decline every offer to fix it. macOS will offer to initialise an unreadable disk, Windows will offer to format it or to scan and fix errors, and the camera will offer to format the card. All three write to the media. Say no to all of them.
- Do not run recovery software on the original. Consumer scavenging tools write their cache, their logs and sometimes their output back to the same device, and they can overwrite the exact frames you are trying to save. If a drive is clicking, buzzing or spinning up and down, software is the wrong tool entirely and every power-on costs you.
- Write down the sequence while it is fresh. Which body, which slot, which card, what the error message said, what you tried next. That history genuinely changes how a case is approached.
Then get it looked at before you commit to anything. Diagnosis at our laboratory is free and you get a fixed written quote before any work begins, so you can find out what is actually wrong without spending money to find out.
The six failure points in a wedding workflow
A wedding job is not one storage device. It is a chain of six, and the file only becomes safe once it exists in two places you have actually verified. Each link fails in its own characteristic way.
| Stage | Typical cause of failure | What you see |
|---|---|---|
| In-camera card (SD, UHS-II, CFexpress) | Controller or firmware fault, card pulled while the buffer was still flushing, damaged pin row | “Card error” in body, card reads as unformatted or 0 bytes, or only part of the ceremony is present |
| Card reader and cable | Bus-powered reader losing contact mid-copy, an overloaded hub, a knock to the desk | Copy stops part way, or files land with correct names and sizes but will not open |
| Ingest drive (bus-powered portable) | Knocked or dropped while spinning, connector torn from the board, head or motor failure | Clicking, not detected, or “you need to format this disk before you can use it” |
| Working SSD | Controller failure with no warning signs, firmware corruption, sudden loss of the volume | The volume simply is not there on the next boot |
| Backup drive | Never verified, or kept in the same bag as the original so both are lost together | You discover at the worst moment that the backup is stale, partial or empty |
| NAS or RAID archive | A second disk failing during rebuild, a volume dropped after an update, a share deleted in error | Array degraded or crashed, volume unmountable, shares gone |
Different links need genuinely different handling. A knocked portable drive is mechanical work covered on our external hard drive recovery page, a dead working drive is SSD data recovery, and a failed archive volume is NAS data recovery where the disk order and array parameters matter as much as the disks themselves.
Why ingest is the most dangerous moment of the whole job
Ingest is the only point in the workflow where a wedding exists in exactly one place while it is being moved. It also happens at the worst possible time: late at night, after a fourteen-hour day, with a second shoot the following morning.
The specific trap is the reformat. A card gets cleared for the next job before anyone has confirmed the copy is complete and readable. If the transfer stalled, if the reader dropped out for two seconds, or if a folder was still writing when the drive was pulled, you now have a set of files that look right in Finder or Explorer and are not right at all. Free space on a reformatted card is not a safe place to store a wedding.
A safer ingest is not complicated, it is just disciplined:
- Copy, never move. Nothing should be deleted from a card as part of a transfer.
- Two copies to two physically separate drives before any card is touched again, and keep those two drives in different bags.
- Verify by checksum, not by file count. Matching counts tell you nothing about whether the bytes arrived intact. Use an ingest tool that verifies.
- Keep the cards untouched until the gallery is delivered and accepted. Buy enough cards for the season so you are never forced to reformat mid-job.
- Format in the camera, not the computer, and only after all of the above.
If ingest is where you lost the files, say so plainly when you bring the media in. Whether a card was reformatted, and whether anything was shot to it afterwards, is the first thing that determines what is realistically retrievable.
CFexpress, UHS-II and what modern cards do when they fail
Modern camera media is not simple flash any more. CFexpress Type A and Type B are effectively NVMe solid-state drives in a card shell, with their own controller, firmware and wear-levelling tables. UHS-II SD adds a second row of contacts to carry the faster bus. Both facts change what a failure looks like.
When a high-speed card dies, the memory itself is often still holding your frames. What has usually failed is the controller, or the mapping tables that tell it where each block lives. The card enumerates as 0 bytes, or does not appear at all, or runs unusually warm. Nothing on the normal interface can reach the data, which is why consumer software finds nothing and why repeated retries are wasted effort. Getting to it is laboratory work at the controller and chip level, in an environment built for it.
The physical failures are worth knowing too. The second pin row on UHS-II cards is delicate and easily scored. CFexpress Type B sockets bend pins if a card goes in at an angle in a dark ballroom. Cards that have been in a pocket through a poolside shoot or a rainy outdoor tea ceremony come in with contact corrosion.
One thing to be clear about: any physical work we do on a card or drive is done purely to make it readable long enough to image it. You receive your recovered files, not a working card. A card that has failed once should never be trusted with a wedding again. The same principles apply across formats, and the detail for smaller media is on our SD card recovery page.
RAW files: what a good recovery actually returns
“We found some JPEGs” is not a recovery. For a wedding delivery, the quality of the result is measured on four things.
- Structural completeness of the RAW. CR3, NEF, ARW, RAF, ORF and DNG are container formats. A file carved out of free space can look correct in a file listing and still be truncated, so it opens as an embedded preview only, or the last third shows banding and colour breakup. That difference only shows up when you open it in your editor.
- Names, folders and sequence. Recovering at filesystem level rather than by carving keeps original filenames, folder structure, timestamps and frame numbering. That matters enormously when you deliver in ceremony order and match against a shot list. A folder of 4,000 files called REC0001 onwards is technically a recovery and practically a nightmare.
- Video. Wedding films sit in large MOV or MP4 files that are fragmented across the volume. If the camera never closed the file because a battery was pulled, the index that makes it playable may be missing even though the picture data is there.
- Sidecars and catalogues. Your edits live in XMP sidecars, Lightroom catalogues or Capture One session files, and they are often on a different drive from the images. If the editing work matters as much as the originals, say so at the start so it is included in scope.
Before you pay anything, you see the recovered file list and confirm it. For a photographer that is the useful part: you can check that the ceremony folder, the vows and the family formals are genuinely there, at the counts you expect, before money changes hands. Recovery of images from a laptop or working system follows the same route, covered on our Mac data recovery page.
The delivery deadline, and how to manage it
Turnaround pressure is what pushes photographers into the damaging move. The DIY attempt at 1am, the third-party utility, the “let me just try one more format”. Deadline pressure is real, but the sequence that protects the delivery is the boring one.
- Get a diagnosis before you promise anything. Free diagnosis gives you a realistic picture of what is recoverable and how long it will take, which is the only honest basis for a conversation with the couple or the planner.
- Tell the couple early, in plain terms. A calm holding message from you beats a silent missed deadline. You do not need to explain head-stack work to them, only that the files are with a laboratory and you will confirm the delivery date.
- Triage the folders. Tell us which parts matter first: ceremony, vows, first dance, family formals. Work can be ordered around a delivery, and a partial gallery on time is often worth more than everything a week late.
- Bring the whole set. The card and the reader you used, the external drive with its own enclosure and power supply, and for a NAS the unit plus every disk, labelled in bay order. Missing one disk from an array stalls the whole job.
- Do not run anything else while you wait. Every additional attempt narrows the options.
On cost: NAS, RAID and server recovery runs roughly SGD 400 to SGD 4,000 depending on the array and the failure. For everything else, diagnosis is free and you get a fixed written quote before any work starts, with no recovery, no fee on standard cases. There is more detail on our data recovery cost page. Our laboratory is at 6 Harper Road beside Tai Seng MRT, open Monday to Friday 9.30am to 6.30pm, with weekends and public holidays by appointment.
Confidentiality: it is your clients’ wedding, not just your files
Wedding media is other people’s private life. Identifiable guests, family moments, sometimes public figures or corporate hosts, and a contract between you and the couple that almost certainly says something about who may handle the images.
Handle a failure the way you would handle any other client confidentiality issue. Do not post the failing card with the couple’s name in a photographers’ group chat looking for advice. Do not hand media to a shop that ships work overseas to a partner laboratory, because you lose sight of where the files are and who has touched them.
Everything here is handled in line with Singapore’s PDPA, work is done in our own Singapore laboratory and nothing is sent overseas. NDAs are available on request, and if your own client contract requires one, ask for it before you hand anything over rather than afterwards. It is a reasonable, routine request and no serious laboratory will blink at it.
It is also worth reading your own terms before you say anything to the couple. Most photographer contracts carry a clause about backup and data loss, and knowing what yours says shapes how you frame the conversation. If you are ready to bring media in, or you want to describe the failure first and get a view on it, contact us here.
