VOB Sector Layout vs MP4 Atom Structure Explained

This article examines the core architectural differences between the sector-based layout of DVD Video Object (VOB) files and the hierarchical, atom-based structure of MPEG-4 Part 14 (MP4) files. While VOB files rely on continuous, fixed-size physical and logical sector alignments inherited from the MPEG-2 Program Stream specification, MP4 utilizes a tree of variable-length, object-oriented containers known as atoms or boxes. Understanding these contrasting methodologies reveals how storage media constraints historically shaped media containers, transitioning from physical disk sector streaming to modern, random-access index-based playback.

VOB: Sector-Aligned Stream Architecture

A VOB file is essentially an extended MPEG-2 Program Stream (PS) strictly tailored for physical optical media, specifically DVDs. The entire structural philosophy of a VOB revolves around fixed data blocks that correspond directly to the physical storage architecture of a DVD disc.

• Fixed 2048-Byte Sectors: DVDs are read in physical sectors of 2,048 bytes (2 KB). VOB files adhere strictly to this constraint: every "Pack" within a VOB is exactly 2,048 bytes in length, matching a single logical DVD sector.
• Packs and PES Packets: Each 2048-byte pack begins with a 14-byte Pack Header containing a System Clock Reference (SCR). The remainder of the sector contains one or more Packetized Elementary Stream (PES) packets carrying multiplexed video, audio, or subtitle payloads.
• Video Object Units (VOBUs): Content is grouped into VOBUs, which represent an autonomous unit of video playback (typically 0.4 to 1.0 seconds, beginning with a closed Group of Pictures, or GOP). Every VOBU begins with a specialized Navigation Pack (NV_PCK) that contains Presentation Control Information (PCI) and Data Search Information (DSI).
• Interleaved Metadata: VOB files do not possess an isolated, global index table for the entire video stream. Instead, timing information, stream IDs, and playback navigation pointers are distributed continuously across every sector throughout the entire file.

MP4: Hierarchical Atom (Box) Architecture

The MP4 container format, derived from Apple’s QuickTime File Format and standardized under the ISO Base Media File Format (ISOBMFF / ISO/IEC 14496-12), abandons physical media alignment in favor of an object-oriented, hierarchical structure composed of "atoms" (officially termed "boxes").

• Variable-Length Box Structure: An atom consists of a 4-byte size descriptor, a 4-byte FourCC identifier (such as moov, mdat, or ftyp), and the payload. Unlike VOB sectors, atoms are not restricted to fixed boundaries and can vary in size from 8 bytes to gigabytes.
• Separation of Index and Media Payload: MP4 strictly bifurcates metadata and raw compressed streams:
  • moov (Movie Box): Serves as the central index containing timeline metadata, codec configurations, resolution, frame rates, and sub-boxes such as trak (track) and mdia (media).
  • mdat (Media Data Box): Contains the actual elementary audio and video streams.
• Nested Sample Tables: Within the moov container, the stbl (Sample Table) box holds detailed sub-boxes (stts for time-to-sample, stsz for sample size, and stco/co64 for chunk offsets). These tables map every video frame and audio packet to its exact byte offset within the mdat container.

Key Contrasts Between VOB and MP4 Layouts

1. Allocation: Fixed Alignment vs. Dynamic Nesting

VOB files enforce a rigid, flat, 2048-byte sector structure designed to match DVD-ROM drive buffer sizes and sector-read capabilities. MP4 uses a flexible, tree-like hierarchy of nested containers that can scale arbitrarily, making it independent of any physical storage medium.

2. Metadata Distribution

VOB metadata is interleaved chronologically alongside the media data. Every VOBU contains navigation packs that guide the DVD laser head to relative sector offsets for fast-forwarding or skipping. Conversely, an MP4 file centralizes metadata inside the moov atom. The media in the mdat atom does not require inline headers, timecodes, or packet-level navigation data, as the sample tables in the moov box supply the required byte offsets externally.

3. Seeking and Random Access

Seeking in a VOB requires traversing through sectors to read the DSI tables within navigation packs or reading external DVD control files (.IFO). If an error occurs, playback can often resynchronize at the next recognizable pack header. In an MP4 file, random access is performed mathematically: the playback engine queries the stbl index inside the moov box to identify the exact frame, translates it to a byte offset, and jumps directly to that address in the mdat box. If the moov box is missing or corrupt, an MP4 file cannot be parsed or played without reconstruction.

4. Streaming and Transmission

The sector-based design of VOB makes it a continuous linear stream; a player can begin rendering content the moment it reads consecutive valid packs. In standard MP4 files, the moov atom must be fully read before playback begins. If the moov atom is placed at the end of the file (a common result of single-pass encoding), the entire file must be downloaded before playback can initiate, requiring post-processing tools to shift the moov box to the beginning of the file (progressive download) or fragmenting the layout into moof (Movie Fragment) boxes for adaptive streaming.