How DVD VOBU Boundaries Align with Chapter Markers

In DVD-Video authoring, chapter markers rely directly on Video Object Unit (VOBU) boundaries to enable random access and seamless navigation. Because DVD players decode video starting from self-contained video units rather than arbitrary frames, every chapter marker created by an author must correspond to the start of a VOBU. This article explains the technical relationship between VOBU boundaries and chapter markers, how authoring software enforces this alignment, and what happens when markers do not match encoded video frames.

What Is a VOBU?

A Video Object Unit (VOBU) is the basic structural building block of a DVD-Video program stream. It typically represents between 0.4 and 1.0 seconds of content (up to a maximum of 1.2 seconds for NTSC or 1.4 seconds for PAL).

Every VOBU must begin with a Navigation Pack (NV_PCK) containing presentation and data search information, immediately followed by video, audio, or subpicture packets. Crucially, the video payload inside a VOBU begins with an MPEG-2 Group of Pictures (GOP) that starts with an intra-coded frame (I-frame). Because I-frames are compressed independently of other frames, they serve as the only valid entry points where video decoding can immediately commence.

The Chapter Marker (PTT) Requirement

In DVD terminology, a chapter marker defines an entry point called a Part of Title (PTT). When a viewer presses the "Next Chapter" button or selects a chapter from a menu, the DVD player looks up the target cell and sector address in the disc’s navigation tables (IFO files).

Because decoding cannot begin on predictive frames (P-frames or B-frames) without referencing previous data, a chapter marker cannot be placed on an arbitrary frame. It must point to a byte address on the disc containing an NV_PCK and a subsequent I-frame. Consequently, a chapter marker is inherently bound to the exact start boundary of a VOBU.

How Alignment Is Handled During Authoring

To align chapter markers with VOBU boundaries, DVD authors use one of two primary workflows:

1. Pre-Encoding Flagging (Forced I-Frames)

The most precise approach requires setting chapter points before or during the MPEG-2 encoding stage.

• The author generates a chapter list containing specific timecodes.
• The MPEG-2 encoder reads this list and forces a closed GOP and an I-frame at each specified timecode, closing the previous GOP regardless of standard GOP length.
• When the DVD multiplexer processes this stream, it places a new VOBU header at every forced I-frame.
• The resulting chapter marker lands on the exact frame intended by the author.

2. Post-Encoding Snapping

If an author attempts to insert a chapter marker into a pre-encoded MPEG-2 stream that lacks an I-frame at the desired timecode:

• The DVD authoring software cannot split an existing GOP without re-encoding.
• The software automatically "snaps" the chapter marker to the nearest existing VOBU boundary (the closest preceding or succeeding I-frame).
• Depending on GOP length, this can shift the actual chapter playback point by up to half a second or more from the author's intended cut.

Structural Relationship in the DVD File System

When the disc layout is compiled, the authoring system writes the correlation between chapters and VOBUs into the control files:

• The PTT Search Table (VTS_PTT_SRPT): Maps the chapter number to a specific Title and Program Chain (PGC).
• The Cell Playback Information Table (C_PBIT): Defines the start address of the cell representing the chapter.
• The VOBU Address Map (VOBU_ADMAP): Lists the starting sector address of every VOBU on the disc.

The chapter entry point references the first cell of that chapter, and that cell's starting sector corresponds directly to the start sector of the target VOBU's NV_PCK. Through this structure, DVD players achieve instantaneous chapter jumps without decoding intermediate frames.