Memory cards · bench notes · CDF-2025-0907
Big Files, and Not a Frame Would Play.
A videographer from Caerphilly came in with a card that had turned on him overnight, just back from a wedding marquee in a field outside Machen: the last three files of the night just will not import
. What he said next set out the whole job — they are full size when I look at the card, so the footage must still be on there
— and the couple were flying out in under a fortnight.
Does that match yours? Call us.
0800 6890668
Why that happens.
He had it right, and the reason is worth setting out. While a camera is rolling it streams picture data straight onto the card. The index — a table of where every frame sits — only gets written when the file closes cleanly. Take the card out while the light is still on and the pictures remain, but that table is never made. The same thing happens if the battery dies or the card runs out of room. Editing software finds its way around a clip using nothing but that table, so a file packed with frames and missing its map looks, to the software, exactly like an empty one.
What we used on this one.
Every step, in order →| The tools | What it did here | Why it helps |
|---|---|---|
| Klennet Carver | Assembled the surviving frames into containers his edit suite could open | It understands how a video file is built; blunter tools hand back fragments that will not play |
| PC-3000 Flash | Made the complete sector-level copy of the card before any work started | Bypasses the controller to read the NAND itself, against a manufacturer-ID library kept up to date |
| UFS Explorer Professional Recovery | Lifted the intact clips and the card’s folder tree out of that image | Reads the volume formats that defeat most software: APFS, ReFS, XFS, ZFS, Btrfs |
How it went.
Image it first, before anything else
Nobody offered a view on the fault until a copy of the card existed. A deadline the customer has already given a client alters how a job is paced; it does not entitle anyone to run experiments on the one and only original. After that, every step ran against the copy. The card itself went into an antistatic sleeve and stayed there, kept back in case the original ever needed a second read.
Get the working clips out of harm’s way first
Most of the evening lifted off the image with no difficulty and went straight somewhere safe before another thing was tried. The three awkward files were then taken apart and studied. His theory stood: plenty of frame data on all three, a truncated closing index on two, and none at all on the third. That finding shaped everything that followed.
Let the healthy clips supply the pattern to follow
Camera footage is rarely written in one continuous stretch, and fragmentation is precisely where one-click recovery software gives up — whatever it produces will not open anywhere. Undamaged clips shot earlier that same evening, same body, same codec, same settings, gave us a template. Each broken file was rebuilt around the frames that remained, then tested the only way that means anything: watched right through, start to finish.
Signing it off.
Two of the three went back whole, and each one was watched from start to finish before it left the bench. The third plays cleanly until the last few seconds, which never made it out of the camera’s buffer before the card was removed. The finished film reached the couple before their flight.
Pages people often read next.
Filed under Memory cards as well.
Is that what yours is doing?
Power it down, get it to us, and hold off on any decision until the diagnosis says what can still be read.