Google Announces Guetzli Open Source JPEG Encoder That Reduces Image Size By 35 Percent
According to Google, “It is our hope that webmasters and graphic designers will find Guetzli useful and apply it to their photographic content, making users’ experience smoother on image-heavy websites in addition to reducing load times and bandwidth costs for mobile users”.
Guetzli, or “cookie” in Swiss German”, allows developers to produce JPEGs that are smaller, but still compatible with existing browsers, image processing applications, and the JPEG standard. Guetzli is very similar to Google’s Zopfli algorithm, which creates smaller PNG and gzip files. Recurrent Neural Network (RNN)-based image compression, Rapid and Accurate Image Super-Resolution (RAISR), and WebP vary slightly from Guetzli and Zopfli because they require a client change.
How does Guetzli work? The visual quality of a JPEG image is related to its compression process. Guetzli focuses upon the quantization stage which compresses a range of values to a single quantum value. Color quantization lessens the number of colors used in a single image. Guetzli’s psychovisual model “approximates color perception and visual masking in a more thorough and detailed way than what is achievable by simpler color transforms and the discrete cosine transform. “ It utilizes a search algorithm that attempts to reconcile the differences between the the JPEG format’s and Guetzli’s psychovisual model.
According to Google, testers actually preferred the Guetzli JPEGs despite their smaller size. Testers found the Guetzli JPEGs to be of a higher quality. Like Zopfli, the downside is that search algorithms take significantly longer to create compressed images. Google argues, however, that the longer processing time is well worth the better quality and smaller image file sizes.