MP4 timed text (tx3g)

tx3g identifies a 3GPP timed-text sample entry whose container timing and text modifiers describe a subtitle track.

tx3g is the sample-entry identifier for the 3GPP timed-text format commonly encountered in MP4 and 3GP subtitle tracks. Inspection tools can also label the decoded codec mov_text.

This is not a standalone SRT file hidden inside MP4. The track has container timing and a sample description, while its timed samples contain text and optional presentation modifiers.

Component Responsibility
tx3g sample entry Default text box, style, font-table references and presentation flags
Container sample timeline When each sample starts and how long it is presented
Text sample A byte-length field, encoded text and any modifier boxes
Renderer Decode and lay out the sample using its defaults and modifiers

Bytes and characters are counted differently

A text sample starts with a 16-bit text-length field. That length counts the text bytes, not all modifier bytes and not the number of visible glyphs.

For the UTF-8 spelling café, the text occupies five bytes although it has four composed Unicode characters. Its two-byte length field plus text therefore occupy seven bytes before optional modifier boxes.

The 3GPP format supports UTF-8 and signaled UTF-16 text. Styling ranges use their defined character positions. A byte index into a UTF-8 string cannot be casually used as the position of a displayed character.

Modifiers can describe styled spans, highlighting, links or other supported behavior. The receiving player chooses the implementation of those defined features; the identifier tx3g alone does not mean that the track has only plain unstyled sentences.

Empty samples can represent the gaps

Take a track on a 1,000-unit-per-second time scale:

Sample start Duration Text
0 12,500 Empty
12,500 2,500 The last train leaves at midnight.
15,000 1,200 Empty
16,200 1,800 Then we still have time.

The empty sample starting at 15 seconds prevents the first sentence from remaining on screen until the next sentence. The timing comes from the container's sample schedule. Any edit-list mapping or other applicable track timing also needs to be applied.

A simple text extraction reconstructs cues from that schedule. It must account for empty text and sample duration rather than copying every packet into a visible subtitle line.

For an Emby MP4, container recognition and subtitle rendering are separate capabilities. A route can retain the track, convert its decoded text for the app, or render its presentation upstream. Its styled ranges and other modifiers determine which presentation details a simpler text conversion can retain.

See Also

References