Cloudinary Uses AI to Further Improve Image Quality Online, Announces JPEG XL Support for Apple Ecosystem

Cloudinary, the image and video platform that powers many of the world’s top brands, today launched a new AI-enhanced image quality feature and expanded JPEG XL support. These two new features enable developers to build and deliver high-quality and high-performing visual experiences.

Cloudinary’s automatic quality selection capability (q_auto) is one of the company’s most popular product features, used by tens of thousands of users worldwide. It analyzes the pixels of an image to find the most suitable quality level and encoding settings based on user device and bandwidth. The AI-enhanced q_auto update announced today takes this process to a whole new level beyond Cloudinary’s previous industry leading version.

Training AI for even smarter image compression

The evolution of q_auto has been fueled by Cloudinary’s extensive research into the impact of compression on perceived image quality. A key output of this is the Cloudinary Image Dataset (CID22) of human-annotated compressed images. Critical to building this dataset, humans annotated their perception of quality in relation to different compression scenarios. For the enhanced q_auto transformation, an AI model was trained on these observations. Through this Cloudinary found that the average image size could be reduced even further than previously calculated without any visible impact on quality, further minimizing bandwidth.

Apple’s support for JPEG XL helps fuel adoption

Cloudinary’s lead image researcher, Dr. Jon Sneyers, along with several researchers at Google, helped create JPEG XL, which delivers high-quality images on the modern web, offering high-fidelity compression and progressive rendering. Following Apple’s announcement last June that it would support JPEG XL in iOS 17, macOS Sonoma and the Safari browser, Cloudinary’s customers can now deliver unparalleled experiences to their users on Apple devices.

The benefits of JPEG XL include:

• Significant reduction in file size while maintaining high quality, providing a critical balance between performance and quality;
• Faster encoding/decoding (compared to AVIF) for increased performance;
• Progressive decoding to satisfy impatient web and mobile application visitors while providing a smoother experience;
• Designed for high-fidelity photographic use;
• Flexible lossless or lossy compression modes; and
• Existing JPEG images can be converted to JPEG XL with no generation loss – no other new image format can do this. JPEG XL files are 20% smaller and can be converted back to the exact same JPEG file.

“We’re always committed to applying the latest technologies and formats that empower our users to deliver superior visual content that helps them meet their business goals,” said Shelby Britton, senior director of Product Marketing at Cloudinary. “The enhancements we’re introducing today exemplify this commitment: an AI-driven quality selection that further boosts image quality consistency and performance, along with the seamless integration of JPEG XL for Apple users. Together these set a higher bar for image optimization, driving higher resource efficiency, performance, and visual appeal.”

Photo credit: Cloudinary