Lankio Tools

What Metadata Is Stored in MP4, MOV, and WebM Files?

Published September 1, 2026

What Metadata Is Stored in Videos? covers the general picture — container tags vs. stream-level details. This one gets specific about how the major container formats actually differ under the hood.

MP4 and MOV: close cousins

MP4 and MOV (QuickTime) are both built on the same underlying ISO Base Media File Format structure — a series of nested "boxes" (MP4 terminology) or "atoms" (QuickTime's original terminology for the same concept). This is why tools like ffprobe/ffmpeg handle both with the same underlying logic, and why our Video Metadata Checker and Video Metadata Cleaner treat them consistently.

Metadata commonly found in this family:

  • A moov box holding format-level tags: title, comment, creation time, encoder
  • QuickTime-specific location tags (historically written as ©xyz, encoding GPS as an ISO 6709 string) on videos recorded with location services enabled
  • Per-track metadata (a trak box per video/audio stream) including codec details and handler names
  • Structural ftyp box fields (major_brand, minor_version, compatible_brands) — these identify the file format itself rather than describing the content, and are not something a metadata-removal tool is expected to (or should) strip

WebM: a different foundation entirely

WebM is built on Matroska, an open container format using a binary tree structure called EBML (Extensible Binary Meta Language) rather than ISO's box structure. Metadata in WebM/Matroska files lives in dedicated EBML elements — a Tags element for descriptive metadata, SegmentInfo for format-level details like the writing application. Despite the different underlying structure, the same class of information (title, encoder, creation date) is typically present in an equivalent form, which is why our tools can report on WebM files using the same normalized categories (container tags, per-stream details) as MP4/MOV.

MKV: WebM's more permissive sibling

MKV (Matroska) is the general-purpose format WebM was based on — WebM is essentially a constrained profile of Matroska limited to a specific set of codecs. Metadata-wise, they work identically since they share the same EBML foundation.

What this means for cleaning

Because our Video Metadata Cleaner uses ffmpeg's -map_metadata -1 on a stream copy rather than a per-format custom implementation, it applies uniformly across all four containers — but the same caveats apply to all of them equally: structural identification fields survive (they're not metadata in the descriptive sense), and vendor-specific private boxes/elements ffmpeg doesn't recognize as metadata aren't guaranteed to be removed. See What Metadata Is Stored in Videos? for the full explanation of that limitation.