AVIF Baseline Profile: Codec and Dimension Limits

The AVIF Baseline Profile establishes standardized limits on AV1 codec configurations and image dimensions to guarantee widespread interoperability across web browsers and resource-constrained hardware. By restricting the underlying AV1 bitstream to specific profiles and performance levels, the specification ensures that decoders can reliably process images without excessive memory or compute overhead. This article breaks down the explicit constraints enforced by the AVIF Baseline Profile, including codec profile requirements, bit depth, chroma subsampling limits, and dimensional ceilings governed by AV1 Level 5.1.

AV1 Codec Profile Restriction

The AVIF specification defines two primary profiles: Baseline (avif) and Advanced (avis). The Baseline Profile strictly requires the underlying AV1 bitstream to conform to the AV1 Main Profile (seq_profile = 0).

By mandating Profile 0, the Baseline Profile deliberately excludes:

• AV1 High Profile (Profile 1): Which adds support for 4:4:4 chroma subsampling.
• AV1 Professional Profile (Profile 2): Which adds support for 12-bit color depths and 4:2:2 chroma subsampling.

Restricting baseline files to the Main Profile ensures that common hardware video decoders—which overwhelmingly target Profile 0 for streaming video—can decode AVIF still images without falling back to slower software implementations.

Bit Depth and Chroma Subsampling Constraints

Because the Baseline Profile requires AV1 Main Profile conformance, it imposes strict boundaries on pixel formats:

• Bit Depth: Allowed bit depths are limited to 8-bit and 10-bit. 12-bit images are prohibited in the Baseline Profile and must be encoded using the Advanced Profile.
• Chroma Subsampling: Only 4:2:0 and 4:0:0 (Monochrome) formats are permitted.
• Prohibited Formats: 4:2:2 and 4:4:4 sampling modes cannot be used in baseline AVIF files.

For full-color photographic images, 4:2:0 is the default. If an image requires an alpha channel (transparency), it is encoded as an auxiliary monochrome (4:0:0) image item that runs parallel to the primary color item, matching its dimensions.

Decoder Level and Resolution Limits

The Baseline Profile constrains the maximum complexity of an image by capping the operating point level. Specifically, the AV1 sequence header parameter seq_level_idx_0 must be set to Level 5.1 or lower (index 19 or below).

AV1 Level 5.1 enforces strict bounds on single-frame dimensions and total pixel count:

• Maximum Picture Size (MaxPicSizeLuma): 8,912,896 luma samples (approximately 8.9 megapixels). A standard 4K frame (3840 × 2160 = 8,294,400 pixels) falls safely within this limit.
• Maximum Frame Width (MaxHSizeLuma): 8,192 pixels.
• Maximum Frame Height (MaxVSizeLuma): 4,352 pixels.

Single-coded frames exceeding these limits cannot be marked with the baseline AVIF brand.

Bypassing Resolution Caps via Grid Items

While a single coded frame is limited to approximately 8.9 megapixels under Level 5.1, the AVIF specification allows baseline files to display higher overall resolutions through image grids (grid derived image items).

When an image exceeds Level 5.1 dimensions:

1. The image is split into smaller rectangular tiles.
2. Each tile is encoded as an independent AV1 image item conforming to Main Profile and Level 5.1 limits.
3. The container's grid item instructs the parser to reconstruct the complete image from the individual tiles at display time.

This mechanism enables the baseline profile to support ultra-high-resolution images while preserving compatibility with hardware decoders that have strict single-frame buffer limits.