Videos take up a lot of space, pile up quickly and end up scattered across phones, memory cards, external drives and cloud services. The risk is not only losing them to a failure, but also finding yourself in a few years with files no program opens properly any more, or with a mountain of clips with no names or dates. Archiving well means choosing formats carefully, organising files and, above all, having a method for checking that everything is still intact.
Masters and access copies
The first principle is to separate two kinds of file:
- preservation master: the highest-quality version available, which is never modified and not used for everyday viewing;
- access copies: light, compatible versions for watching, sharing or uploading, which can be regenerated from the master at any time.
This way the access format can change over the years, following the devices of the day, without touching the original.
What to keep as the master
Native digital footage
For files produced by smartphones and cameras, the best master is almost always the original file as it is. Re-encoding it to another format, even losslessly, adds no quality and introduces an extra step. Just keep it intact, with its original name or renamed according to a consistent scheme.
Tape and film digitisations
For captures from VHS, Hi8, MiniDV or film, the master must preserve all the information captured. Here a lossless codec makes sense. The one most used by archives is FFV1, an open, documented codec, often in a Matroska (MKV) container. It has several advantages:
- lossless compression, with files much smaller than uncompressed video;
- public specifications and a free implementation in FFmpeg;
- the option of embedding checksums in individual slices, to detect corruption.
An example conversion:
ffmpeg -i capture.avi -c:v ffv1 -level 3 -g 1 -slicecrc 1 -c:a flac master.mkv
The -g 1 option makes every frame independent; -slicecrc 1 adds integrity checks; FLAC keeps the audio lossless.
Editing projects
For a finished project it is worth keeping the exported master in a high-quality or lossless intermediate codec, the original files used and, if possible, the project file with a list of materials. The project file alone is not enough: in ten years the program may no longer exist.
Open, documented formats
A format is suitable for preservation when:
- it has public specifications, so that a program to read it can be written in future too;
- it is widely used, with several independent implementations;
- it does not depend on licences or online services to be opened.
FFV1, Matroska, FLAC, and also H.264 in MP4 for access copies, meet these criteria well.
Names, folders and metadata
An archive is only useful if you can find what you are looking for. A few simple rules:
- date-based names in year-month-day format, for example
2024-07-14_wedding_ceremony_01.mkv, which sort correctly on any system; - no spaces or special characters in names, to avoid problems across systems;
- a stable folder structure, by year and event;
- an accompanying text file for each collection, with a description, people, place, origin of the tape and notes on the digitisation.
Metadata embedded in files is useful, but a simple text file alongside remains readable with any tool.
The copies rule
A widely used backup rule is 3-2-1: three copies of the data, on two different media, one of them kept in a different place. For a home video archive this might mean:
- one copy on the computer or home NAS;
- one copy on an external drive kept disconnected;
- one copy in a cloud service or at a relative's house.
A disconnected drive also protects against accidental deletion and against malicious software that encrypts connected files.
Checking integrity
Files can become corrupted silently: a bad sector, a copy error, an ageing medium. To notice in time:
- calculate a checksum (for example SHA-256) for each file when archiving it and save it in a file alongside;
- repeat the check periodically, for example once a year, and after every copy;
- if a file does not match, restore it from another copy.
On Linux and macOS you can use sha256sum or shasum -a 256; on Windows the certutil -hashfile command. For FFV1 files with per-slice checks, FFmpeg can also report damaged areas when decoding.
Migrating media
No medium lasts forever. Hard drives, SSDs, optical discs and memory cards have different and hard-to-predict lifespans. Plan a periodic migration of the archive to new media, verifying checksums after every transfer.
In short
- Separate the preservation master from access copies.
- For digital footage keep the original intact; for digitisations use a lossless format such as FFV1 in MKV.
- Prefer open, documented and widely used formats.
- Use date-based names, stable folders and accompanying notes.
- Keep three copies, verify checksums and migrate media over time.
