3GP asset metadata

TS 26.244 clause 8 defines asset information: boxes in a user data box (udta) of the movie or of a track, describing the content. QuickTime and MP4 readers know many of the same codes (see the MOV registered user data items).

Common rules

Every asset box is a full box: size, type, version (uint8, 0) and flags (3 bytes, 0), so its data starts at offset 12. Most then have a 16-bit language field: one pad bit (0) and a packed ISO 639-2/T code, three 5-bit letters each stored as the letter’s ASCII value minus 0x60 (eng is 0x15C7). Strings are null-terminated UTF-8, or UTF-16 starting with the byte-order mark 0xFEFF. There may be several boxes of a type in different languages.

Text boxes

Box	Fields after version and flags
`titl`	language, title (string)
`dscp`	language, description
`cprt`	language, copyright notice
`perf`	language, performer
`auth`	language, author
`gnre`	language, genre
`coll`	language, name of the collection the media belongs to
`albm`	language, album title, then an optional uint8 track number

Example titl, English, “Hi”:

00 00 00 11 74 69 74 6C 00 00 00 00 15 C7 48 69 00   ....titl......Hi.

rtng and clsf

Offset	Size	Name	Meaning (`rtng`)
12	4	rating entity	FourCC of the rating body, such as `BBFC`.
16	4	rating criteria	FourCC of the rating, such as `PG13`.
20	2	language	Pad bit and packed language.
22	varies	rating info	String.

Offset	Size	Name	Meaning (`clsf`)
12	4	classification entity	FourCC of the classifying body.
16	2	classification table	Which of its tables; 0 means none in particular.
18	2	language	Pad bit and packed language.
20	varies	classification info	String.

kywd

Offset	Size	Name	Meaning
12	2	language	Pad bit and packed language.
14	1	keyword count	Number of keywords; more than 0.
15	varies	keywords	Each: a size byte (bytes in the keyword string) and the string.

loci

Offset	Size	Name	Meaning
12	2	language	Pad bit and packed language.
14	varies	name	Place name (string).
…	1	role	0 shooting location, 1 real location, 2 fictional location; others reserved.
…	4	longitude	16.16 fixed-point degrees.
…	4	latitude	16.16 fixed-point degrees.
…	4	altitude	16.16 fixed-point metres above sea level.
…	varies	astronomical body	String, such as `earth`.
…	varies	additional notes	String.

A longitude outside −180 to 180, or a latitude outside −90 to 90, means the coordinates are unspecified. The table lists these fields as unsigned 32-bit, but west, south and below sea level need negative values, so readers take them as signed two’s complement.

yrrc, urat and thmb

Box	Fields after version and flags
`yrrc`	Recording year, uint16. No language field.
`urat`	3 reserved bytes (0), then a star rating byte: 0 not rated, or 10 to 50 for 1 to 5 stars.
`thmb`	Format (FourCC, must be `jpeg`), then the image to the end of the box, following TS 26.234 clause 7.5.

orie

Offset	Size	Name	Meaning
12	2	digital zoom	8.8 fixed enlargement factor.
14	2	optical zoom	8.8 fixed magnification factor.
16	1 bit	pan indication	1 true north, 0 magnetic north.
…	…	pan	Fixed-point compass direction: north 0°, east 90°, south 180°, west −90°. Undefined when the camera points straight up or down.
…	4	rotation	16.16 fixed rotation about the direction the camera points.
…	4	tilt	16.16 fixed angle from the horizontal; 90 is straight up.

The clause’s table gives pan as 32 bits after the 1-bit indication, while its text describes a 16.15 fixed-point value; that suggests the indication and pan share one 32-bit word, but the exact packing could not be confirmed and is left open here.

ID32

Clause 8.3 carries an ID3 version 2 tag. The ID32 box sits in a meta box whose handler is ID32, either directly or as the primary item stored elsewhere.

Offset	Size	Name	Meaning
8	1	version	0.
9	3	flags	0.
12	2	language	Pad bit and packed language; `mul` or `und` if the tag’s own language fields disagree.
14	varies	ID3v2 data	A complete ID3v2 tag with its frames, the same ID3 version throughout, no footer; the size comes from the box size. See the MP3 ID3v2 page for its layout.

.3GP File Format – Asset metadata