Container: the wrapper, and what rides in it
A container is the file structure holding a video's streams and its blocks of information about them. Whether a mark is inside the picture or inside the container predicts everything about its durability. As of 2026-09-22.
| The term | Container |
|---|---|
| What it names | The file structure wrapping a video's streams and metadata blocks |
| What it is not | The picture, or the compression applied to it |
| Where the register uses it | Explaining why the two mark layers behave oppositely |
Inclusion rule. Words this site uses in a narrow sense, where the ordinary sense would lead a reader to misread a cell. No vendor statement appears on this page. Order. Fixed order: what the word names, what it excludes, then where it is used here.
1One structural fact explains the whole subject
A video file has a picture inside it and information wrapped around it. A burned-in mark and an in-picture signal are in the first place; a credential manifest and metadata keys are in the second. Every difference in how those behave follows from that.
Operations on the picture, such as cropping, threaten the first group and leave the second untouched. Operations on the container, such as re-encoding and stripping, threaten the second and leave the first alone. Nothing else needs to be known to predict most outcomes.
2Why the readable layer is the fragile one
A manifest is readable precisely because it is a structured block in a container that ordinary tools already parse. That is also why an encoder can drop it without anybody noticing. The properties are two sides of one design decision rather than an oversight.
An in-picture signal has the reverse trade: it needs a scheme-specific detector to read and survives handling that removes attached blocks. One entry in this register names both kinds on the same product, so one file can carry both profiles at once.
3What this means for a claim in the register
A vendor saying its output carries a manifest is describing the file as delivered. Whether that block is still present three steps later is a question about the handling in between, which no page here addresses except one caveat.
So the file-contents column records what a vendor writes and not what arrives. The readings say which layer each claim concerns, because that is the part a reader needs in order to know what to protect.
Nothing on this page is a vendor statement; the values it helps read are on the support table, with the page and the date each one was read from. See also the machine-readable column, what survives an export. Nearby terms: crop, reframe, lossy compression.