Nokia has outlined new HEIF capabilities for everything from HDR images to map data, and reports an average 32.39% bitrate reduction for a lower-overhead version tested on small images. It has also made reference source code available, giving developers a practical way to explore the format rather than merely admire the acronym.
Nokia Watch analysis
What happened
HEIF is a container for images and their related information, designed to work with different compression technologies. Nokia says recent work includes support for VVC/H.266 and Low-overhead HEIF, aimed at small assets such as icons, stickers and thumbnails, where container metadata can take up a meaningful share of the file.
The third edition adds progressive rendering, region-linked annotations and camera parameters. Other described capabilities include HDR gain maps, image collections and non-destructive operations such as cropping and rotation. Extensions for tiled, multi-resolution imagery let a viewer request a particular region and resolution rather than processing a very large image in full.
Nokia’s evaluation compared Low-overhead HEIF using compact VVC configuration with conventional HEIF using regular VVC configuration on images from the JPEG AI test dataset. The reported bitrate reduction averaged 32.39% across the tested sizes, ranging from 67.32% at 8 × 8 pixels to 3.12% at 256 × 256 pixels. These are Nokia’s test results, not a claim that every image or workflow will see the same saving.
Nokia has also published reference code, including reader and writer libraries, Java APIs and support for VVC-coded images and Low-overhead HEIF. Its accompanying conformance files and examples are intended to help developers test implementations.
Why it matters
Image formats have to do more than squeeze pixels into fewer bytes. They also carry the metadata and related content needed for modern photography, HDR, maps and other imaging workflows. A single container that can support those uses, while adopting different codecs, could make it easier to add capabilities without rebuilding the surrounding workflow.
The small-image results address a less glamorous but useful problem: when an image is tiny, metadata overhead can be a sizeable part of the total file. The reported savings are strongest at the smallest tested resolutions, so the figures are most relevant to lightweight assets, not a blanket promise of smaller photo libraries.
Our read
The interesting part is the breadth: HEIF is being presented as a flexible imaging framework, not just another file extension. Public code and test files make the work easier for developers to examine and try. Nokia’s benchmark is promising, but the practical test will be whether independent implementations interoperate and whether the size gains hold for real-world images.
What to watch
- Whether developers adopt the reference implementation and report interoperability results.
- How Low-overhead HEIF performs on common web and app assets beyond the test dataset.
- Whether support for the format and its newer features spreads across software and devices. Activity teaser: HEIF is growing beyond image storage, with support for features including HDR gain maps, progressive rendering and very large tiled imagery. Nokia also reports that its lower-overhead version cut bitrate by an average 32.39% in tests on small images, and has published reference code for developers. The savings look most striking at tiny resolutions, where metadata can be a sizeable slice of the file. Should developers back one flexible image container for all these jobs, or keep specialised formats for specialised uses?
Discussion spark: Should developers back one flexible image container for everything from HDR to tiny web images, or keep specialised formats for specialised jobs?
Sources and evidence
- Source update (28 October 2026, 07:59 UTC)
not affiliated with or endorsed by Nokia