Six sections, eighteen pieces
Where everything is
The whole site in one place: six sections, and every piece filed under the one it belongs to.
The codec is the argument
What gets thrown away, and who decided.
A whole film on one disc, and a codec that was never meant to ship
DivX ;-) 3.11 Alpha was a hacked build of Microsoft's MPEG-4 implementation. It made a feature fit on a CD-R, and a company formed around what followed.
Long
The open-source answer had to be written twice
XviD was rebuilt from scratch after the earlier open project's licensing collapsed, and it stayed compatible with the same decoders.
Medium
H.264 is still everywhere, and that is the achievement
Standardised in 2003, H.264/AVC remains the format with the widest hardware decode support of any video codec, two decades after ratification.
Long
AV1 was built to be free, and free took years
The Alliance for Open Media published AV1 in 2018 as a royalty-free codec, explicitly to get out from under the patent pools.
Medium
The file is not the format
AVI holding what it should not have, and what replaced it.
AVI held things it was never designed to hold
The AVI container came from 1992 and predates most of what people put in it, which is why variable bitrate audio and modern subtitles fit so badly.
Long
Matroska was designed for exactly what AVI could not do
MKV was built as an open, extensible container with real support for multiple tracks, chapters and subtitle formats.
Medium
MP4 came out of a committee and won anyway
The MP4 container derives from Apple's QuickTime file format by way of the MPEG process, and it is what the web settled on.
Short
What people actually used
One person's editor, a library under everything, and a front end for the hard part.
One person's editor became everybody's
VirtualDub was written by Avery Lee and released free. For years it was the default way to cut and filter AVI on Windows.
Medium
FFmpeg is the layer under almost everything else
Most tools that claim to convert video are calling FFmpeg. It is the piece the ecosystem depends on and the piece nobody markets.
Long
HandBrake put a front end on the difficult part
HandBrake wraps encoder settings in presets, which is why people who never learned the parameters could still get a usable file.
Short
Every setting is a trade
Bitrate against resolution, keyframes against size, and generation loss.
Bitrate against resolution, and something has to give
A fixed bitrate spread over more pixels means less for each one. Every encode is a decision about where that shortage should show.
Long
A keyframe is expensive and you need them anyway
Keyframes are stored whole rather than as differences. Too few and seeking breaks; too many and the file grows.
Medium
Encoding an encode is always worse
Lossy compression discards information permanently, so re-encoding compounds. The second pass cannot recover what the first threw out.
Medium
Four generations, four arguments
MPEG-2 on disc, H.264 everywhere, HEVC in the pools, AV1 built to escape them.
MPEG-2 was good enough because the disc was big
The DVD-Video spec could afford MPEG-2's inefficiency because it had 4.7 gigabytes to spend. The constraint moved when the disc did.
Medium
HEVC arrived tangled in more than one patent pool
HEVC's licensing fragmented across competing pools, which slowed adoption enough to create the opening AV1 was built for.
Long
The disc stopped being the unit of measurement
When delivery moved to streaming, the question changed from what fits on a disc to what survives a variable connection.
Medium
The people and foundations behind the formats
Where the code came from and who kept it free.
One person's hack, and the company that grew on top of it
The DivX ;-) build is credited to Jérôme Rota. DivXNetworks formed afterwards to make a legitimate codec out of the attention.
Medium
Xiph kept building formats nobody had to license
The Xiph.Org Foundation produced Vorbis, Theora, FLAC and Opus, on the position that a format everybody uses should not be owned.
Short