How Media Files Go Bad While Sitting Still

Files that are never opened can still become unreadable. Storage media age, errors accumulate quietly, and video suffers disproportionately when they do.

Storage media have finite lifespans

Magnetic media lose signal strength over time and mechanical drives have moving parts that eventually fail. Neither is designed for indefinite passive storage.

Solid-state storage holds charge in cells that leak gradually, and leakage accelerates at higher temperatures. Unpowered drives are not a long-term archive.

Optical media vary enormously in quality, and the layers in consumer discs degrade at rates that depend heavily on manufacturing and storage conditions.

Errors accumulate without any warning

A stored bit can flip without the file being touched. The file system reports the file as present and normal, because nothing checked its contents.

This is often called silent corruption, and the word silent is the important part. Damage is discovered when the file is finally opened, which may be years later.

Backups made after corruption faithfully preserve the corrupted version. A backup rotation can quietly replace every good copy with a damaged one.

Video is unusually vulnerable

Because most frames are described relative to others, a small error propagates until the next self-contained frame. A handful of bad bytes can wreck seconds of playback.

Damage to a container's index or header can render an entire file unopenable, even though the picture data behind it is intact.

Compression amplifies the effect generally, since a compressed file has less redundancy for anything to recover from.

Checksums turn silence into evidence

Recording a checksum when a file is created allows later verification. A mismatch proves corruption, and a match proves the file is unchanged.

Verification must be run deliberately and periodically, because nothing prompts it. An archive that is never verified provides no assurance.

Some file systems compute and check this automatically and can repair damage when a redundant copy is available. That combination is what makes damage self-correcting.

Copying is its own risk

Transfers can fail partway or introduce errors, particularly across networks or from failing media. A copy that reports success is not proof of a faithful copy.

Verifying checksums after transfer is the only reliable confirmation. Comparing file sizes catches truncation and nothing else.

Deleting an original immediately after copying removes the ability to retry. Keeping it until verification completes costs a little space and prevents an unrecoverable loss.