A Look Inside Google’s New Image Format
Last week, Google announced that it has developed a new image file format that promises to speed up not only Google, but the rest of the Internet. The new format, dubbed WebP, could reduce the size of image downloads significantly compared to the currently widespread JPEG format. If, as Google claims, 65 percent of the traffic on the Internet consists of image files, WebP could eventually give the entire online world a speed boost.

WebP is based on a new data compression algorithm for image files. According to Google product manager Richard Rabbat, the formula was originally developed to compress video clips, such as those Google serves on YouTube. Engineers then found that the same math could be used on static images.
It works like this: software converts an image stored in another file format to the WebP format by looking at a block of 16 pixels in the original image, and noting the color of each pixel in that block. The software then predicts what the pixel colors in an adjacent block of 16 pixels will be, based on the first block. Finally, it compares the predicted values to the actual colors of the pixels in the next block. If the WebP algorithm’s educated guess is correct, no additional data needs to be added to the compressed image file. Only those pixels whose colors are different from the prediction are recorded in the file.
The result is that fewer pixel values need to be stored in most WebP image files than in a JPEG file of the same image. When an Internet user with a WebP-enabled browser downloads the image, his browser would use the same formula to predict many of the pixel values, and then fill in those pixels stored in the WebP file that didn’t match the algorithm’s prediction. Instead of downloading a lot of pixels, the WebP browser would mathematically calculate their colors. Predicting pixel colors turns out to be faster than downloading them, even on a high-speed Internet connection.
That’s the theory, but how does WebP fare in real-world practice? In tests on one million images selected from the Internet at random, Google software developers claim that converting them to WebP format led to an average 40 percent reduction in the image’s file size. The company has set up a demo page that shows typical Web images and lists how much smaller the same image would be in WebP, without visible changes to the quality of the image.
For now, you can’t use WebP. The predictive algorithms need to be built into your browser. Google is adding WebP support to its own browser, Chrome. But realistically, WebP will need to be supported by nearly all of the world’s browsers - that means Internet Explorer, Firefox and Safari - before website operators will have an incentive to start serving WebP images en masse. It’s a long-term project, not a quick fix.
Speed has always been a mantra at Google. Faster Web pages lead to more Googling by Web users. This, in turn, leads to more ads served by Google, which leads to potentially more ad revenue for the company. In a 2007 test conducted live on the site for several weeks, Google claims to have found that slowing its response to a search query from the average 400 milliseconds to 500 milliseconds lowered test subjects’ search engine usage enough that (had the delay been real across the entire site) it would have cost Google $900 million in lost revenue over a year.
That’s one of the reasons the company recently began serving search results to users’ screens while they are still typing their search terms. If WebP gets built into everyone’s browsers and the format is embraced by website operators, Google’s traffic - and everyone else’s – seems almost sure to go up.
Keep Reading
Most Popular
The inside story of how ChatGPT was built from the people who made it
Exclusive conversations that take us behind the scenes of a cultural phenomenon.
How Rust went from a side project to the world’s most-loved programming language
For decades, coders wrote critical systems in C and C++. Now they turn to Rust.
Design thinking was supposed to fix the world. Where did it go wrong?
An approach that promised to democratize design may have done the opposite.
Sam Altman invested $180 million into a company trying to delay death
Can anti-aging breakthroughs add 10 healthy years to the human life span? The CEO of OpenAI is paying to find out.
Stay connected
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.