Kagami / avif.js

:shipit: AVIF polyfill for the browser
https://kagami.github.io/avif.js/
Creative Commons Zero v1.0 Universal
681 stars 31 forks source link
av1 avif image-compression javascript polyfill


AVIF (AV1 Still Image File Format) polyfill for the browser
Start using superior image compression today! :shipit:
:confetti_ball: :tada: DEMO :tada: :confetti_ball:

Features

Supported browsers

With native decoder:

With AV1 polyfill:

Usage

npm install avif.js
// Put this to reg.js and serve avif-sw.js from web root
// Both scripts should be transpilled (either manually with e.g. browserify or
// automatically by parcel)
require("avif.js").register("/avif-sw.js");
<body>
  <!-- Register worker -->
  <script src="https://github.com/Kagami/avif.js/raw/master/reg.js"></script>

  <!-- Can embed AVIF with IMG tag now -->
  <img src="https://github.com/Kagami/avif.js/raw/master/image.avif">

  <!-- Or via CSS property -->
  <div style="background: url(image2.avif)">
    some content
  </div>
</body>

That's it! Service worker will detect all fetch requests for AVIF files and decode them on the fly. It works with any complex scenarios of image embedding you might have, e.g. background-image in external CSS or XMLHttpRequest from a script.

See demo directory for the usage example.

To generate AVIF files you may use go-avif CLI utility.

Technical details for nerds

AVIF file is basically an AV1 keyframe packed inside ISOBMFF container, almost identical to the HEIF structure, except AV1 video format instead of HEVC is used. Latest versions of Chrome and Firefox support AV1 video decoding, but still can't display AVIF images, it usually takes some time before new format will be added. See e.g. Firefox issue.

Though abovementioned technical aspects of AVIF make it quite easy to implement as a tiny polyfill. All we need to do is repack AVIF as a single-frame AV1 video and decode it using native decoder. This is exactly what avif.js does. First it fetches the AVIF file into binary buffer, then parses the ISOBMFF structure, then searches and extracts the actual frame data (OBUs) and finally embeds it into MP4 video file. Now we can decode that video with standard <video> element and dump raw pixel data to temporary <canvas>.

Instead of forcing users to call some function every time they need to display AVIF file, fetch event interceptor powered by Service Worker API is used. It allows us to replace image data that browser doesn't know how to decode with some known one. avif.js uses BMP to avoid spending time on second compression of already decoded pixel data. It's very fast operation, we just need to write BMP header and copy color values in order. Finally we can deliver that created on the fly .bmp back to the browser and our image will appear on the page.

The actual process is a bit more complex, e.g. we can't create <video> element in a worker, so pass decoding request to the main thread and get result back. Also container fields of the video should correspond to properties of the still image, some offsets need to be fixed and so on. But you got the general idea.

Limitations of Service Worker API

TODO

License

avif.js is licensed under CC0.
Demo images are taken from av1-avif repo.