JPEG XL: Encoding, decoding and animation with FFmpeg

Introduction

JPEG XL (.jxl) represents the future of image compression, offering superior quality at reduced file sizes. Since version 5.1 (with full animation support added in late 2024/early 2025), FFmpeg natively integrates this format via the libjxl library.

This guide details the essential commands to leverage this modern codec for both static images and animations, including a dedicated section on lossless workflows.

Technical Prerequisites

Before starting, ensure your FFmpeg installation is compatible:

• Required Version: FFmpeg 6.0 or higher is recommended for stable animation support ( libjxl_anim ).
• Compilation: FFmpeg must be compiled with --enable-libjxl.
• Verification: Run ffmpeg -codecs | grep jxl. You should see libjxl (static) and libjxl_anim (animation).

1. Encoding Static Images

Converting to JPEG XL significantly reduces PNG or JPG file sizes while preserving visual quality.

1.1 Standard Lossy Encoding

ffmpeg -i input.png -c:v libjxl output.jxl

• -distance: Controls quality. 1.0 is default (high quality), 0.5 is near-perfect, 15.0 is lowest.
• -effort: Controls compression speed/size trade-off (1 = fastest, 9 = slowest/smallest).

1.2 Lossless Encoding (Static)

For archival or pixel-perfect requirements, set -distance to 0. This forces the encoder into Modular mode, ensuring bit-exact reconstruction.

ffmpeg -i input.png -c:v libjxl -distance 0 -effort 9 output_lossless.jxl

• -distance 0: Mandatory for true lossless compression.
• -effort 9: Highly recommended for lossless. Since quality is fixed, higher effort significantly reduces file size without any quality trade-off, though it increases encoding time.
• Note: You can optionally add -modular 1 to force modular mode explicitly, though -distance 0 does this automatically.

2. Decoding and Conversion

FFmpeg handles conversion from JPEG XL to standard formats seamlessly, preserving color space and bit depth.

ffmpeg -i image.jxl output.png

3. Creating JPEG XL Animations

Replace heavy GIFs with lightweight, high-color-depth animated JPEG XL files. This requires the libjxl_anim codec and the image2pipe format.

3.1 Standard Lossy Animation

ffmpeg -i video.mp4 -c:v libjxl_anim -f image2pipe -distance 0.5 -effort 5 animation.jxl

3.2 Lossless Encoding (Animation)

Creating a truly lossless animation (e.g., for screen recordings or pixel art) follows the same logic as static images but uses the animation codec.

ffmpeg -i source.mov -c:v libjxl_anim -f image2pipe -distance 0 -effort 9 animation_lossless.jxl

• -distance 0: Ensures every frame is encoded without loss.
• -effort 9: Crucial for animations. Lossless video data is large; maximum effort provides the best compression ratio to keep file sizes manageable.
• -f image2pipe: Required to stream frames correctly into the single animated .jxl container.

Performance Tip: Lossless animation encoding is computationally intensive. Using -effort 9 may be slow for long videos. If speed is critical, -effort 7 offers a good compromise with minimal size increase.

4. Summary of Key Parameters

Conclusion

JPEG XL combined with FFmpeg offers a robust, future-proof alternative to GIF and WebP, delivering both high-efficiency lossy compression and true lossless capabilities in a single format.

FFmpeg Options Explanation

-distance: Controls the visual quality (fidelity) of the image.

• 0.0: Mathematically Lossless. The output is bit-for-bit identical to the input; no data is discarded.
• 0.5: Near-Perfect. Higher quality than default; artifacts are virtually non-existent even under close inspection (often used for archival).
• 1.0: Default. High quality; generally considered "visually lossless" for standard viewing (artifacts are imperceptible to the human eye).
• 15.0: Lowest Quality. Maximum compression with visible artifacts and blurring.

-effort: Controls the compression speed vs. file size trade-off (how hard the encoder works).

• 1: Fastest. Minimal compression; largest file sizes. Ideal for quick previews or real-time encoding.
• 4–5: Balanced. The default range; offers a good compromise between encoding speed and file size reduction.
• 7: High Compression. Significantly smaller files with a noticeable increase in encoding time. Recommended for web delivery and archives where size matters.
• 9: Slowest. Maximum compression; smallest possible file sizes. Encoding can be extremely slow. Best for "write-once, read-many" archival storage where file size is the only priority.

Note: Unlike -distance, increasing -effort never reduces quality; it only increases the time required to encode.

Regarding Alpha (Transparency): If your input image has transparency (like a PNG), JPEG XL preserves it automatically. There is a separate parameter, -alpha_distance, if you wish to compress the transparency channel differently from the color data, but -distance only affects the color/visual fidelity.

Links

• Official Website
• Community Website
• Wikipedia