This is not the first time when Google has done something new regarding image compression, however, among all of its other algorithms, this one is compatible with most of the browser, apps, and JPEG standards.
According to Ars Technica, the encoder technically works by improving the quality of the JPEG images while keeping the size of the file constant. The new compression algorithm, called Guetzli, can also create images that are higher quality than existing compression algorithms for a given file size, the company said.
In the example below, you can see the uncompressed original image on the left, libjpeg (another image compression tool) in the middle and Guetzli on the right.
Google has come up with a new algorithm that aims at reducing the JPEG file size by 35%. "This implies the Butteraugli psychovisual image similarity metric which guides Guetzli is reasonably close to human perception at high-quality levels".
The new encoder is called Guetzli - Swiss German for "cookie", apparently - and according to Google, it can create "high quality JPEG images with file sizes 35 percent smaller than now available methods". Their Guetzli algorithm targets the quantization stage to reduce the size of image files by swapping it out for loss in visual quality.
The JPEG encoder is open-source and available for you to download and implement in your own projects from this GitHub repository. The time between shots might be reduced if Guetzli compression is used.
The underlying compression algorithms behind popular implementations of JPEG have always been lossy; once a web version has been generated from a high-quality original master, the lost pixel information can not be recovered (except, arguably by AI - another field of interest for Google's image researchers). Right: Guetzli. Google claims that Guetzli has fewer artifacts without a larger file size. The only tradeoff: Guetzli takes a little longer to run than compression options like libjpeg.
Guetzli employs a search algorithm that tries to overcome the difference between the psychovisual modeling of JPEG's format and Guetzli's psychovisual model. "Guetzli is rather slow to encode", the researchers said, suggesting it's most likely useful on image-heavy websites.