Through 2021 we steadily saw the WebP image format grow in popularity. We process thousands of jobs daily across image analysis, conversion and sanitising, and while JPEG remains by far the most common format, the rise in WebP files is unmistakable.
What is the WebP image format?
WebP is an image format developed by Google, first announced in September 2010 as an open format for lossy compressed images. In 2011 Google extended it to support lossless compression as well.
It supports lossy and lossless compression, animation, transparency, colour profiles and embedded metadata.
What is so good about it?
WebP files are generally smaller than the equivalent JPEG or PNG.
JPEG is the workhorse for displaying content on the web, and WebP files are typically around 25 to 35% smaller. That comparison only means something if you hold quality constant, which is why the SSIM index matters — it compares images on the similarity a human eye actually perceives, rather than on file size alone.
PNG files, although larger than JPEG, remain popular where part of the image needs to be transparent — where you want background content to show through when images are stacked in layers. WebP supports transparency too, and is generally around 25% smaller.
As of May 2021, over 94% of web browsers supported WebP. That makes it a serious contender against the established formats, because it will display for the overwhelming majority of visitors.
There is some irony in the fact that at the time of writing, every image on our own clearnet and dark web sites is a JPEG or PNG. Moving that content to WebP is on the list, but it has to work everywhere — desktop, mobile, and the Tor Browser via the onion site — before it is worth doing.
What about metadata, privacy and steganography?
This is the part I care most about. Whatever the format, the same questions apply:
Photos and videos may contain personal and private information. Is it being tracked, or leaking personal information? Does it contain metadata or steganography? Have copyrights and digital watermarks been added? Is the media in the best optimised format before sharing online? Are you sharing content safely?
WebP can carry metadata, and it does one thing rather better than its rivals: the WebP header declares up front whether metadata exists in the file. That is a marked contrast to JPEG and PNG, where you have to go looking, with no consistent specification for where such sections should appear or what they should contain.
WebP supports EXIF, ICCP and XMP metadata, with payloads much like their JPEG and PNG equivalents.
Steganography concerns are unchanged by the format, because the same principles apply. For anyone unfamiliar with the term, steganography in digital media means concealing a file, message, image, sound or video inside another file. It is quite different from hidden metadata: the data is encoded into the image data itself, often encrypted, and effectively invisible. I have written more about that in how photo steganography works.
WebP and the future
If the trend continues, WebP will overtake JPEG for new content. The argument is simply smaller files at the same image quality — which means less bandwidth and a faster experience, because pages render sooner when their images arrive sooner. With browser support essentially universal since mid-2021, there is little standing in the way.
If you have observations about WebP, particularly around metadata and privacy, I would be interested to hear them. You can also submit images for analysis at snapWONDERS.