Role of Auxiliary Items in the AVIF iloc Table

In the AV1 Image File Format (AVIF), the Item Location Box (iloc) acts as the central index detailing exactly where each data payload resides within the container. Secondary auxiliary items—such as alpha channels, depth maps, and HDR gain maps—are stored as discrete items separate from the primary color image. The iloc table defines the physical byte offsets and lengths for these auxiliary payloads, enabling decoders to identify, locate, and access auxiliary data independently without parsing the entire file.

Structural Indexing in ISOBMFF

AVIF is built on the ISO Base Media File Format (ISOBMFF) and HEIF specifications, which treat image components as independent items identified by a unique item_ID. The primary image item contains the visual color payload (typically YUV or RGB), while secondary auxiliary items contain monochrome AV1 bitstreams representing auxiliary data.

The iloc box maps each item_ID to its storage location within the file, whether the data exists inside a Media Data Box (mdat), an Item Data Box (idat), or an external reference. For auxiliary items, the iloc box records:

• Base Offset: The base memory or file position for the item's data.
• Extent Count: The number of separate fragments making up the item.
• Extent Offset and Length: The precise start position and byte length of each payload fragment.

Enabling Non-Destructive Extensions

Auxiliary items are linked to the primary image using the Item Reference Box (iref) with a reference type such as auxl. While the iref box defines the semantic relationship between the auxiliary item and the master image, the iloc table provides the physical access mechanism.

Because the iloc table assigns separate storage extents to auxiliary items:

• Alpha Channels: Transparency data is encoded as an independent monochrome AV1 image with its own item_ID and iloc entry, preventing the need to encode an uncompressed 4-channel image.
• Gain Maps and Tone Mapping: HDR gain maps can be embedded alongside standard dynamic range (SDR) base images, allowing the file to remain backwards-compatible.
• Depth Maps: Spatial and stereoscopic data can be packaged into the same file without altering the primary color stream.

Selective and Progressive Retrieval

A primary benefit of indexing auxiliary items within the iloc table is random and conditional access. Network-based decoders can read the file header, parse the iloc box, and execute targeted byte-range requests.

If a client device does not support or require a secondary auxiliary feature—such as a rendering engine that ignores depth maps or a display that does not render HDR—the parser can skip the byte ranges allocated to those auxiliary items in the iloc table. Conversely, when transparency or HDR is required, the decoder uses the explicit offsets to fetch only the required auxiliary streams.

Facilitating Multi-Threaded Decoding

Because the iloc table provides independent offset and length pairs for the primary image and its auxiliary items prior to decoding, software can dispatch payloads to separate decode pipelines in parallel. A system can decode the primary AV1 color bitstream on one thread while concurrently decoding the alpha or gain map bitstream on another, merging the reconstructed planes in memory prior to display.