MP4 metadata atoms
MP4 metadata uses structured boxes and metadata conventions whose keys, payload types and scope must be interpreted separately from the coded media streams.
MP4 packages media and associated descriptions through boxes, often called atoms in related terminology. The movie's coded tracks and a descriptive title occupy different structures.
One common tag convention uses ilst
An illustrative iTunes-style hierarchy is:
moov
udta
meta
ilst
©nam
data
The visible key ©nam is commonly associated with a title in that convention. Other keys describe artist information or cover artwork.
| Example key | Conventional role |
|---|---|
| ©nam | Title |
| ©ART | Artist |
| aART | Album-artist relationship |
| covr | Embedded cover data |
These labels name a metadata convention. They should not be read as a promise that every MP4 uses this hierarchy or every reader exposes the same fields.
Other metadata arrangements use another contract
Keyed metadata can use a key table and namespace relationship instead of interpreting every entry as one familiar ilst key. FFmpeg documents an option to write mdta metadata.
A reader supporting one arrangement can still fail to map another arrangement into the expected library title. Renaming a file does not convert those boxes.
The payload also has a type. A string, number and picture bytes need their appropriate interpretation rather than one blanket “decode every data box as text”.
The description does not replace stream properties
A title or album tag can describe the work represented by an M4A or video file. Track/sample structures establish the actual codec, duration and media layout.
For Emby, compare the descriptive atom the file carries with the imported field and the selected media source. Correct playback and correct title import can succeed independently.
A rewrite has preservation choices
A remux can retain coded media while changing which metadata boxes are copied or rewritten. An export can produce a useful new audio file while omitting embedded artwork or custom fields.
Embedded cover art explains the picture-selection boundary; Unicode metadata develops the text-decoding boundary.